pulsecat (OP)
|
|
July 23, 2013, 10:25:02 AM Last edit: February 22, 2014, 04:55:19 PM by pulsecat |
|
ATC - is a competition where trading algorithms compete against each other at some period of time. The algorithm that performs better than others during the contest trading period, will earn the prize. Round #5 "Best automated strategy of 2013"12/01 - 12/31Prize: 0.5 BTC----------------------------- Rules- Any Bitcointalk user with forum registration date on or before 11/30 is eligible for pariticipation - In order to participate you need to submit a link to the algorithm created at CryptoTrader.org- All code is backtested on historical MtGox data at the trading frequency of your choice (5m,10m,15m,etc). - Each participant is allowed to submit one algorithm only- Only algorithms created before 12/01 00:00 GMT are eligible for participation. - Any Cryptotrader API functionality can be used. The documentation is available under the "Help" tab at the website. - The code should not contain any references to specific date or price. - An algorithm must be profitable at the period 2013/01/01 00:00 GMT - 2014/01/01 00:00 GMT. - Until 12/01 00:00 GMT participants post links to their backtest results (link should look like: cryptotrader.org/backtests/<session-id>) to this thread so that the winner can be further determined. At 2014/01/01 all participating scripts will be re-run with initial cash deposit set to 5000 USD and fee 0.5%. The author of the algorithm that performed best will be awarded with 0.5 BTC + donated coins if any. - Any kind of cheating is against the rules. Feel free to ask any questions here. If you want to report a problem, please don't forget to provide link to the backtest session you used. Have fun! Round #5 participant list Grinny https://cryptotrader.org/backtests/AzhmEhWnxuPHSFadXZirconiumX https://cryptotrader.org/backtests/n8mxhhMcfviBjCLcoexuals https://cryptotrader.org/backtests/bQzucyJursAiv2cgFmedji https://cryptotrader.org/backtests/Wq8vb2a4LzGWeW79QKiri11 https://cryptotrader.org/backtests/mkLghthyTHuTXt3Te]sukiho https://cryptotrader.org/backtests/uKQgjtPd2XqXLjCEritod https://cryptotrader.org/backtests/69cLtTudmkebzFLG2MeatPopsicle https://cryptotrader.org/backtests/Wuz7Tgstb3XNqezjq
|
|
|
|
|
|
|
Each block is stacked on top of the previous one. Adding another block to the top makes all lower blocks more difficult to remove: there is more "weight" above each block. A transaction in a block 6 blocks deep (6 confirmations) will be very difficult to remove.
|
|
|
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction.
|
|
|
Diabolicus
Member
Offline
Activity: 90
Merit: 10
|
|
July 23, 2013, 11:12:06 AM |
|
Count me in :-) May we submit more than one algorithm?
|
|
|
|
pulsecat (OP)
|
|
July 23, 2013, 11:20:12 AM Last edit: August 12, 2013, 06:16:22 AM by pulsecat |
|
Yes, this is allowed. Everything which is not forbidden is allowed :-) Updated 08/12 Not alllowed for Round #3 and further rounds
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
July 24, 2013, 01:00:37 AM |
|
Yay! time to 'hurry-up-n-wait' for 07/28 + a few hours so I dont derp and miss it again
|
|
|
|
|
pulsecat (OP)
|
|
July 25, 2013, 10:46:03 AM Last edit: July 25, 2013, 11:03:01 AM by pulsecat |
|
Such a high frequency strategy is something fresh. You are in!
|
|
|
|
Diabolicus
Member
Offline
Activity: 90
Merit: 10
|
|
July 26, 2013, 10:20:24 AM |
|
Everyone trying hard not to blink before 07/28 23:59 GMT, huh? Here's my first entry: https://cryptotrader.org/backtests/6hHNDRtRmy5j9mAnPBasic 1hr EMA cross strategy, combined with a more or less elegant RSI implementation to filter out weak signals. There's also a rudimentary stop loss and trailing stop loss which I find quite useful. It took me a while to get the hang of this whole scripting thing (after all I'm an architect, not a programmer), so I left in a few comments and unused indicators that might be useful to others. I left in my statistics summary, too. Maybe someone else can improve the code and pulsecat can add such a statistic as a default output? Feel free to canibalize! Now let's see if MACD and MFI or CCI can be more efficient.
|
|
|
|
pulsecat (OP)
|
|
July 26, 2013, 04:54:06 PM |
|
Everyone trying hard not to blink before 07/28 23:59 GMT, huh? Here's my first entry: https://cryptotrader.org/backtests/6hHNDRtRmy5j9mAnPBasic 1hr EMA cross strategy, combined with a more or less elegant RSI implementation to filter out weak signals. There's also a rudimentary stop loss and trailing stop loss which I find quite useful. It took me a while to get the hang of this whole scripting thing (after all I'm an architect, not a programmer), so I left in a few comments and unused indicators that might be useful to others. I left in my statistics summary, too. Maybe someone else can improve the code and pulsecat can add such a statistic as a default output? Feel free to canibalize! Now let's see if MACD and MFI or CCI can be more efficient. A good example of how various TA-Lib indicators can be used. And i agree that reporting needs to be improved. Thanks.
|
|
|
|
cp1
|
|
July 26, 2013, 10:58:09 PM |
|
Do you have to sell at the end, or can your end balance be in BTC?
|
|
|
|
medji
Newbie
Offline
Activity: 31
Merit: 0
|
|
July 26, 2013, 11:41:56 PM Last edit: July 26, 2013, 11:58:23 PM by medji |
|
Can you only access the price data within the testing time frame (7 days)? It might not be enough data to bootstrap some indicators until they are reliable
|
|
|
|
pulsecat (OP)
|
|
July 27, 2013, 06:07:15 AM |
|
Do you have to sell at the end, or can your end balance be in BTC?
You don't have to sell. Conveniently, the resulting amount of BTC will be displayed in USD. Can you only access the price data within the testing time frame (7 days)? It might not be enough data to bootstrap some indicators until they are reliable
At the start, each algorithm is pre-initialized with 100 bars of history trading data. This should be enough for most indicators. If not, please let us know.
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
July 28, 2013, 04:54:58 AM |
|
Ok, I'll throw mine in now. Didnt have as much time as I thought to play with it Anyway, half arsed attempt at following a really short ema slope's rate of change, and direction change. Thought process was the rate of increase / decrease would have to slow down before it switches direction. Watch for the slow-down in the rise or fall and buy/sell when the direction change happens. 30 min window. https://cryptotrader.org/backtests/dQy9CXc5CLiqadvsn
|
|
|
|
cp1
|
|
July 28, 2013, 05:25:05 AM Last edit: July 28, 2013, 05:54:59 AM by cp1 |
|
|
|
|
|
medji
Newbie
Offline
Activity: 31
Merit: 0
|
|
July 28, 2013, 09:24:29 AM |
|
|
|
|
|
pulsecat (OP)
|
|
July 28, 2013, 04:49:56 PM Last edit: July 29, 2013, 03:39:16 AM by pulsecat |
|
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
July 30, 2013, 12:47:27 AM |
|
lol not gonna lie one bit... I kinda feel a bit dense and ham-fisted comparing mine to the others submitted Some day I'll write decent code... some day
|
|
|
|
cp1
|
|
July 30, 2013, 12:54:30 AM |
|
lol not gonna lie one bit... I kinda feel a bit dense and ham-fisted comparing mine to the others submitted Some day I'll write decent code... some day If it works it works! I meant to ask earlier -- are there switch/case statements in this language?
|
|
|
|
pulsecat (OP)
|
|
July 30, 2013, 09:17:30 AM |
|
lol not gonna lie one bit... I kinda feel a bit dense and ham-fisted comparing mine to the others submitted Some day I'll write decent code... some day If it works it works! I meant to ask earlier -- are there switch/case statements in this language? CoffeeScript provides all of JavaScript's functionality and it has great documentation http://coffeescript.org/#switch
|
|
|
|
cp1
|
|
July 30, 2013, 08:54:30 PM |
|
Thanks for that link, figured out my problem. But I found a stupid error in my code! I'll just have to hope that BTC dips in the next few days.
|
|
|
|
cp1
|
|
July 31, 2013, 05:36:46 AM |
|
Do any of the api functions calculate first and second derivative? All the talibs that I looked at had weird names.
|
|
|
|
pulsecat (OP)
|
|
July 31, 2013, 05:31:58 PM Last edit: July 31, 2013, 05:44:12 PM by pulsecat |
|
Do any of the api functions calculate first and second derivative? All the talibs that I looked at had weird names.
Well, I'm not an expert in Ta-lib, so i can't really help. These links might be useful http://tadoc.org/http://www.ta-lib.org/d_api/d_api.html
|
|
|
|
dexX7
Legendary
Offline
Activity: 1106
Merit: 1024
|
|
July 31, 2013, 05:43:59 PM |
|
So this ended already? Will there be a round 3? Very interesting.. How is cheating defined? Over optimization which essentially would mirror the historical price movement or something?
|
|
|
|
cp1
|
|
July 31, 2013, 05:59:12 PM |
|
So this ended already? Will there be a round 3? Very interesting.. How is cheating defined? Over optimization which essentially would mirror the historical price movement or something? He runs it on data that came out after you submitted your algorithm.
|
|
|
|
1Pakis
|
|
August 01, 2013, 06:08:19 AM |
|
One more clarification because it makes some deference to my algorithm to when it starts. Is it 07/29 00:00 GMT as in BST or 07/29 00:00 GMT as in UTC? Why nobody gives times in UTC in this forum? Every body gives times in their local timezone which causes grade confusion now that is summer in the northern hemisphere. Try to stick with UTC next time.
|
Tips are welcome at this address 18DVZkpSwmejPjekX3QMKvRRtR8Bfx65LN.
|
|
|
pulsecat (OP)
|
|
August 01, 2013, 03:37:24 PM |
|
|
|
|
|
cp1
|
|
August 01, 2013, 03:39:10 PM |
|
Looks like buy and hold is going to win unless there's a downward dip.
|
|
|
|
pulsecat (OP)
|
|
August 01, 2013, 03:50:21 PM |
|
The market is unpredictable, so who knows.
|
|
|
|
AstroBoy
Newbie
Offline
Activity: 50
Merit: 0
|
|
August 02, 2013, 10:27:17 AM |
|
The market is unpredictable, so who knows.
True, but I bet you it will take a downward dip in less then 7 days. You can't time the market, I know this, but I just have a feeling. I hope it is sooner though and gets captured in this contest. I'd like to start running an algo soon, this has been really informative. Cheers!
|
|
|
|
Diabolicus
Member
Offline
Activity: 90
Merit: 10
|
|
August 02, 2013, 01:58:55 PM |
|
I mixed up previous_price with buy_price, so the trailing stop loss isn't really trailing :-/ ah well ... next time
|
|
|
|
genydeny
Newbie
Offline
Activity: 50
Merit: 0
|
|
August 04, 2013, 08:27:52 AM |
|
Hi to all,this is a great contest! As you can see, all of the above strategies behave incorrectly when there is a flat trend, which was established on August 2th. Apparently, they may be improved by filtration (to reduce the number of losing trades). Any thoughts?
|
|
|
|
Diabolicus
Member
Offline
Activity: 90
Merit: 10
|
|
August 05, 2013, 06:52:02 AM |
|
Buy & Hold did it again, I presume?
Give us round 3 before my vacation :-)
|
|
|
|
genydeny
Newbie
Offline
Activity: 50
Merit: 0
|
|
August 05, 2013, 07:18:54 AM |
|
|
|
|
|
|
Diabolicus
Member
Offline
Activity: 90
Merit: 10
|
|
August 05, 2013, 09:19:26 AM |
|
A single algorithm per participant would be wise, otherwise people might just submit two completely different solutions, one for bear markets and one for bull (only laziness stopped me from doing so last time). However that would also require the contest being limited to already registered forum members only (as of today), to prevent people from cheating by registering new accounts.
With 4 weeks trading period we should definitely see a winner next time.
Additional rule suggestion: - winning algorithm must not only outperform Buy & Hold, but also Not-Buy-At-All, i.e. the result must exceed the initial 1000$. (I think a strategy should not win if it doesn't increase the value. Just losing less than B&H is simply not enough imo.)
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
August 05, 2013, 10:50:52 AM |
|
However that would also require the contest being limited to already registered forum members only (as of today), to prevent people from cheating by registering new accounts. Hey, give me a chance to register, I didn't already cause I thought you guys are going to nail it, who new you wont succeed and there will be a third round? Thanks. Edit: Registered.
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
August 05, 2013, 10:57:45 AM |
|
May I also ask for a few days (or a week) to study previous rounds algorithms and why they failed?
|
|
|
|
pulsecat (OP)
|
|
August 05, 2013, 03:14:22 PM |
|
A single algorithm per participant would be wise, otherwise people might just submit two completely different solutions, one for bear markets and one for bull (only laziness stopped me from doing so last time). However that would also require the contest being limited to already registered forum members only (as of today), to prevent people from cheating by registering new accounts.
With 4 weeks trading period we should definitely see a winner next time.
Additional rule suggestion: - winning algorithm must not only outperform Buy & Hold, but also Not-Buy-At-All, i.e. the result must exceed the initial 1000$. (I think a strategy should not win if it doesn't increase the value. Just losing less than B&H is simply not enough imo.)
These are all worthy suggestions. Thanks! May I also ask for a few days (or a week) to study previous rounds algorithms and why they failed?
Don't worry, before a trading period begins, each participant will have a few days for developing his strategy.
|
|
|
|
safinaskar
Newbie
Offline
Activity: 41
Merit: 0
|
|
August 10, 2013, 10:58:34 AM |
|
Please run round #3!
|
|
|
|
pulsecat (OP)
|
|
August 10, 2013, 02:40:17 PM |
|
First i need to make sure we will have enough entries to start. Who is in round 3?
|
|
|
|
cp1
|
|
August 10, 2013, 04:15:46 PM |
|
In!
|
|
|
|
Diabolicus
Member
Offline
Activity: 90
Merit: 10
|
|
August 10, 2013, 07:56:25 PM |
|
Me too, as long as I got Wi-Fi on Tenerife
|
|
|
|
safinaskar
Newbie
Offline
Activity: 41
Merit: 0
|
|
August 11, 2013, 06:16:48 AM |
|
in!
|
|
|
|
2dogs
Legendary
Offline
Activity: 1267
Merit: 1000
|
|
August 11, 2013, 05:18:00 PM |
|
checking it out
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
August 11, 2013, 11:25:34 PM |
|
I'll throw my mess in again, whether or not I get it cleaned up at all
|
|
|
|
|
Diabolicus
Member
Offline
Activity: 90
Merit: 10
|
|
August 13, 2013, 06:02:42 PM |
|
Guess I'll have to be the first to submit for round 3, since my flight leaves tomorrow and I won't take my notebook with me (and typing scripts for cryptotrader.org on a cell phone is kind of a tedious task ...) So here we go: https://cryptotrader.org/backtests/mwpRw9xgA7qbqt8x3The basic idea was to have a 2h 10/21 EMA cross strategy, but use 1h candles and 20/42 EMAs instead to react faster to sudden price changes. RSI and ADX help filter out false signals. Had a few more ideas, but couldn't bring it all together in time, maybe next round.
|
|
|
|
pulsecat (OP)
|
|
August 17, 2013, 08:43:53 AM |
|
Guess I'll have to be the first to submit for round 3, since my flight leaves tomorrow and I won't take my notebook with me (and typing scripts for cryptotrader.org on a cell phone is kind of a tedious task ...) So here we go: https://cryptotrader.org/backtests/mwpRw9xgA7qbqt8x3The basic idea was to have a 2h 10/21 EMA cross strategy, but use 1h candles and 20/42 EMAs instead to react faster to sudden price changes. RSI and ADX help filter out false signals. Had a few more ideas, but couldn't bring it all together in time, maybe next round. Thanks for submission. Hope you are having a nice trip. Just to remind everyone - the trading period starts in less than 2 days.
|
|
|
|
sukiho
|
|
August 17, 2013, 09:02:26 PM |
|
thanks for setting this up. will you be using a 0.55% trading fee for the reruns?
|
|
|
|
medji
Newbie
Offline
Activity: 31
Merit: 0
|
|
August 18, 2013, 01:14:15 AM |
|
|
|
|
|
|
panic
|
|
August 18, 2013, 01:53:27 PM |
|
|
...too much panic and too little reason
|
|
|
pulsecat (OP)
|
|
August 18, 2013, 06:04:06 PM |
|
thanks for setting this up. will you be using a 0.55% trading fee for the reruns?
Correct, all of strategies will be backtested using 0.55% fee. Four entries submitted so far, six hours until the submission period ends.
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
August 18, 2013, 11:54:52 PM |
|
https://cryptotrader.org/backtests/qyYudoimmtTEupchJCCI enable with [high|low]-value ROC trigger mostly learning how to use talib functions in coffeescript :-) edit : damn forgot about including a method for stop-loss lol
|
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
August 19, 2013, 01:09:43 AM |
|
I'm enjoying seeing the variety of styles people use, both from a technical standpoint, and methodology standpoint. Also nice being able to see real examples of specific methods in use.
|
|
|
|
pulsecat (OP)
|
|
August 19, 2013, 05:31:53 AM |
|
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
August 20, 2013, 11:22:58 PM |
|
for what it's worth, the following is the same as my actual entry for the contest, but with added round debug info. https://cryptotrader.org/traders/q75spqHf6SWTDaid9Strategy is the same, but I found being able to step through the rounds and review what it was doing for each was enlightening. example: 013-08-19 09:30 Trader started. Balance: 1000.00 USD 2013-08-19 09:45 ROUND ==> 1 | HOLDING: usd=1000 btc=0 2013-08-19 09:45 cciValue =13.65474704470624 2013-08-19 09:45 highRocValue = -8.403361344537785 | lowRocValue = -9.888532448760534 2013-08-19 09:45 high =118 | low =117.25 | close =117.39002 2013-08-19 09:45 @ @ hold cci @ @ 2013-08-19 09:45 ************* 2013-08-19 09:45 ************* 2013-08-19 10:00 ROUND ==> 2 | HOLDING: usd=1000 btc=0 2013-08-19 10:00 cciValue =9.35064812180757 2013-08-19 10:00 highRocValue = -9.075714285714298 | lowRocValue = -7.711210914329358 2013-08-19 10:00 high =117.91999 | low =117.1 | close =117.25 2013-08-19 10:00 @ @ hold cci @ @
|
|
|
|
pulsecat (OP)
|
|
August 21, 2013, 04:58:13 PM Last edit: August 21, 2013, 06:21:23 PM by pulsecat |
|
for what it's worth, the following is the same as my actual entry for the contest, but with added round debug info. https://cryptotrader.org/traders/q75spqHf6SWTDaid9Strategy is the same, but I found being able to step through the rounds and review what it was doing for each was enlightening. example: 013-08-19 09:30 Trader started. Balance: 1000.00 USD 2013-08-19 09:45 ROUND ==> 1 | HOLDING: usd=1000 btc=0 2013-08-19 09:45 cciValue =13.65474704470624 2013-08-19 09:45 highRocValue = -8.403361344537785 | lowRocValue = -9.888532448760534 2013-08-19 09:45 high =118 | low =117.25 | close =117.39002 2013-08-19 09:45 @ @ hold cci @ @ 2013-08-19 09:45 ************* 2013-08-19 09:45 ************* 2013-08-19 10:00 ROUND ==> 2 | HOLDING: usd=1000 btc=0 2013-08-19 10:00 cciValue =9.35064812180757 2013-08-19 10:00 highRocValue = -9.075714285714298 | lowRocValue = -7.711210914329358 2013-08-19 10:00 high =117.91999 | low =117.1 | close =117.25 2013-08-19 10:00 @ @ hold cci @ @ If you look at updated log, you will notice that the round counter has been reset. This happened because i restarted the trading engine after the software upgrade, so that all context data have been lost. Here's interesting problem - whether the engine should restore trader's context up on restart or not. That is not obvious to me.
|
|
|
|
luicon
|
|
August 22, 2013, 04:47:54 PM |
|
how to use the algoritms is real life trading ?
|
|
|
|
pulsecat (OP)
|
|
August 23, 2013, 02:40:28 PM |
|
how to use the algoritms is real life trading ?
The next update will add real trading feature. Meanwhile, you can run any algorithm using simulated trading account. For example, this code https://cryptotrader.org/backtests/PqS7WC4NXv6PiF3RD sends an alert to your email upon EMA crossovers when being run in the trading mode.
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
August 24, 2013, 01:17:15 AM Last edit: August 26, 2013, 06:32:26 PM by tabbek |
|
anyone else having an issue getting the graphs to display? It does the loading circle animation for me, but when it finishes, it's still just a blank graph section. the code and backtest simulation areas appear working, just the graph is broken.
edit : problem appears to have gone away. havent had any issues the past couple days... except with my fail contest entry :-P
|
|
|
|
|
sukiho
|
|
August 26, 2013, 08:36:45 PM |
|
looks like mtgox is winning so far
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
August 27, 2013, 11:54:38 PM |
|
That buy'n'hold link appears to actually be going to the EMA crossover example. Edit: LoL buy'n'hold looks like it'll be the winner yet again... unless we get a big slide back down. 2013-08-18 17:00 Simulation started. Balance: 1000.00 USD 2013-08-18 17:00 BUY 8.770 BTC at 113.39 2013-08-27 16:00 Simulation completed. Balance: 8.770 BTC (1141.63 USD)
|
|
|
|
pulsecat (OP)
|
|
September 02, 2013, 08:38:49 AM |
|
|
|
|
|
Diabolicus
Member
Offline
Activity: 90
Merit: 10
|
|
September 05, 2013, 10:36:42 AM Last edit: September 05, 2013, 12:40:42 PM by Diabolicus |
|
I just realized that some of the latest changes to the platform changed the way how the code is handled, see also here: https://bitcointalk.org/index.php?topic=222696.msg3087218#msg3087218If you open my entry you see for example the statistcs at the bottom displayed correctly, but if you click backtest again all is zeroed out. There has obviously been some change in how the storing of variables in the context object is handled, and this doesn't only affect the stats, even if for the July timeframe the result stays the same. It all worked fine when I posted my entry, so please change it back? Can't just change the rules in the middle of the contest ... ;-)
|
|
|
|
pulsecat (OP)
|
|
September 05, 2013, 02:50:48 PM Last edit: September 05, 2013, 06:46:34 PM by pulsecat |
|
OK, i see what happened. The problem is caused by careless merging of the code related to real trading mode, where orders are processed asynchronously. Should be fixed by now. Sorry for confusion.
|
|
|
|
pulsecat (OP)
|
|
September 07, 2013, 05:28:38 PM |
|
|
|
|
|
Mogumodz
|
|
September 07, 2013, 06:24:19 PM |
|
Kicking myself for not entering. Good luck everyone, count me in next round.
2013-08-19 00:00 Simulation started. Balance: 1000.00 USD 2013-09-03 13:00 SELL - Balance: 1225.59 USD
|
|
|
|
marvinrouge
|
|
September 08, 2013, 01:09:57 AM |
|
Kicking myself for not entering. Good luck everyone, count me in next round.
2013-08-19 00:00 Simulation started. Balance: 1000.00 USD 2013-09-03 13:00 SELL - Balance: 1225.59 USD
nice! hope indeed seeing your entry on the next round
|
|
|
|
marvinrouge
|
|
September 08, 2013, 01:23:01 AM Last edit: September 08, 2013, 05:06:14 PM by marvinrouge |
|
It seems marvinrouge's MACD algo started to overcome Buy'n'Hold strategy.
It's a narrow margin, it could not be enough if the trend reverse in the next hours, as it seems (my algo will buy if we'll exceed 131 in the next hours).
|
|
|
|
StarenseN
Legendary
Offline
Activity: 2478
Merit: 1362
|
|
September 08, 2013, 01:23:49 AM |
|
following
|
|
|
|
marvinrouge
|
|
September 09, 2013, 04:04:46 PM |
|
Buy and hold take the lead !
|
|
|
|
sukiho
|
|
September 10, 2013, 11:44:12 AM |
|
difficult to beat the market at the moment, but it is possible. the question is how will it work in future conditions? I think there is a basic flaw in standard backtesting which is easy to overcome (not so much a flaw but a disadvantage), but my programming isnt quite up to it, Im working on it
|
|
|
|
capoeira
Legendary
Offline
Activity: 2388
Merit: 1003
|
|
September 10, 2013, 07:19:20 PM |
|
where is next round topic?
|
|
|
|
ironstove
Member
Offline
Activity: 111
Merit: 10
Possibilities are limitless
|
|
September 11, 2013, 03:44:41 AM |
|
Is the default trading fee that you're backtesting with the .6% of mtgox? Or are you using 0% commission fee?
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
September 11, 2013, 08:58:56 PM |
|
the difficulty with beating the market currently is the fact that most of the substantial jumps and losses were caused by huge volume swings with no real precursors. There's no good way to predict someone coming along in a slow moving market and in one swoop making a huge purchase/sale. No trend gets established, just BAM, huge change. the sharper the angle of the upswing or downswing, the harder to 'trend'. the more curvy, the easier. the shorter your timescale, the more you can see the curve to a mass purchase / sell, but then you also subject yourself to other issues. tis a balance between looking at big picture, missing finer detail, vs looking in close, seeing details, but also being thrashed by false signals Is the default trading fee that you're backtesting with the .6% of mtgox? Or are you using 0% commission fee?
If you click on the "settings" tab next to the "log" tab, you will see that the bot defaults to 0.55 trading fee. One step up from the 'base' gox fee.
|
|
|
|
capoeira
Legendary
Offline
Activity: 2388
Merit: 1003
|
|
September 11, 2013, 09:55:22 PM |
|
the difficulty with beating the market currently is the fact that most of the substantial jumps and losses were caused by huge volume swings with no real precursors. There's no good way to predict someone coming along in a slow moving market and in one swoop making a huge purchase/sale. No trend gets established, just BAM, huge change.
the sharper the angle of the upswing or downswing, the harder to 'trend'. the more curvy, the easier.
the shorter your timescale, the more you can see the curve to a mass purchase / sell, but then you also subject yourself to other issues.
tis a balance between looking at big picture, missing finer detail, vs looking in close, seeing details, but also being thrashed by false signals
agreed, I used to trade on 15min, now i trade on the 4h......anything below is gambling now/atm
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
September 17, 2013, 11:36:44 PM |
|
agreed, I used to trade on 15min, now i trade on the 4h......anything below is gambling now/atm
I like to watch the 15 minute charts, but base my technicals on 1 or 4hr. 9/21 EMA -> 36/84 sort of thing, using a multiplier.
|
|
|
|
bitjoint
Sr. Member
Offline
Activity: 333
Merit: 250
Commander of the Hodl Legions
|
|
September 18, 2013, 02:48:45 AM |
|
Watching...
|
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
September 19, 2013, 12:01:46 PM |
|
Though Buy'n'Hold wins over all again, remember that Buy'n'Hold won't save you from market crash and there are lots of more profitable strategies than those that took part in this competition.
Bad news for Round #3 contestants, but I'm sure everyone can't wait for Round #4! Even lazy ones like myself who were late last time, hope to make it for the next round!
|
|
|
|
Stinky_Pete
|
|
September 19, 2013, 03:14:36 PM |
|
Would it be possible to run these strategies on historic data, such as starting Jan 1 2012? There's a background of steady climb so buy-and-hold will continue to be a good strategy, but also wild fluctuations in the summer months which might show if some of the other strategies out-perform buy-and-hold.
|
|
|
|
|
sukiho
|
|
September 19, 2013, 05:29:11 PM |
|
thats interesting to see, but how do you do it? I get a too much data message if I try myself
|
|
|
|
Stinky_Pete
|
|
September 19, 2013, 06:50:14 PM |
|
Thanks for doing that - now I just have to work out how marvinrouge does it!
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
September 21, 2013, 01:17:31 AM |
|
lol ouch
|
|
|
|
pulsecat (OP)
|
|
September 22, 2013, 10:40:55 AM Last edit: September 22, 2013, 01:08:10 PM by pulsecat |
|
thats interesting to see, but how do you do it? I get a too much data message if I try myself
I've used admin command line tool, that allows to run batch of simulations at once. Later the tool will be possibly integrated into the web, meanwhile another tool for automated trading is available at Github: https://github.com/pulsecat/cryptrade.
|
|
|
|
1Pakis
|
|
September 23, 2013, 07:21:44 AM |
|
Can you explain why if you apply the same algorithm for the same time period to bitstamp the results are completely opposite? https://cryptotrader.org/backtests/iALkrpS9HpDXMLTxvBalance: 0.421 BTC (53.53 USD)
|
Tips are welcome at this address 18DVZkpSwmejPjekX3QMKvRRtR8Bfx65LN.
|
|
|
cp1
|
|
September 23, 2013, 07:44:47 AM |
|
The bitstamp data seems to only be integers. Early on this makes a huge difference because there are a lot of bad trades where the algorithm sells lower than it buys. On the Mtgox data there are 2 decimals, so trades go like buy at 5.52 sell at 5.48, which is a loss, but not a big deal. But the bitstamp data is rounded to integers, so it goes buy at 6, sell at 5, which is a 16% loss. After 10 or so of these trades the Mtgox data is down by 10% but the bitstamp is down by 90% due to the rounding errors. Then the bitstamp is crippled and can't make very much money during the big april bubble.
|
|
|
|
Diabolicus
Member
Offline
Activity: 90
Merit: 10
|
|
September 23, 2013, 07:49:39 AM |
|
Interesting find, might have something to do with bitstamp data missing all decimal places? Prices seem to be integer only ..
|
|
|
|
pulsecat (OP)
|
|
September 23, 2013, 08:12:33 AM Last edit: September 23, 2013, 08:25:08 AM by pulsecat |
|
Right, wrong rounding made Bitstamp much less volatile. Should be fixed, once the import script completes its work.
|
|
|
|
|
Diabolicus
Member
Offline
Activity: 90
Merit: 10
|
|
September 26, 2013, 06:44:11 AM |
|
Will there be a 4th round?
|
|
|
|
pulsecat (OP)
|
|
September 27, 2013, 12:32:59 PM |
|
Will there be a 4th round?
Before we start, i want to make sure we have at least 5 participants. Who is in?
|
|
|
|
capoeira
Legendary
Offline
Activity: 2388
Merit: 1003
|
|
September 27, 2013, 12:34:58 PM |
|
Will there be a 4th round?
Before we start, i want to make sure we have at least 5 participants. Who is in? I want to join (1st time)
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
September 27, 2013, 12:49:05 PM |
|
I'm in also.
|
|
|
|
IncreaseMyT
|
|
September 27, 2013, 05:32:04 PM |
|
Someone teach me how to run that thing! Seriously
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
September 27, 2013, 09:42:15 PM |
|
im in again.
|
|
|
|
cp1
|
|
September 28, 2013, 12:22:16 AM |
|
I'll be in, I had some personal stuff come up last time.
|
|
|
|
Diabolicus
Member
Offline
Activity: 90
Merit: 10
|
|
September 28, 2013, 03:37:24 AM |
|
Me too, of course.
|
|
|
|
ZirconiumX
|
|
September 28, 2013, 08:50:11 AM |
|
I'll join in.
Are we allowed to modify one of the example scripts and enter it in?
Matthew:out
|
|
|
|
overunity
|
|
September 28, 2013, 09:40:07 AM |
|
I would love to be able to do what you guys can do .
To get my computer to automatically do things based on some information I have wrote .
Is there any were that these scripts ,I think you call them? can be learned by a very interested nooby .?
If there is does anyone know of a video format since I learn quicker watching instructional commentated video footage ?
|
|
|
|
nitrous
|
|
September 28, 2013, 10:18:32 AM |
|
I would love to be able to do what you guys can do .
To get my computer to automatically do things based on some information I have wrote .
Is there any were that these scripts ,I think you call them? can be learned by a very interested nooby .?
If there is does anyone know of a video format since I learn quicker watching instructional commentated video footage ?
In order to write these scripts, you need to know a programming language. For CryptoTrader.org, the language is Coffeescript. I don't know how quickly you can learn to do this. Essentially, programming is about translating logic into a form that a computer can understand. I've been programming in various languages for about 9 years, but you should definitely be able to get to a position where you can write simple programs or scripts, even to do trading, after a much shorter time. There may be videos, but I don't think that videos are very appropriate for learning to program, I suggest you search for tutorials in an interpreted language, like Javascript or Python. Coffeescript compiles into Javascript, so once you know Javascript it should be fairly straightforward to get to know Coffeescript (I'm not sure whether there are resources to help you learn Coffeescript without JS). It may sound daunting, but if you make a concerted effort to try you will get great results, and you will find that being able to program will be an invaluable skill in unexpected areas, and will allow you to fully take advantage of the power of modern technology. Another thing you might find scary is that there are so many different languages out there. What you will find though is that they share many features in common, and that with quick access to Google you should be able to get started with other languages in the same paradigm fairly quickly. If you just want me to tell you what to start with though, in case you can't decide what to do with all the variety out there, then I would recommend C if you're seriously interested in programming. Firstly, many other languages you will find use a very similar syntax, so it will be a very useful foundation. Secondly, it's a compiled language that is used for many purposes from your operating system to the programs you use, so it's versatile. Thirdly, it isn't object-orientated, so that's one less thing to learn as you start out. Fourthly, it will introduce you to the concept of compiling. Fifthly, it is quite strict so it will help you develop good practices before you start using more lenient languages. After a short while, if you can write some basic C programs and are aching to start on trading scripts you can use with CryptoTrader.org, then you will be prepared to start learning Javascript (which has similar syntax to C, but some things are different). Then you can learn Coffeescript and use cryptotrader.org. (Note that you will now have 1/3 of the tools to write basic websites as well! Just learn HTML and CSS (which aren't programming languages themselves, more just ways to format text), and you'll be ready to go!) Alternatively, if you want to actually use scripts to trade, you might want to instead learn Python, as many bitcoin trading libraries have been written in Python. Python, you will find, has a strikingly different appearance and syntax to C, but many of the underlying concepts are the same. With its interpreter as well, you may find it much quicker to work with than C, JS or Coffeescript. Btw, JS, Coffeescript, and Python all include Object-Orientated programming. This is an important concept to learn about at some stage, but you don't need to concern yourself with it when you start out, and it's not difficult to learn when you're more comfortable with the basics. Note: If you're on Windows, you might find it helpful to install Cygwin. If you have started learning these languages but are unsure of how to use them to start trading, feel free to PM me. I can be quite busy at times (and I'm going to get even more busy over the next few years !), but if I'm able to I'd be happy to help you learn this stuff
|
|
|
|
Diabolicus
Member
Offline
Activity: 90
Merit: 10
|
|
September 28, 2013, 01:08:22 PM |
|
You should start by taking apart step by step some of the scripts linked in this thread. Try to understand what each line does, alter some values and see how that affects the result. If you run into any trouble just ask here.
|
|
|
|
capoeira
Legendary
Offline
Activity: 2388
Merit: 1003
|
|
September 28, 2013, 01:52:07 PM |
|
I don't know nothing about programing, but playing around with the examples i could achieve what i wanted
|
|
|
|
overunity
|
|
September 28, 2013, 02:14:59 PM |
|
Awesome replies more than I ever hoped for .
I'am going to be googling the advice given here Thank you all
|
|
|
|
pulsecat (OP)
|
|
September 28, 2013, 06:47:04 PM Last edit: September 28, 2013, 07:24:12 PM by pulsecat |
|
I'll join in.
Are we allowed to modify one of the example scripts and enter it in?
Matthew:out
Yes, modified example scripts are allowed. I would love to be able to do what you guys can do .
To get my computer to automatically do things based on some information I have wrote .
Is there any were that these scripts ,I think you call them? can be learned by a very interested nooby .?
If there is does anyone know of a video format since I learn quicker watching instructional commentated video footage ?
In addition to the great explanation by nitrous, i would only add that even if you don't have programming skills, you still can run any algorithm of your choice to trade for bitcoins at Mtgox or Bitstamp. For example, let's say you want to trade EMA method at Bitstamp: 1) At Home page you will find links to some basic algorithms, one of those is EMA 10/21. 2) Go to the link to open the backtest tool. Configure backtest settings: change platform to Bitstamp, tune other settings if needed. Then click 'Backtest' button. Once the simulation is finished, you will get a log that shows how the strategy performed on historical market data at Bitstamp. Backtesting is an important procedure to test validity of any trading strategy. 3) Let's say you are satisfied with the backtest results and now you want configure bot to do trading for you at Bitstamp Install cryptrade tool from https://github.com/pulsecat/cryptrade. This can be tricky part if you don't have experience with nodejs/git or if you are on windows. 4) Run bot by using the following command ./cryptrade.sh -p bitstamp https://cryptotrader.org/backtests/PqS7WC4NXv6PiF3RD
Thats it. When bot is running, the algorithm, fed with marked data received from the server, generates trading signals that may result in profit if your strategy is good.
|
|
|
|
|
|
Adriano
Staff
Legendary
Offline
Activity: 1285
Merit: 1085
|
|
September 28, 2013, 07:40:33 PM |
|
- Until 10/07 00:00 GMT participants post links to their backtest results (link should look like: cryptotrader.org/backtests/<session-id>) to this thread so that the winner can be further determined. At 09/20 all participating scripts will be re-run with initial cash deposit set to 1000 USD and fee 0.5%. The author of the algorithm that performed best will be awarded with 2 BTC.
I think you have forgotten to update the end of the competition. Regards, Adriano
|
--
|
|
|
pulsecat (OP)
|
|
September 28, 2013, 07:44:03 PM |
|
- Until 10/07 00:00 GMT participants post links to their backtest results (link should look like: cryptotrader.org/backtests/<session-id>) to this thread so that the winner can be further determined. At 09/20 all participating scripts will be re-run with initial cash deposit set to 1000 USD and fee 0.5%. The author of the algorithm that performed best will be awarded with 2 BTC.
I think you have forgotten to update the end of the competition. Regards, Adriano Fixed, thanks.
|
|
|
|
IncreaseMyT
|
|
September 30, 2013, 06:33:43 PM |
|
Why don't you guys offer a service for newbs like us that interested. For instance, I would be willing to fund, someone writes the script and both make money? I would love to try and run one or even two, but I don't have enough time to learn how to code, nor do I want As everyone I have worked with in the digital industry usually hates writing code.
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
October 01, 2013, 03:58:57 PM |
|
Can you give us start/end dates for Round #2 & Round #3? It was in the opening post, but now is overwritten with Round #4 info. Need it to simulate how would new algorithms do in previous competition rounds.
|
|
|
|
|
pulsecat (OP)
|
|
October 07, 2013, 12:28:52 PM |
|
Looks like our beloved Bitcointalk is back? If no one objects, i changed Round #4 start date to 14 Oct due to obvious reasons. To avoid such confusions in the future, a discussion section has been added to the site.
|
|
|
|
pulsecat (OP)
|
|
October 07, 2013, 06:41:05 PM |
|
Why don't you guys offer a service for newbs like us that interested. For instance, I would be willing to fund, someone writes the script and both make money? I would love to try and run one or even two, but I don't have enough time to learn how to code, nor do I want As everyone I have worked with in the digital industry usually hates writing code. Seems like the feature you are asking for https://bitcointalk.org/index.php?topic=222696.msg3291331#msg3291331Good news! Cryptotrader now allows fully automated trading at MtGox & Bitstamp. Technically, everyone subscribed to the paid service gets his virtual machine in our VPS cloud, so that scalability is not an issue anymore.
|
|
|
|
dddbtc
|
|
October 07, 2013, 10:17:16 PM |
|
Just bought Pro access, I'm trying to trade on bitstamp, after I ran backtests and decided which algorithm I want to use, I clicked "trade" and attempted to login with my bitstamp user/pass as well as the bitstamp API key/secret. Neither of them work. I have the API key created with all permissions except for withdraws enabled under Security--->API access. I confirmed the email to activate the key I created. Why can't I trade on bitstamp I really hope this gets resolved quickly since I just purchased Pro access. my partner was hoping to mess with algorithms for the contest, but we're anti-gox because they won't accept my verification documents (however, bitstamp will haha)
|
|
|
|
cp1
|
|
October 08, 2013, 01:29:44 AM |
|
|
|
|
|
pulsecat (OP)
|
|
October 08, 2013, 02:00:46 PM |
|
Just bought Pro access, I'm trying to trade on bitstamp, after I ran backtests and decided which algorithm I want to use, I clicked "trade" and attempted to login with my bitstamp user/pass as well as the bitstamp API key/secret. Neither of them work. I have the API key created with all permissions except for withdraws enabled under Security--->API access. I confirmed the email to activate the key I created. Why can't I trade on bitstamp I really hope this gets resolved quickly since I just purchased Pro access. my partner was hoping to mess with algorithms for the contest, but we're anti-gox because they won't accept my verification documents (however, bitstamp will haha) Done. Now trading bots use clientId/key/secret to access Bitstamp API. Thanks.
|
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
October 10, 2013, 12:32:54 AM |
|
|
|
|
|
Grinny
|
|
October 10, 2013, 05:33:26 PM |
|
|
|
|
|
|
cp1
|
|
October 11, 2013, 04:29:11 PM |
|
- Given enough time, BnH can be beaten by EMA.
- I have definitely optimised my code for this dataset.
The contest is only run over a short term period. I thought it was still 2 weeks, but it's a month now. But still short term.
|
|
|
|
jago25_98
|
|
October 12, 2013, 12:00:40 AM |
|
Is the ZirconiumX one a typo? How come so much better? Is this something that's a good bet to try running live?
|
Bitcoiner since the early days. Crypto YouTube Channel: Trading Nomads | Analyst | News Reporter | Bitcoin Hodler | Support Freedom of Speech!
|
|
|
cp1
|
|
October 12, 2013, 04:03:33 AM |
|
Is the ZirconiumX one a typo? How come so much better? Is this something that's a good bet to try running live?
It's good to go back in time and use Jan. 1
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
October 12, 2013, 05:48:00 AM |
|
Well, one reason mine isnt doing well is that it was not intended to be running on a 4 hour time scale. I had it running on 15 minutes specifically.
|
|
|
|
sukiho
|
|
October 12, 2013, 05:54:57 AM |
|
Well, one reason mine isnt doing well is that it was not intended to be running on a 4 hour time scale. I had it running on 15 minutes specifically. the other two are set up for 1 hour time scales as well so those results arent really valid
|
|
|
|
ZirconiumX
|
|
October 12, 2013, 01:51:38 PM |
|
Is the ZirconiumX one a typo? How come so much better? Is this something that's a good bet to try running live?
Some minor tuning, that's all. Matthew:out
|
|
|
|
pulsecat (OP)
|
|
October 13, 2013, 08:39:27 AM |
|
Just to remind: 15 hours until Round #4 starts. Anyone can still join in!
|
|
|
|
Grinny
|
|
October 13, 2013, 09:20:08 AM |
|
All code is backtested on historical Bitstamp data at the trading frequency of your choice (5m,10m,15m,etc). Erhhh...I just clicked on my bookmark and the preset shows me a period of 4h - but this algorithm was design for 1h - how can i change and save this???
|
|
|
|
sukiho
|
|
October 13, 2013, 09:23:39 AM |
|
All code is backtested on historical Bitstamp data at the trading frequency of your choice (5m,10m,15m,etc). Erhhh...I just clicked on my bookmark and the preset shows me a period of 4h - but this algorithm was design for 1h - how can i change and save this??? shows 1 hour when I click on your entry, I think you are ok
|
|
|
|
sukiho
|
|
October 13, 2013, 09:33:34 AM |
|
|
|
|
|
Grinny
|
|
October 13, 2013, 09:41:17 AM |
|
All code is backtested on historical Bitstamp data at the trading frequency of your choice (5m,10m,15m,etc). Erhhh...I just clicked on my bookmark and the preset shows me a period of 4h - but this algorithm was design for 1h - how can i change and save this??? shows 1 hour when I click on your entry, I think you are ok thanks! shows 4h for me, but maybe it's because i have no ct account..(yet) or some cookie issue. nvm. good luck
|
|
|
|
|
|
|
medji
Newbie
Offline
Activity: 31
Merit: 0
|
|
October 14, 2013, 01:42:28 AM |
|
|
|
|
|
pulsecat (OP)
|
|
October 14, 2013, 05:29:32 AM |
|
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
October 14, 2013, 10:07:10 AM |
|
Just for fun, I ran all eight algorithms through periods for round #2 (07/29 - 08/5) & round #3 (08/19 09/19). Below every current contest entry is a line with result from these two previous rounds: ZirconiumX https://cryptotrader.org/backtests/NQckFniuZqZZJbrWX 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 10.727 BTC (1036.12 USD) 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 9.705 BTC (1241.02 USD)
cp1 https://cryptotrader.org/backtests/Z9tdtS6A3Er5aCCHY 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 2.668 BTC (256.95 USD)+750.00 (1006.94 USD) 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 2.504 BTC (320.52 USD)+750.00 (1070.52 USD)
tabbek https://cryptotrader.org/backtests/etFTmediwTGYDBfed 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 9.638 BTC (928.36 USD) 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 7.859 BTC (1005.86 USD)
Grinny https://cryptotrader.org/backtests/yjNjgwmXQn7Seju4q 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 10.676 BTC (1028.31 USD) 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 1185.76 USD
sukiho https://cryptotrader.org/backtests/oxMSRvfR6e7q8Sdcq 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 1037.85 USD 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 9.994 BTC (1279.16 USD)
Diabolicus https://cryptotrader.org/backtests/dR9XrbryXZZkRd7SC 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 959.36 USD 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 9.840 BTC (1259.44 USD)
itod https://cryptotrader.org/backtests/T7P9xTKj3JS95yKbY 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 10.727 BTC (1036.12 USD) 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 9.927 BTC (1269.39 USD)
mejdi https://cryptotrader.org/backtests/3mm5uMLajton9fCS9 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 1017.36 USD 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 9.762 BTC (1249.46 USD) sukiho's algorithm based on this is looking really, really good, came first in both backtests. Good luck to all contestants.
|
|
|
|
sukiho
|
|
October 14, 2013, 11:17:37 AM |
|
Just for fun, I ran all eight algorithms through periods for round #2 (07/29 - 08/5) & round #3 (08/19 09/19). Below every current contest entry is a line with result from these two previous rounds: ZirconiumX https://cryptotrader.org/backtests/NQckFniuZqZZJbrWX 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 10.727 BTC (1036.12 USD) 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 9.705 BTC (1241.02 USD)
cp1 https://cryptotrader.org/backtests/Z9tdtS6A3Er5aCCHY 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 2.668 BTC (256.95 USD)+750.00 (1006.94 USD) 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 2.504 BTC (320.52 USD)+750.00 (1070.52 USD)
tabbek https://cryptotrader.org/backtests/etFTmediwTGYDBfed 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 9.638 BTC (928.36 USD) 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 7.859 BTC (1005.86 USD)
Grinny https://cryptotrader.org/backtests/yjNjgwmXQn7Seju4q 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 10.676 BTC (1028.31 USD) 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 1185.76 USD
sukiho https://cryptotrader.org/backtests/oxMSRvfR6e7q8Sdcq 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 1037.85 USD 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 9.994 BTC (1279.16 USD)
Diabolicus https://cryptotrader.org/backtests/dR9XrbryXZZkRd7SC 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 959.36 USD 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 9.840 BTC (1259.44 USD)
itod https://cryptotrader.org/backtests/T7P9xTKj3JS95yKbY 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 10.727 BTC (1036.12 USD) 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 9.927 BTC (1269.39 USD)
mejdi https://cryptotrader.org/backtests/3mm5uMLajton9fCS9 2013/07/29 00:00 - 2013/08/05 00:00, Balance: 1017.36 USD 2013/08/19 00:00 - 2013/09/19 00:00, Balance: 9.762 BTC (1249.46 USD) sukiho's algorithm based on this is looking really, really good, came first in both backtests. Good luck to all contestants. yes, but not so good on longer term tests, also not sure if it beat buy and hold. my feeling is that the btc market is changing quickly as it grows so long term isnt so important, and also there is a lot of luck involved. Diabolicus looks good to me at the moment as well
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
October 14, 2013, 12:09:26 PM |
|
sukiho's algorithm based on this is looking really, really good, came first in both backtests. Good luck to all contestants.
yes, but not so good on longer term tests, also not sure if it beat buy and hold. my feeling is that the btc market is changing quickly as it grows so long term isnt so important, and also there is a lot of luck involved. Diabolicus looks good to me at the moment as well I think we need just one nice price drop during the contest period to beat buy & hold. It hasn't got to be like silkroad bust, but something similar would be nice Long term tests don't mean much IMHO, they have small resemblance to actual contest conditions. Diabolicus came first also in some of my favorite test-periods, so you two guys are the favorites.
|
|
|
|
dddbtc
|
|
October 15, 2013, 07:05:32 PM Last edit: October 15, 2013, 11:15:06 PM by dddbtc |
|
I'll offer a BTC bounty to anyone who can create a modified Ichimoku Strategy that includes the Heikin-Ashi candlestick modification. This would be a very beneficial indicator strategy to use on cryptotrader. I'd love to do this myself, unfortunately I don't know this scripting language well enough. More details on this strategy and its differences and benefits can be found on the following page. http://www.dailyfx.com/forex/education/trading_tips/daily_trading_lesson/2013/06/04/Ichimoku_Buy_AUDCAD.html
|
|
|
|
ZirconiumX
|
|
October 17, 2013, 04:32:01 PM |
|
Here you go: https://cryptotrader.org/backtests/f5quDEd3thWNZbshNThe Heikin-Ashi candles are the heikin_ashi_<blah> variables. I didn't know what you wanted to do with them, so I just print them at the moment. If you wanted to physically draw the candles, then speak to pulsecat. Matthew:out
|
|
|
|
pulsecat (OP)
|
|
October 18, 2013, 09:13:01 AM Last edit: October 18, 2013, 09:07:42 PM by pulsecat |
|
|
|
|
|
Grinny
|
|
October 18, 2013, 09:21:55 AM |
|
Yay! Thanks for the info I was really "pissed" yesterday as the threshold went down to -0.374% - just 0.026% ahead of my selling point... and then it went up again. So damn close
|
|
|
|
ZirconiumX
|
|
October 18, 2013, 05:21:53 PM |
|
Most algorithms there don't sell, so don't assume this is drawn. I doubt I'll win this... Matthew:out
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
October 18, 2013, 06:01:04 PM |
|
Most algorithms there don't sell, so don't assume this is drawn. I doubt I'll win this... Matthew:out We are all doing much worse than expected. Nobody got that selling point on Oct 16th right. Buy'n'Hold proves to be hard to beat. I think there's so much to learn about trading logic, maybe we just need some more competition rounds to have a chance to figure it out properly.
|
|
|
|
tabbek
Member
Offline
Activity: 116
Merit: 10
|
|
October 18, 2013, 07:40:11 PM |
|
|
|
|
|
pulsecat (OP)
|
|
October 18, 2013, 09:08:14 PM |
|
Ooops, sorry that was a copy paste issue.
|
|
|
|
cp1
|
|
October 19, 2013, 03:46:02 AM |
|
Now I just need a large dip and recovery, hopefully we're in a temporary bubble right now.
|
|
|
|
ZirconiumX
|
|
October 19, 2013, 07:35:19 AM |
|
Most algorithms there don't sell, so don't assume this is drawn. I doubt I'll win this... Matthew:out We are all doing much worse than expected. Nobody got that selling point on Oct 16th right. Buy'n'Hold proves to be hard to beat. I think there's so much to learn about trading logic, maybe we just need some more competition rounds to have a chance to figure it out properly. On inspection, I think pulsecat may have misunderstood my entry, which seems to be causing me a major headache. My data was from the beginning of 2013 until now, which required 4h candles, since cryptotrader couldn't handle the 1h candles. Because of how my EMA system worked, I only get "correct" entries from the October 17th 12:00 candle, by which point the bubble has disappeared. Still, a retest at 1h candles makes me slightly worse off. In the interests of progress, here is an experimental bot I worked on in the past, before giving up. Maybe the ideas behind it will inspire someone. Matthew:out
|
|
|
|
|
mobile
Sr. Member
Offline
Activity: 400
Merit: 250
the sun is shining, but the ice is still slippery
|
|
October 26, 2013, 02:00:55 PM |
|
watching as this looks interesting....
|
1 MoBi1eNbqh8QMuvtZjYzQGV8NEckJJYcT rep| GnuPG <3 CLAM <3
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
October 26, 2013, 02:17:33 PM |
|
I'm on the verge of buying again, need few stable 4-hour ticks till the end of the day to secure the lead and insure against large price surge. Here's the version with debug info: https://cryptotrader.org/backtests/cXagSEFirNxwzGFgtWish me luck
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
October 27, 2013, 11:33:45 PM |
|
Bummer. Unbelievable bad luck, 4 too small candle rises and than one too big: https://cryptotrader.org/backtests/XwFG3EnHdMyYDzANSIf any of these last 4/5 candles where a bit bigger it would buy for sure. Or a fifth bit smaller, a price was too high for $0.79 Well, maybe that was fair, there was nothing new in that algorithm, just the old algo from featured in Backtest optimized to try to beat B&H. But I have something completely new, can't wait if there's gonna be Round #5. What would you say if I tell you I've got the algo that wins both Round #2 & Round #3? Maybe you guys may still get this Buy&Hold bugger in this round? Good luck.
|
|
|
|
ZirconiumX
|
|
October 28, 2013, 08:43:31 AM |
|
Bummer. Unbelievable bad luck, 4 too small candle rises and than one too big: https://cryptotrader.org/backtests/XwFG3EnHdMyYDzANSIf any of these last 4/5 candles where a bit bigger it would buy for sure. Or a fifth bit smaller, a price was too high for $0.79 Well, maybe that was fair, there was nothing new in that algorithm, just the old algo from featured in Backtest optimized to try to beat B&H. But I have something completely new, can't wait if there's gonna be Round #5. What would you say if I tell you I've got the algo that wins both Round #2 & Round #3? Maybe you guys may still get this Buy&Hold bugger in this round? Good luck. Be wary of the Gambler's fallacy, itod. Past results give *zero* indication of what happens in the future... As for me, I've got 4 or 5 different algos to play with. Or maybe I've just got one. You'll have to wait and see. Matthew:out
|
|
|
|
monsterer
Legendary
Offline
Activity: 1008
Merit: 1000
|
|
October 29, 2013, 12:01:23 PM |
|
Been trying out your site, created a simple algo, but got:
"TypeError: Uncaught, unspecified "error" event."
What am I supposed to do with this message?
|
|
|
|
ZirconiumX
|
|
October 29, 2013, 12:37:05 PM |
|
Been trying out your site, created a simple algo, but got:
"TypeError: Uncaught, unspecified "error" event."
What am I supposed to do with this message?
Look over your code again. Usually that message means you are comparing a variable to something before you have assigned it a value - perhaps you typed blah instead of context.blah? Matthew:out
|
|
|
|
jambola2
Legendary
Offline
Activity: 1120
Merit: 1038
|
|
October 29, 2013, 05:51:26 PM |
|
Wait , do any of you know a place where I can invest my money in these overpowered algorithms , to earn monies ?
|
No longer active on bitcointalk, however, you can still reach me via PMs if needed.
|
|
|
|
Gimmoroy
Legendary
Offline
Activity: 1372
Merit: 1000
|
|
October 29, 2013, 09:01:56 PM |
|
2013-10-14 03:00 Simulation started. Balance: 1000.00 USD 2013-10-14 03:00 BUY 1.873 BTC at 132.82 2013-10-26 11:00 Simulation completed. Balance: 1.873 BTC (345.28 USD)+750.00 USD Buy and Hold: $1311.7015509712394 This Strategy: $1095.2759373588315
Can someone explain me this? What difference between "Buy and Hold" and single Buy and Hold made in this simulation?
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
October 29, 2013, 11:03:41 PM |
|
2013-10-14 03:00 Simulation started. Balance: 1000.00 USD 2013-10-14 03:00 BUY 1.873 BTC at 132.82 2013-10-26 11:00 Simulation completed. Balance: 1.873 BTC (345.28 USD)+750.00 USD Buy and Hold: $1311.7015509712394 This Strategy: $1095.2759373588315
Can someone explain me this? What difference between "Buy and Hold" and single Buy and Hold made in this simulation? Isn't it obvious? The real "Buy and Hold" means buying for full USD position, not only 1/4 of it.
|
|
|
|
cp1
|
|
October 29, 2013, 11:39:18 PM |
|
Can someone explain me this? What difference between "Buy and Hold" and single Buy and Hold made in this simulation?
I had a typo when calculating buy and hold -- I entered 0.95 to account for the fee instead of 0.995.
|
|
|
|
Gimmoroy
Legendary
Offline
Activity: 1372
Merit: 1000
|
|
October 30, 2013, 01:22:05 AM |
|
counting fingers (311.7 / 4) *.95 =74,02875 not 95.2
|
|
|
|
|
ZirconiumX
|
|
November 11, 2013, 04:27:56 PM |
|
It looks like my tuning effort went well! Matthew:out
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
November 11, 2013, 05:31:15 PM |
|
Well, Round 4 is over.
We couldn't ask for a better chance, had two big price dips and nobody properly handled it. May I suggest something for (eventual) future round: Maybe the wining algorithm should be the one which handles the whole year 2013 best? I know that comparing past instead of future performance opens the way for cheating, but any references to specific date or price should be forbidden. IMHO the algorithm which is most profitable in the very long term is the most valuable to everybody.
|
|
|
|
Grinny
|
|
November 11, 2013, 06:58:19 PM |
|
8h! So close. But I'm happy with the 3rd place.. Grats to the winner!
|
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
November 16, 2013, 01:09:57 PM |
|
Nice one-liner with avg = instrument.vwap(20) , can't get shorter than that, you win Don't have an idea what "Turtle Trading System" is, off to read some google links to find out. Regarding the contest idea for the whole year 2013 I've suggested, how many of you are for it that we organize our own winning pool for it? We can each, for instance, give 0.1 BTC in the pool and if there are 10 of us that may make things interesting. I understand pulsecat for not willing to risk 2BTC again after 4 unsuccessful rounds, but we can make our own fun with this, and might actually learn something valuable for 0.1BTC invested. Please tell me what you think even if you are not for it.
|
|
|
|
Grinny
|
|
November 16, 2013, 01:16:30 PM |
|
Regarding the contest idea for the whole year 2013 I've suggested, how many of you are for it that we organize our own winning pool for it? We can each, for instance, give 0.1 BTC in the pool and if there are 10 of us that may make things interesting. I understand pulsecat for not willing to risk 2BTC again after 4 unsuccessful rounds, but we can make our own fun with this, and might actually learn something valuable for 0.1BTC invested. Please tell me what you think even if you are not for it.
great idea! count me in for a contest beginning at 2014-01-01 and ending at 2014-12-31 (or for any other period, but more interested in a long time contest)
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
November 16, 2013, 01:37:45 PM |
|
Regarding the contest idea for the whole year 2013 I've suggested, how many of you are for it that we organize our own winning pool for it? We can each, for instance, give 0.1 BTC in the pool and if there are 10 of us that may make things interesting. I understand pulsecat for not willing to risk 2BTC again after 4 unsuccessful rounds, but we can make our own fun with this, and might actually learn something valuable for 0.1BTC invested. Please tell me what you think even if you are not for it.
great idea! count me in for a contest beginning at 201 4-01-01 and ending at 201 4-12-31 (or for any other period, but more interested in a long time contest) I suppose you've meant 2013, not 2014? It would be a bit too long to wait a whole year for the results
|
|
|
|
Grinny
|
|
November 16, 2013, 01:53:53 PM |
|
Regarding the contest idea for the whole year 2013 I've suggested, how many of you are for it that we organize our own winning pool for it? We can each, for instance, give 0.1 BTC in the pool and if there are 10 of us that may make things interesting. I understand pulsecat for not willing to risk 2BTC again after 4 unsuccessful rounds, but we can make our own fun with this, and might actually learn something valuable for 0.1BTC invested. Please tell me what you think even if you are not for it.
great idea! count me in for a contest beginning at 201 4-01-01 and ending at 201 4-12-31 (or for any other period, but more interested in a long time contest) I suppose you've meant 2013, not 2014? It would be a bit too long to wait a whole year for the results nope, I meant 2014 of course. a one year lasting contest - the iron man of all automated trading contest
|
|
|
|
pulsecat (OP)
|
|
November 16, 2013, 02:12:20 PM |
|
Nice one-liner with avg = instrument.vwap(20) , can't get shorter than that, you win Don't have an idea what "Turtle Trading System" is, off to read some google links to find out. Regarding the contest idea for the whole year 2013 I've suggested, how many of you are for it that we organize our own winning pool for it? We can each, for instance, give 0.1 BTC in the pool and if there are 10 of us that may make things interesting. I understand pulsecat for not willing to risk 2BTC again after 4 unsuccessful rounds, but we can make our own fun with this, and might actually learn something valuable for 0.1BTC invested. Please tell me what you think even if you are not for it. Right, the goal of the contest was to find out what strategies could outperform B&H in the short term, because obviously in the long term it is less profitable than even basic strategies. But i really like the idea of donation-based contest and can help with organization if there are enough people willing to join in.
|
|
|
|
Mavi
Member
Offline
Activity: 80
Merit: 10
|
|
November 16, 2013, 02:16:07 PM |
|
Regarding the contest idea for the whole year 2013 I've suggested, how many of you are for it that we organize our own winning pool for it? We can each, for instance, give 0.1 BTC in the pool and if there are 10 of us that may make things interesting. I understand pulsecat for not willing to risk 2BTC again after 4 unsuccessful rounds, but we can make our own fun with this, and might actually learn something valuable for 0.1BTC invested. Please tell me what you think even if you are not for it.
I'm interested, i can chip in but i don't know how to code. I like to watching you guys
|
|
|
|
ZirconiumX
|
|
November 16, 2013, 04:31:30 PM |
|
Regarding the contest idea for the whole year 2013 I've suggested, how many of you are for it that we organize our own winning pool for it? We can each, for instance, give 0.1 BTC in the pool and if there are 10 of us that may make things interesting. I understand pulsecat for not willing to risk 2BTC again after 4 unsuccessful rounds, but we can make our own fun with this, and might actually learn something valuable for 0.1BTC invested. Please tell me what you think even if you are not for it.
I'm interested, i can chip in but i don't know how to code. I like to watching you guys Conversely, I can code, but am really poor (check my tipjar address in my sig if you don't believe me). Matthew:out
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
November 16, 2013, 04:36:47 PM |
|
Regarding the contest idea for the whole year 2013 I've suggested, how many of you are for it that we organize our own winning pool for it? We can each, for instance, give 0.1 BTC in the pool and if there are 10 of us that may make things interesting. I understand pulsecat for not willing to risk 2BTC again after 4 unsuccessful rounds, but we can make our own fun with this, and might actually learn something valuable for 0.1BTC invested. Please tell me what you think even if you are not for it.
great idea! count me in for a contest beginning at 201 4-01-01 and ending at 201 4-12-31 (or for any other period, but more interested in a long time contest) I suppose you've meant 2013, not 2014? It would be a bit too long to wait a whole year for the results nope, I meant 2014 of course. a one year lasting contest - the iron man of all automated trading contest Sorry, to wait a whole year for a result is a bit too long for me. It's better idea to try one more month-long round IMHO. Let's see what others think. Maybe pulsecat offers to double the amount we collect
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
November 16, 2013, 04:38:39 PM |
|
I'm interested, i can chip in but i don't know how to code. I like to watching you guys Conversely, I can code, but am really poor (check my tipjar address in my sig if you don't believe me). Matthew:out Why don't you two join forces, ZirconiumX code & Mavi invest 0.1BTC and you split the winnings?
|
|
|
|
ZirconiumX
|
|
November 16, 2013, 04:48:42 PM |
|
I'm interested, i can chip in but i don't know how to code. I like to watching you guys Conversely, I can code, but am really poor (check my tipjar address in my sig if you don't believe me). Matthew:out Why don't you two join forces, ZirconiumX code & Mavi invest 0.1BTC and you split the winnings? I'm up for it, not sure what Mavi thinks, though. Could we have some sort of special stipulation to try and bring some novelty into the competition, for instance "Be as profitable as you can using the RSI indicator" or whatever? Matthew:out
|
|
|
|
Mavi
Member
Offline
Activity: 80
Merit: 10
|
|
November 16, 2013, 05:12:55 PM |
|
Fine by me, but i don't have any btc right now, i'm waiting for price goes down.
|
|
|
|
pulsecat (OP)
|
|
November 16, 2013, 07:36:10 PM Last edit: November 17, 2013, 10:29:08 AM by pulsecat |
|
OK, who is willing to participate in the next round, beside itod and ZirconiumX? Unlike the previous rounds, strategies will be backtested on historical market data with start date set to 2013/01/01 plus one month of future performance. Entry fee is 0.1 BTC which goes to the winning pot. Also i'll double the prize - for example if there will be 5 participants, i add 0.5 BTC to the pot. Voluntary donations are welcome, though they won't be doubled. If you wish to join in send entry fee (0.1 BTC) and your nickname to this address admin@cryptotrader.org using Coinapult service. contact me in PM first. If you would like to contribute to the prize pool, please send your bitcoin donations to this address 1BueD4kDjkPHVWH7btwt6GNz7Ki3wyeA7q
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
November 16, 2013, 09:19:34 PM |
|
Entry fee is 0.1 BTC which goes to the winning pot. Also i'll double the prize - for example if there will be 5 participants, i add 0.5 BTC to the pot. Voluntary donations are welcome, though they won't be doubled.
Great pulsecat! Withdrawing coins to send it. Although this Coinapult service looks like a pain in the ass. What will launch when I do the transaction? It's a bit scary, aren't there more elegant ways to generate unique payment addresses?
|
|
|
|
pulsecat (OP)
|
|
November 23, 2013, 01:43:41 PM Last edit: November 25, 2013, 06:05:13 PM by pulsecat |
|
OK, because there is very few of those who would risk their bitcoins, the Round #5 will be sponsored by myself! As compared to previous rounds: - Strategies will be backtested over the period of 1 year instead of one month. - There is no "outperform Buy and Hold" rule, so someone should guaranteed win. - All strategies must be pure algorithms, read more in the rules link.Please review updated rules for Round 5 and suggest changes, as these rules are subject to changes until Monday. Anyone who wishes to donate to the prize fund please send Bitcoins to 1BueD4kDjkPHVWH7btwt6GNz7Ki3wyeA7q. The bigger will be the prize, the more participants join the contest. As the result everyone will benefit from the variety of good strategies!
|
|
|
|
Gimmoroy
Legendary
Offline
Activity: 1372
Merit: 1000
|
|
November 23, 2013, 05:59:59 PM Last edit: November 24, 2013, 05:48:40 PM by Gimmoroy |
|
- There is no "outperform Buy and Hold" rule, so someone should guaranteed win.
Sure? My supermegaalgorithm will make single buy at the start point. May I send the address to get my prize ? Is there someone, who will publish profitable algorithm? I think the challenge should keep an algorithm in secret. And I am sure, it will be a good occasion to sell the good things, and make service idea more reliable. Also the announcement should be crossposted at least in coders section.
|
|
|
|
wamatt
Member
Offline
Activity: 85
Merit: 10
|
|
November 24, 2013, 05:06:41 AM |
|
It's probably too complicated, but have you considered plans to support Bitfinex via their API?
Complexity would due to the leverage and shorting. Additionally there are loan fee's on margin trades, all of which, would affect the P+L calculation
|
|
|
|
Grinny
|
|
November 24, 2013, 10:45:25 AM |
|
|
|
|
|
ZirconiumX
|
|
November 24, 2013, 04:08:23 PM |
|
Which does not beat Buy'N'Hold. Bear in mind that modified example scripts are allowed. As for me, though I'm not making an official entry yet, I am experimenting with some drunk science. I have no idea how changing the VWAP from 20 periods to 1 period increases the income tenfold. Matthew:out
|
|
|
|
pulsecat (OP)
|
|
November 24, 2013, 06:51:10 PM |
|
Which does not beat Buy'N'Hold. Bear in mind that modified example scripts are allowed. As for me, though I'm not making an official entry yet, I am experimenting with some drunk science. I have no idea how changing the VWAP from 20 periods to 1 period increases the income tenfold. Matthew:out Great! Finally we have first volume based strategy. The only dissappointing detail is that it went short at $494.26 and didn't buy after.
|
|
|
|
ZirconiumX
|
|
November 25, 2013, 07:17:48 AM |
|
Which does not beat Buy'N'Hold. Bear in mind that modified example scripts are allowed. As for me, though I'm not making an official entry yet, I am experimenting with some drunk science. I have no idea how changing the VWAP from 20 periods to 1 period increases the income tenfold. Matthew:out Great! Finally we have first volume based strategy. The only dissappointing detail is that it went short at $494.26 and didn't buy after. Does the volume really have that much of an effect if I'm only calculating the VWAP for one period? As for the short selling, there are multiple things I could try to solve the problem, based around ignoring the VWAP, such as checking if we have hit a 20-period high, or using the ADX. I could also try something purely volume based. Matthew:out
|
|
|
|
lonsharim
|
|
November 25, 2013, 10:14:21 AM |
|
Is there any place for us to see the winning strategies of the previous (4?) rounds?
|
|
|
|
ZirconiumX
|
|
November 25, 2013, 04:24:57 PM |
|
Which does not beat Buy'N'Hold. Bear in mind that modified example scripts are allowed. As for me, though I'm not making an official entry yet, I am experimenting with some drunk science. I have no idea how changing the VWAP from 20 periods to 1 period increases the income tenfold. Matthew:out Great! Finally we have first volume based strategy. The only dissappointing detail is that it went short at $494.26 and didn't buy after. Does the volume really have that much of an effect if I'm only calculating the VWAP for one period? As for the short selling, there are multiple things I could try to solve the problem, based around ignoring the VWAP, such as checking if we have hit a 20-period high, or using the ADX. I could also try something purely volume based. Matthew:out As it turns out, one of these strategies is very successful indeed. I'll throw you a bone: 2013-11-25 00:00 Simulation completed. Balance: 103.9711 BTC (83182.08 USD)Matthew:out
|
|
|
|
ZirconiumX
|
|
November 25, 2013, 05:22:55 PM |
|
After polishing off the code somewhat, I present to you, the B'n'H beater! This is my official entry. Matthew:out
|
|
|
|
ZirconiumX
|
|
November 26, 2013, 05:23:48 PM |
|
After polishing off the code somewhat, I present to you, the B'n'H beater! This is my official entry. Matthew:out Note that this does not seem to be as effective on MtGox or BTC-E. It's almost a fluke. Matthew:out
|
|
|
|
Casimir1904
Full Member
Offline
Activity: 209
Merit: 100
Radix-The Decentralized Finance Protocol
|
|
November 27, 2013, 12:16:42 PM |
|
After polishing off the code somewhat, I present to you, the B'n'H beater! This is my official entry. Matthew:out looks nice, only to bad it didn't sell on the $900 up from 19.11 for mtgox i did test with butter bot ema 10/21 2h and buy threashold on 0% and sell threashold on 5% made 9 trades in 2y and beated B'n'H 0.6% fee. Profit:36929.98%Max. Profit:37153.50%Min. Profit:-0.60%Trades:9 Time Type Price Value Profit 21.11.2013 11:00:00 buy 670.000 $37029.98 36929.98% 55.26862866 BTC 20.11.2013 13:00:00 sell 500.000 $37253.50 37153.50% 04.05.2013 22:00:00 buy 110.120 $8254.24 8154.24% 02.05.2013 08:00:00 sell 116.380 $8304.06 8204.06% 17.04.2013 18:00:00 buy 90.162 $6472.16 6372.16% 11.04.2013 04:00:00 sell 169.064 $6511.22 6411.22% 21.08.2012 18:00:00 buy 10.082 $390.64 290.64% 19.08.2012 12:00:00 sell 10.099 $392.99 292.99% 28.11.2011 17:00:00 buy 2.539 $99.40 -0.60% 39.14927137 BTC +16.11935729 BTC +36929.98 USD Starting fund was $100. Can't make so long backtesting with cryptotrader but should be nearly same result.
|
|
|
|
exuals
Newbie
Offline
Activity: 3
Merit: 0
|
|
November 27, 2013, 06:53:47 PM |
|
Not sure if I'm eligible to contest, doing these more out of interest. Based off a similar data-set that ZirconiumX used to show off his BnH beater. I can't reach quite the same results, I use a cheesy minimum price gain percentage mixed with a volume indicator for buying during bubbles. First actual attempt I suppose: 2013-11-25 16:00 Simulation completed. Balance: 75.9959 BTC (63075.88 USD) (Starting from $1,000, MtGox, BTC/USD @ 4hr) https://cryptotrader.org/backtests/bQzucyJursAiv2cgFI'm honestly not sure how "vwap" works but changing the param from a 5 to a twenty increased output dramatically. Also I'm just a game programmer with little to no market knowledge but this is hell of a lot of fun to dick around with.
|
|
|
|
exuals
Newbie
Offline
Activity: 3
Merit: 0
|
|
November 27, 2013, 07:09:48 PM |
|
After polishing off the code somewhat, I present to you, the B'n'H beater! This is my official entry. Matthew:out I learned a lot from your code and attempted to implement my simple system into yours, turned out alright: https://cryptotrader.org/backtests/kqPNdaRruEz2edpiK2013-11-24 16:00 Simulation completed. Balance: 181.2825 BTC (145035.09 USD) I haven't added selling based of previous highs beating new high * percentage but I think it would sell at the peak with some adjustment for potentially more profit.
|
|
|
|
ZirconiumX
|
|
November 27, 2013, 09:56:26 PM |
|
After polishing off the code somewhat, I present to you, the B'n'H beater! This is my official entry. Matthew:out I learned a lot from your code and attempted to implement my simple system into yours, turned out alright: https://cryptotrader.org/backtests/kqPNdaRruEz2edpiK2013-11-24 16:00 Simulation completed. Balance: 181.2825 BTC (145035.09 USD) I haven't added selling based of previous highs beating new high * percentage but I think it would sell at the peak with some adjustment for potentially more profit. Well, my first place spot was fun while it lasted, but I don't stand a hope in hell against that. Matthew:out
|
|
|
|
sukiho
|
|
November 28, 2013, 11:08:27 AM |
|
just a mention to clarify the rules, November has only 30 days
|
|
|
|
pulsecat (OP)
|
|
November 28, 2013, 03:27:07 PM |
|
After polishing off the code somewhat, I present to you, the B'n'H beater! This is my official entry. Matthew:out Please note this round is on Mtgox, not Bitstamp. just a mention to clarify the rules, November has only 30 days
Thanks for pointing this out. Fixed.
|
|
|
|
ZirconiumX
|
|
November 28, 2013, 04:33:03 PM |
|
After polishing off the code somewhat, I present to you, the B'n'H beater! This is my official entry. Matthew:out Please note this round is on Mtgox, not Bitstamp. just a mention to clarify the rules, November has only 30 days
Thanks for pointing this out. Fixed. In that case, I haven't a hope in hell. Matthew:out
|
|
|
|
pulsecat (OP)
|
|
November 29, 2013, 07:08:28 AM |
|
After polishing off the code somewhat, I present to you, the B'n'H beater! This is my official entry. Matthew:out Please note this round is on Mtgox, not Bitstamp. just a mention to clarify the rules, November has only 30 days
Thanks for pointing this out. Fixed. In that case, I haven't a hope in hell. Matthew:out You still can tune your algo, as the entries are not tracked until 1 Dec.
|
|
|
|
medji
Newbie
Offline
Activity: 31
Merit: 0
|
|
November 29, 2013, 09:41:03 AM |
|
|
|
|
|
bernard75
Legendary
Offline
Activity: 1316
Merit: 1003
|
|
November 29, 2013, 09:54:17 AM |
|
Why in the world would you tell here if you found the sorcerers stone?
|
|
|
|
Grinny
|
|
November 29, 2013, 10:30:58 AM |
|
Why in the world would you tell here if you found the sorcerers stone?
Why not? Algorithms are no indivisible goods, right?
|
|
|
|
bernard75
Legendary
Offline
Activity: 1316
Merit: 1003
|
|
November 29, 2013, 10:40:24 AM |
|
Why in the world would you tell here if you found the sorcerers stone?
Why not? Algorithms are no indivisible goods, right? If everybody uses them they loose value.
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
November 29, 2013, 11:59:21 AM |
|
Mind blowing. 15.65 BTC -> 115.5319 BTC. If someone told me this is possible I wouldn't have believed. This one will be tough to beat.
|
|
|
|
medji
Newbie
Offline
Activity: 31
Merit: 0
|
|
November 29, 2013, 09:11:03 PM |
|
It bought 63 @ $15.5 in the beginning mate, so just below twice the btc.
|
|
|
|
pulsecat (OP)
|
|
November 30, 2013, 05:58:10 PM |
|
Just to remind: the next round starts in only 6 hours!
|
|
|
|
|
sukiho
|
|
November 30, 2013, 07:40:16 PM |
|
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
November 30, 2013, 08:50:01 PM |
|
Cryptotrader down just when polishing algo for submission, can't work now and need every minute. It would be fair to get few hours extension because of this. Thanks. Edit: 502 Bad Gateway
nginx/1.5.7
|
|
|
|
pulsecat (OP)
|
|
November 30, 2013, 09:20:11 PM |
|
Cryptotrader down just when polishing algo for submission, can't work now and need every minute. It would be fair to get few hours extension because of this. Thanks. Edit: 502 Bad Gateway
nginx/1.5.7 Thanks for letting know. Because of these circumstances entries will be accepted until 2:00 GMT.
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
November 30, 2013, 09:25:17 PM |
|
Cryptotrader down just when polishing algo for submission, can't work now and need every minute. It would be fair to get few hours extension because of this. Thanks. Edit: 502 Bad Gateway
nginx/1.5.7 Thanks for letting know. Because of these circumstances entries will be accepted until 2:00 GMT. Thanks!
|
|
|
|
silversurfer1958
|
|
December 01, 2013, 12:41:12 AM |
|
Stagger buy and sell orders at +/- trade fee +1% above current price, wait 30 mins. Cancel all trades. Rinse and repeat.
You're buying low, selling high.
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
December 01, 2013, 01:59:16 AM |
|
|
|
|
|
|
|
|
pulsecat (OP)
|
|
December 01, 2013, 07:04:11 AM |
|
|
|
|
|
Mavi
Member
Offline
Activity: 80
Merit: 10
|
|
December 01, 2013, 09:56:20 AM |
|
Kiri11's algorithm starts with $5k.
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
December 01, 2013, 08:29:54 PM |
|
Kiri11's algorithm starts with $5k.
It doesn't matter because all should start with $5k, pulsecat overrides all settings someone inputs with defaults while testing. We will probably soon see first run and results with official settings for all algorithms.
|
|
|
|
Coin.Karma
|
|
December 03, 2013, 02:03:50 PM |
|
Hey guys, Not part of the contest.....and not possible in real because so much bitcoins don't exist...but some crazy profit on a 6 day trading for an EMA based strategy.....check it out, also, it just takes the ideal situations....wont be possible in any case...even a little bit of profit of this amount is not possible because of the amount of noise present and trouble created by other bots....have fun. Wish I could code... too much of a hassle https://cryptotrader.org/backtests/9JAtfub6bhXCdoGsZStrategy started with 1 BTC and in 6 days ended with 1472660896026965 BTC (hehe) @pulsecat, you should also create a fun contest for a strategy which makes more than the amount of bitcoins that will ever be in existence.just for kicks.....
|
|
|
|
|
pulsecat (OP)
|
|
December 03, 2013, 02:38:44 PM Last edit: December 03, 2013, 08:27:12 PM by pulsecat |
|
Hey guys, Not part of the contest.....and not possible in real because so much bitcoins don't exist...but some crazy profit on a 6 day trading for an EMA based strategy.....check it out, also, it just takes the ideal situations....wont be possible in any case...even a little bit of profit of this amount is not possible because of the amount of noise present and trouble created by other bots....have fun. Wish I could code... too much of a hassle https://cryptotrader.org/backtests/9JAtfub6bhXCdoGsZStrategy started with 1 BTC and in 6 days ended with 1472660896026965 BTC (hehe) @pulsecat, you should also create a fun contest for a strategy which makes more than the amount of bitcoins that will ever be in existence.just for kicks..... Obviously the main reason why this strategy does not work is a lack of volume on CEX.IO these days.
|
|
|
|
medji
Newbie
Offline
Activity: 31
Merit: 0
|
|
December 04, 2013, 01:56:58 AM |
|
Nice. If someone gets to profit or learned something interesting from my entries be sure to send me a tip
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
December 04, 2013, 12:06:39 PM |
|
What is a real surprise for me revealed from round#5 is that there are 3 totally different strategies (Ichimoku, MFI/RSI, EMA) which successfully beat Buy&Hold by 50%-100% in the long range. Probably there are even more of them. I honestly didn't expect these strategies to be so successful.
|
|
|
|
pulsecat (OP)
|
|
December 07, 2013, 07:24:37 PM |
|
|
|
|
|
Mavi
Member
Offline
Activity: 80
Merit: 10
|
|
December 07, 2013, 07:28:57 PM |
|
medji lost $150,000 in a week. shame, i was kinda like it
|
|
|
|
pulsecat (OP)
|
|
December 07, 2013, 07:41:02 PM |
|
medji lost $150,000 in a week. shame, i was kinda like it It is just 16% of the portfolio, mr. Buy&Hold lost 32%
|
|
|
|
Mavi
Member
Offline
Activity: 80
Merit: 10
|
|
December 07, 2013, 08:26:10 PM |
|
medji lost $150,000 in a week. shame, i was kinda like it It is just 16% of the portfolio, mr. Buy&Hold lost 32% B&H was already lost the game These algorithms are good in long term but useless in short term. I need passive income, monthly
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
December 07, 2013, 09:12:49 PM |
|
medji lost $150,000 in a week. shame, i was kinda like it It is just 16% of the portfolio, mr. Buy&Hold lost 32% B&H was already lost the game These algorithms are good in long term but useless in short term. I need passive income, monthly Four contest rounds with 2BTC bounty each failed to produce good short-term (2-4 weeks) algorithm, which can beat B&H. Maybe there's no such an algorithm, or maybe the bounty was not high enough for people to publish theirs. If you really want such an algorithm maybe you should try help pulsecat organize higher bounty than 2BTC. I think there will be much more and much better algorithms published in that case, if for no other reason then it would enable much more time devoted to developing them. It's not exactly easy to do everything right, and it takes time.
|
|
|
|
Mavi
Member
Offline
Activity: 80
Merit: 10
|
|
December 07, 2013, 10:43:57 PM |
|
medji lost $150,000 in a week. shame, i was kinda like it It is just 16% of the portfolio, mr. Buy&Hold lost 32% B&H was already lost the game These algorithms are good in long term but useless in short term. I need passive income, monthly Four contest rounds with 2BTC bounty each failed to produce good short-term (2-4 weeks) algorithm, which can beat B&H. Maybe there's no such an algorithm, or maybe the bounty was not high enough for people to publish theirs. If you really want such an algorithm maybe you should try help pulsecat organize higher bounty than 2BTC. I think there will be much more and much better algorithms published in that case, if for no other reason then it would enable much more time devoted to developing them. It's not exactly easy to do everything right, and it takes time. I don't think this is about bounty, if developer has the "super-algorithm", he won't need any bounty
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
December 07, 2013, 11:29:33 PM |
|
medji lost $150,000 in a week. shame, i was kinda like it It is just 16% of the portfolio, mr. Buy&Hold lost 32% B&H was already lost the game These algorithms are good in long term but useless in short term. I need passive income, monthly Four contest rounds with 2BTC bounty each failed to produce good short-term (2-4 weeks) algorithm, which can beat B&H. Maybe there's no such an algorithm, or maybe the bounty was not high enough for people to publish theirs. If you really want such an algorithm maybe you should try help pulsecat organize higher bounty than 2BTC. I think there will be much more and much better algorithms published in that case, if for no other reason then it would enable much more time devoted to developing them. It's not exactly easy to do everything right, and it takes time. I don't think this is about bounty, if developer has the "super-algorithm", he won't need any bounty You got me wrong. Probably nobody has a "super-algorithm", but given enough time many developers may produce some very good ones. Developers don't have few weeks or a month free to experiment and try new things, but given attractive bounty some may decide to devote the time to it. And public contest is an excellent way to test the results.
|
|
|
|
|
kugutsumen
Newbie
Offline
Activity: 15
Merit: 0
|
|
December 10, 2013, 04:44:58 PM Last edit: December 10, 2013, 04:56:21 PM by kugutsumen |
|
Just got my posting rights and this was meant to be a reply to this thread. Kugutsumen IIaBuy the rumor, sell the news... Survive and profit on rocky roads :welp: This Ichimoku algo seems to perform really well now that bitcoin has hit the mainstream: 2013-12-01 12:00 Simulation started. Balance: 5000.00 USD 2013-12-09 22:00 Simulation completed. Balance: 7.7367 BTC (7100.01 USD)
2013-11-01 12:00 Simulation started. Balance: 5000.00 USD 2013-12-09 22:00 Simulation completed. Balance: 41.4457 BTC (38034.69 USD)
2013-10-01 12:00 Simulation started. Balance: 5000.00 USD 2013-12-09 22:00 Simulation completed. Balance: 60.1737 BTC (55221.38 USD)
This is probably too late for round 5 of Pulsecat's contest: "Kugutsumen I" beats Medji R5 entry (sent a link to Pulsecat who can confirm) Kugu 939K — Medji 907K | Before the dip [2013-12-03 16:00] 2013-01-01 00:00 Simulation started. Balance: 5000.00 USD Medji 2013-12-03 16:00 Simulation completed. Balance: 907.2280 BTC (1007023.09 USD) Kugu 2013-12-03 16:00 Simulation completed. Balance: 939.4086 BTC (1042743.57 USD)
Kugu 978K — Medji 854K | After the dip [2013-12-07 18:00] Medji 2013-12-07 18:00 Simulation completed. Balance: 854899.01 USD Kugu 2013-12-07 18:00 Simulation completed. Balance: 978145.93 USD
Kugu 1.3M — Medji 1.14M [2013-12-09 22:00] Medji 2013-12-09 22:00 Simulation completed. Balance: 1245.3286 BTC (1142838.07 USD) Kugu 2013-12-09 22:00 Simulation completed. Balance: 1424.8620 BTC (1307595.88 USD)
Please keep in mind that this is a quick hack, I am completely new to this... I read up on Ichimoku and I fixed a bug in Medji's code and used different parameters.
|
|
|
|
kugutsumen
Newbie
Offline
Activity: 15
Merit: 0
|
|
December 10, 2013, 04:57:01 PM |
|
medji lost $150,000 in a week. shame, i was kinda like it It is just 16% of the portfolio, mr. Buy&Hold lost 32% B&H was already lost the game These algorithms are good in long term but useless in short term. I need passive income, monthly Four contest rounds with 2BTC bounty each failed to produce good short-term (2-4 weeks) algorithm, which can beat B&H. Maybe there's no such an algorithm, or maybe the bounty was not high enough for people to publish theirs. If you really want such an algorithm maybe you should try help pulsecat organize higher bounty than 2BTC. I think there will be much more and much better algorithms published in that case, if for no other reason then it would enable much more time devoted to developing them. It's not exactly easy to do everything right, and it takes time. I don't think this is about bounty, if developer has the "super-algorithm", he won't need any bounty You got me wrong. Probably nobody has a "super-algorithm", but given enough time many developers may produce some very good ones. Developers don't have few weeks or a month free to experiment and try new things, but given attractive bounty some may decide to devote the time to it. And public contest is an excellent way to test the results. medji lost $150,000 in a week. shame, i was kinda like it It is just 16% of the portfolio, mr. Buy&Hold lost 32% B&H was already lost the game These algorithms are good in long term but useless in short term. I need passive income, monthly Four contest rounds with 2BTC bounty each failed to produce good short-term (2-4 weeks) algorithm, which can beat B&H. Maybe there's no such an algorithm, or maybe the bounty was not high enough for people to publish theirs. If you really want such an algorithm maybe you should try help pulsecat organize higher bounty than 2BTC. I think there will be much more and much better algorithms published in that case, if for no other reason then it would enable much more time devoted to developing them. It's not exactly easy to do everything right, and it takes time. I don't think this is about bounty, if developer has the "super-algorithm", he won't need any bounty You got me wrong. Probably nobody has a "super-algorithm", but given enough time many developers may produce some very good ones. Developers don't have few weeks or a month free to experiment and try new things, but given attractive bounty some may decide to devote the time to it. And public contest is an excellent way to test the results. Round 4 was about short term strategy... I just played a bit with round 4 but: 2013/10/14 1:00 AM til 2013/11/10 4:00 PM: Itod 2385.80 USD K IIc 2313.52 USD Griny 2193.82 USD K I 2067.61 USD K IIa 1904.63 USD
However if you kept these running from 2013-10-14 1AM to 2013-12-09 22:00: K IIa 11.1264 BTC (10210.70 USD) K IIc 9.9756 BTC (9154.58 USD) Medji R5 8 .9082 BTC (8175.08 USD) Grinny (5607.27 USD) Itod 6.7131 BTC (5793.42 USD) Medji R4 5.3164 BTC (4689.17 USD)
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
December 10, 2013, 11:25:28 PM |
|
I read up on Ichimoku and I fixed a bug in Medji's code and used different parameters.
kugutsumen, your posts are confusing since it is not clear what are you talking about. If you backtested algorithms from this round on previous rounds - that's cool, and presenting results in clear manner would be appreciated. If you are bragging about some mysterious algorithm of yours which beats everything by a mile, but you don't want to post it - "No comment" on that, when you have some proof we can talk about it. If you are claiming there's a bug in Ichimoku example in Backtest->Featured part of the site, which are there for everyone to use, and you don't want to point which bug is it - that's really lame. Don't do such a thing.
|
|
|
|
kugutsumen
Newbie
Offline
Activity: 15
Merit: 0
|
|
December 11, 2013, 08:50:04 AM Last edit: December 11, 2013, 09:09:24 AM by kugutsumen |
|
I read up on Ichimoku and I fixed a bug in Medji's code and used different parameters.
kugutsumen, your posts are confusing since it is not clear what are you talking about. If you backtested algorithms from this round on previous rounds - that's cool, and presenting results in clear manner would be appreciated. I thought it was clear: I replied to a recent post about round 4 and short term trading. I showed that it was possible to profit on the short term with or without crash, compared my results as well as medji r5 version to the R4 results. If you are bragging about some mysterious algorithm of yours which beats everything by a mile, but you don't want to post it - "No comment" on that, when you have some proof we can talk about it.
Again I was really clear. It isn't a mysterious algo, it is a modified version of the Ichimoku example that Medji is also using. Please keep in mind that this is a quick hack, I am completely new to this... I read up on Ichimoku and I fixed a bug in Medji's code and used different parameters. I have sent a copy to Pulsecat as proof: I have to say the results of your backtest are exciting. Though the round has already started, I would love to have you participate next round. If you are claiming there's a bug in Ichimoku example in Backtest->Featured part of the site, which are there for everyone to use, and you don't want to point which bug is it - that's really lame. Don't do such a thing.
I think that the Ichimoku example has a small bug in it but that is based on my initial reading of the docs at ichimokutrader.com. I'll probably share my fix when round 6 opens. It will give me time to research my theory further, I don't see how this is lame.
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
December 11, 2013, 10:59:41 AM |
|
If you are claiming there's a bug in Ichimoku example in Backtest->Featured part of the site, which are there for everyone to use, and you don't want to point which bug is it - that's really lame. Don't do such a thing.
I think that the Ichimoku example has a small bug in it but that is based on my initial reading of the docs at ichimokutrader.com. I'll probably share my fix when round 6 opens. It will give me time to research my theory further, I don't see how this is lame. Well, it's a peace of opensource code you've got for free to work on. Holding back findings about bugs so everyone can continue to work on that same peace of code is lame in my book. You don't have to offer your fixes to everyone, but you should describe the bug.
|
|
|
|
kugutsumen
Newbie
Offline
Activity: 15
Merit: 0
|
|
December 11, 2013, 07:11:19 PM Last edit: December 11, 2013, 10:25:04 PM by kugutsumen |
|
If you are claiming there's a bug in Ichimoku example in Backtest->Featured part of the site, which are there for everyone to use, and you don't want to point which bug is it - that's really lame. Don't do such a thing.
I think that the Ichimoku example has a small bug in it but that is based on my initial reading of the docs at ichimokutrader.com. I'll probably share my fix when round 6 opens. It will give me time to research my theory further, I don't see how this is lame. Well, it's a peace of opensource code you've got for free to work on. Holding back findings about bugs so everyone can continue to work on that same peace of code is lame in my book. You don't have to offer your fixes to everyone, but you should describe the bug. Why should I do that? Why do you need to insult me? I've already stated that I will probably participate in round 6 and if I do, I will have to release the code. About open source: Nobody should be restricted by the software they use. There are four freedoms that every user should have:
the freedom to use the software for any purpose, the freedom to change the software to suit your needs, the freedom to share the software with your friends and neighbors, and the freedom to share the changes you make.
Hint: Freedom is the keyword... Until recently, my real life job was to find hard-to-find bugs and vulnerabilities in software and systems, mostly banking software (core banking, treasury systems, mobile banking, online banking...) We get paid a lot of money to do that. Open source or closed source software, it doesn't matter, we still get paid. Even the restrictive GPL doesn't force you to publish any changes unless you distribute compiled code. In copyright law, a derivative work is an expressive creation that includes major, copyright-protected elements of an original, previously created first work (the underlying work). The derivative work become a second, separate work independent in form from the first. The transformation, modification or adaption of the work must be substantial and bear its author's personality to be original and thus protected by copyright.In the particular case of cryptotrader.org, I actually paid Alex to use his service and infrastructure. Regarding Medji or the sample code, I don't know their respective history, where the source code comes from and who wrote it. What I know is that Goichi Hosoda, the creator of Ichimoku Kinkō Hyō is not getting anything from Alex or Medji or any of the traders and software engineers who uses his methology. If you had just spent an hour reading up on Ichimoku and took a critical look at the code, you would have noticed the bug right away because it is so obvious!! I even thought it was left there as an exercise. (and in the old days, released exploit code would includes bugs and typos on purpose to make life harder for script kiddies and what we would call lamers...)
|
|
|
|
shadyz
Member
Offline
Activity: 66
Merit: 10
|
|
December 12, 2013, 05:50:50 AM |
|
If people continue sharing great ideas ,the service will continue to get better. That's true for anything, but it should be especially true for the bitcoin community : )) No need to fight
|
|
|
|
pulsecat (OP)
|
|
December 12, 2013, 09:53:08 PM |
|
|
|
|
|
kugutsumen
Newbie
Offline
Activity: 15
Merit: 0
|
|
December 12, 2013, 10:13:38 PM |
|
Kugutsumen I2013-01-01 00:00 Simulation started. Balance: 5000.00 USD 2013-12-12 16:00 Simulation completed. Balance: 1303663.26 USD
n.b. I am not in the contest. However Pulsecat can confirm the results.
|
|
|
|
Grinny
|
|
December 13, 2013, 09:07:49 AM |
|
Kugutsumen I2013-01-01 00:00 Simulation started. Balance: 5000.00 USD 2013-12-12 16:00 Simulation completed. Balance: 1303663.26 USD
n.b. I am not in the contest. However Pulsecat can confirm the results. So you're not willing to share your algo in public like we all did? :/
|
|
|
|
Grinny
|
|
December 13, 2013, 12:12:09 PM |
|
I (and maybe a few others too?) got this PM from frogDog, a new user who has no permission to post yet. He asked me to post his message in this thread, so here it is: Hey, after all the mess kugutsumen created in the contest thread of cryptotrader, I decided to take a deep look into the strategy, and I'm pretty sure, that I found the bug. When you take a look on http://ichimokutrader.com/elements.html#Chikou you see, that the Chikou should be backplaced, and in the http://ichimokutrader.com/signals.html page at the bottom you can see, that it should be compared with the price now. So to state it clearly, its nothing else but comparing the actual price with the price 28 (or whatever) days ago. Putting this into the strategy leads to following results: https://cryptotrader.org/backtests/rjPQ3BPv54wtLbzuSHowever, its not as fancy, as the results look like. Actually, they trade totally the same except on the big crash 4/10. This lead to an advantage of around 14%, that just transfered to now. And you can reassure this, all the results kugutsumen posted, differ by the exact amount, meaning his algorithm doesnt differ much as medjis. Thank you kugutsumen anyway, as I wouldn't have found it, without someone mentioning it. Until today I just read in this forum, so I'm not eligble to post anything, so it would be nice, if you can post this in the contest thread as a quote.
|
|
|
|
kugutsumen
Newbie
Offline
Activity: 15
Merit: 0
|
|
December 13, 2013, 12:20:16 PM |
|
Kugutsumen I2013-01-01 00:00 Simulation started. Balance: 5000.00 USD 2013-12-12 16:00 Simulation completed. Balance: 1303663.26 USD
n.b. I am not in the contest. However Pulsecat can confirm the results. So you're not willing to share your algo in public like we all did? :/ Did you read my post? I said three times that I will wait until round 6 and I am currently writing a clean room implementation. I am trading live with version IIa.
|
|
|
|
kugutsumen
Newbie
Offline
Activity: 15
Merit: 0
|
|
December 13, 2013, 12:33:26 PM Last edit: December 13, 2013, 03:30:19 PM by kugutsumen |
|
I (and maybe a few others too?) got this PM from frogDog, a new user who has no permission to post yet. He asked me to post his message in this thread, so here it is: Hey, after all the mess kugutsumen created in the contest thread of cryptotrader, I decided to take a deep look into the strategy, and I'm pretty sure, that I found the bug. When you take a look on http://ichimokutrader.com/elements.html#Chikou you see, that the Chikou should be backplaced, and in the http://ichimokutrader.com/signals.html page at the bottom you can see, that it should be compared with the price now. So to state it clearly, its nothing else but comparing the actual price with the price 28 (or whatever) days ago. Putting this into the strategy leads to following results: https://cryptotrader.org/backtests/rjPQ3BPv54wtLbzuSHowever, its not as fancy, as the results look like. Actually, they trade totally the same except on the big crash 4/10. This lead to an advantage of around 14%, that just transfered to now. And you can reassure this, all the results kugutsumen posted, differ by the exact amount, meaning his algorithm doesnt differ much as medjis. Thank you kugutsumen anyway, as I wouldn't have found it, without someone mentioning it. Until today I just read in this forum, so I'm not eligble to post anything, so it would be nice, if you can post this in the contest thread as a quote. Well good frogDog Itod, instead of insulting me, frogDog did this the right thing: RTFM & Use the source and say thank you... 2013-09-01 04:00 Simulation started. Balance: 858.00000 BTC (122069.28 USD) Frogdog: 2013-12-13 06:00 Simulation completed. Balance: [b]1313416.44 USD[/b] Kugu I: Simulation completed. Balance: [b]1429824.14 USD[/b]
|
|
|
|
medji
Newbie
Offline
Activity: 31
Merit: 0
|
|
December 13, 2013, 09:58:16 PM Last edit: December 15, 2013, 06:22:50 AM by medji |
|
I (and maybe a few others too?) got this PM from frogDog, a new user who has no permission to post yet. He asked me to post his message in this thread, so here it is: Hey, after all the mess kugutsumen created in the contest thread of cryptotrader, I decided to take a deep look into the strategy, and I'm pretty sure, that I found the bug. When you take a look on http://ichimokutrader.com/elements.html#Chikou you see, that the Chikou should be backplaced, and in the http://ichimokutrader.com/signals.html page at the bottom you can see, that it should be compared with the price now. So to state it clearly, its nothing else but comparing the actual price with the price 28 (or whatever) days ago. Putting this into the strategy leads to following results: https://cryptotrader.org/backtests/rjPQ3BPv54wtLbzuSHowever, its not as fancy, as the results look like. Actually, they trade totally the same except on the big crash 4/10. This lead to an advantage of around 14%, that just transfered to now. And you can reassure this, all the results kugutsumen posted, differ by the exact amount, meaning his algorithm doesnt differ much as medjis. Thank you kugutsumen anyway, as I wouldn't have found it, without someone mentioning it. Until today I just read in this forum, so I'm not eligble to post anything, so it would be nice, if you can post this in the contest thread as a quote. Wow, I was quite perplexed by the claims of a bug, I thought it was in the writing of the ichimoku object (not that there isn't a bug... what the hell do I know?) not a difference in the strategy, the idea was that the chikou would give a bullish/bearish signal if it was above/below the kumo X periods past, but this looks better, thanks!
|
|
|
|
|
speedtrader
Newbie
Offline
Activity: 28
Merit: 0
|
|
December 18, 2013, 06:48:36 PM |
|
about trading, I'm invest 4 days return 100%? only here : I have live video from them
|
|
|
|
kugutsumen
Newbie
Offline
Activity: 15
Merit: 0
|
|
December 19, 2013, 12:41:20 PM |
|
Week 3 intermediate results
Kugutsumen Balance: 1430989.95 (USD)
|
|
|
|
kugutsumen
Newbie
Offline
Activity: 15
Merit: 0
|
|
December 21, 2013, 05:30:55 PM |
|
Starting balance — 1 January 2013: 5000 USD Ending balance — 21 December 2013:
Medji 1108538 USD (1714.45422 BTC) Itod 516255 USD (798.53891 BTC) (bought)
Non-contestants:
Kugutsumen I 1594742 USD (2466.41158 BTC) Frogdog 1309577 (2025.37762 BTC)
|
|
|
|
btcorbust
Newbie
Offline
Activity: 7
Merit: 0
|
|
December 23, 2013, 05:17:02 AM Last edit: December 23, 2013, 06:08:40 AM by btcorbust |
|
I've been playing around with my own version of the Ichimoku algorithm (inspired by medji's) and am finding that there is certainly no one-size-fits-all setting for it. Settings that have been tuned to work great over the full year turn out to be not so great when run over a more recent period such from 1st December.
In hindsight this is pretty obvious, but it highlights the fact you need to continually tweak your algorithm for current market conditions to get best results.
Anyway, I'm too late to enter but my current best for period 1st January to 21st December on MtGox with 0.55% fee is:
$1757649.52 (2630.01512 BTC)
|
|
|
|
sukiho
|
|
December 23, 2013, 05:24:18 PM |
|
what I havent seen yet but seems logical is an exponential backtesting system, similar to the way ema works, so the recent trades in a back test are given more importance then the older ones. how it would work Im not exactly sure and the coding is beyond me but I am sure the results would be much more akin to what the market is currently doing.
|
|
|
|
kugutsumen
Newbie
Offline
Activity: 15
Merit: 0
|
|
December 23, 2013, 08:56:10 PM |
|
I've been playing around with my own version of the Ichimoku algorithm (inspired by medji's) and am finding that there is certainly no one-size-fits-all setting for it. Settings that have been tuned to work great over the full year turn out to be not so great when run over a more recent period such from 1st December.
In hindsight this is pretty obvious, but it highlights the fact you need to continually tweak your algorithm for current market conditions to get best results.
Anyway, I'm too late to enter but my current best for period 1st January to 21st December on MtGox with 0.55% fee is:
$1757649.52 (2630.01512 BTC)
Impressive I haven't played much with the full year performance since my k1b.. which is shy of 1.67M...
|
|
|
|
btcorbust
Newbie
Offline
Activity: 7
Merit: 0
|
|
December 24, 2013, 12:52:46 AM |
|
I've been playing around with my own version of the Ichimoku algorithm (inspired by medji's) and am finding that there is certainly no one-size-fits-all setting for it. Settings that have been tuned to work great over the full year turn out to be not so great when run over a more recent period such from 1st December.
In hindsight this is pretty obvious, but it highlights the fact you need to continually tweak your algorithm for current market conditions to get best results.
Anyway, I'm too late to enter but my current best for period 1st January to 21st December on MtGox with 0.55% fee is:
$1757649.52 (2630.01512 BTC)
Impressive I haven't played much with the full year performance since my k1b.. which is shy of 1.67M... Thanks. I'm not actually trading with that one as its performance over more recent periods isn't as good as other variants I have. Currently I have two bots trading live against MtGox and BTC-e (each tuned to the particular exchange) with around 1 BTC each. I would like to throw some more BTC at them each, but want to see how they get on for a month or so first... it's still only early days so no reports on actual results yet...
|
|
|
|
kugutsumen
Newbie
Offline
Activity: 15
Merit: 0
|
|
December 24, 2013, 03:23:31 AM |
|
I've been playing around with my own version of the Ichimoku algorithm (inspired by medji's) and am finding that there is certainly no one-size-fits-all setting for it. Settings that have been tuned to work great over the full year turn out to be not so great when run over a more recent period such from 1st December.
In hindsight this is pretty obvious, but it highlights the fact you need to continually tweak your algorithm for current market conditions to get best results.
Anyway, I'm too late to enter but my current best for period 1st January to 21st December on MtGox with 0.55% fee is:
$1757649.52 (2630.01512 BTC)
Impressive I haven't played much with the full year performance since my k1b.. which is shy of 1.67M... Thanks. I'm not actually trading with that one as its performance over more recent periods isn't as good as other variants I have. Currently I have two bots trading live against MtGox and BTC-e (each tuned to the particular exchange) with around 1 BTC each. I would like to throw some more BTC at them each, but want to see how they get on for a month or so first... it's still only early days so no reports on actual results yet... Same here... and right now I am working on getting an offline backtest system.
|
|
|
|
btcorbust
Newbie
Offline
Activity: 7
Merit: 0
|
|
December 27, 2013, 03:46:11 AM |
|
Here is the code I also posted on the cryptotrader forum: https://cryptotrader.org/backtests/3uJPMJdhY9Ct4S6i8Results: >2million for the year depending on current exchange rate (the linked backtest is only up until 21 Dec - adjust as appropriate) It combines a rewritten Ichimoku algorithm with Heikin-Ashi candlesticks and Parabolic SAR. I share it here because it is a combination of my own work and that of a few others. Please review and recommend further improvements!
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
December 27, 2013, 07:22:39 AM |
|
Here is the code I also posted on the cryptotrader forum: https://cryptotrader.org/backtests/3uJPMJdhY9Ct4S6i8Results: >2million for the year depending on current exchange rate (the linked backtest is only up until 21 Dec - adjust as appropriate) It combines a rewritten Ichimoku algorithm with Heikin-Ashi candlesticks and Parabolic SAR. I share it here because it is a combination of my own work and that of a few others. Please review and recommend further improvements! Nice Ichimoku re-write, interesting. How much is gained with those Heikin-Ashi candles compared to regular ones, have you intensively back-tested that?
|
|
|
|
btcorbust
Newbie
Offline
Activity: 7
Merit: 0
|
|
December 27, 2013, 07:43:34 AM |
|
Here is the code I also posted on the cryptotrader forum: https://cryptotrader.org/backtests/3uJPMJdhY9Ct4S6i8Results: >2million for the year depending on current exchange rate (the linked backtest is only up until 21 Dec - adjust as appropriate) It combines a rewritten Ichimoku algorithm with Heikin-Ashi candlesticks and Parabolic SAR. I share it here because it is a combination of my own work and that of a few others. Please review and recommend further improvements! Nice Ichimoku re-write, interesting. How much is gained with those Heikin-Ashi candles compared to regular ones, have you intensively back-tested that? I'm still testing various scenarios but am finding that the Heikin-Ashi candles do give a tangible improvement in performance. Note that the HA candles can be easily enabled/disabled in the code via the enable_ha context param if you want to test it for yourself.
|
|
|
|
LoneMiner
Newbie
Offline
Activity: 16
Merit: 0
|
|
December 27, 2013, 11:26:35 PM |
|
Here is the code I also posted on the cryptotrader forum
Quite impressive.
|
|
|
|
leshow
Newbie
Offline
Activity: 48
Merit: 0
|
|
December 30, 2013, 01:54:26 AM |
|
how do you use cryptotrader to do real trading?
|
|
|
|
nitrous
|
|
December 30, 2013, 01:56:57 AM |
|
how do you use cryptotrader to do real trading?
You can get a trading account here https://cryptotrader.org/plans. Free accounts can develop and backtest algorithms, but to actually trade you need to subscribe.
|
|
|
|
leshow
Newbie
Offline
Activity: 48
Merit: 0
|
|
December 30, 2013, 02:19:11 AM |
|
is it a trustworthy service? i just paid for a month of pro, im gonna give it a try on bitstamp. i wrote a 10-21 ema trader in nodejs but this looks like it will be easier (i can turn my computer off) lol
|
|
|
|
autopoiesis
Newbie
Offline
Activity: 3
Merit: 0
|
|
December 30, 2013, 06:46:23 PM |
|
Leshow, I've found the service very reliable, but if you've already rolled your own you might consider running it on an Amazon EC2 instance (or whatever other provider you prefer) if your issue is simply keeping processes alive.
|
|
|
|
leshow
Newbie
Offline
Activity: 48
Merit: 0
|
|
December 31, 2013, 04:35:37 AM |
|
Leshow, I've found the service very reliable, but if you've already rolled your own you might consider running it on an Amazon EC2 instance (or whatever other provider you prefer) if your issue is simply keeping processes alive.
i think i'd have to rewrite it for the ichimoku strategy, it seems to be getting much higher returns in backtesting
|
|
|
|
shadyz
Member
Offline
Activity: 66
Merit: 10
|
|
January 01, 2014, 02:05:24 AM |
|
Leshow, I've found the service very reliable, but if you've already rolled your own you might consider running it on an Amazon EC2 instance (or whatever other provider you prefer) if your issue is simply keeping processes alive.
i think i'd have to rewrite it for the ichimoku strategy, it seems to be getting much higher returns in backtesting Indeed, ichimoku seems to be the one of the most profitable indicator with combination with others I hear some people having trouble backtesting scripts, but i haven't heard a lot of people=customers talking about the cryptotrader service in terms of buying and using the scripts on here?
|
|
|
|
pulsecat (OP)
|
|
January 01, 2014, 02:48:36 PM Last edit: January 01, 2014, 02:59:39 PM by pulsecat |
|
Hi, everyone. It is time for the results of Round #5 itod http://cryptotrader.org/backtests/mmyB3uQGgkfUssL2YBalance: 798.53891 BTC (643286.18 USD) ZirconiumX http://cryptotrader.org/backtests/Ywno99YxczSIznZZpBalance: 278.94306 BTC (224710.67 USD) sukiho http://cryptotrader.org/backtests/h4X7TmoFp2fKHOys9Balance: 1178.11826 BTC (949067.33 USD) exuals http://cryptotrader.org/backtests/dSr8HcASNzCgq7GrlBalance: 376.58086 BTC (303365.64 USD) Kiri11 http://cryptotrader.org/backtests/7a3Ab4DURUwrxe5uvBalance: 640.39500 BTC (515888.76 USD) medji http://cryptotrader.org/backtests/0CKXzdyjGMjUxrJ6HBalance: 1416.25457 BTC (1140904.94 USD)MeatPopsicle http://cryptotrader.org/backtests/SymxrJD5VBibrJEJJBalance: 149.21806 BTC (120206.94 USD) Grinny http://cryptotrader.org/backtests/IM2sVXunG7p5odKyYBalance: 427.99385 BTC (344782.86 USD) Buy'n'Hold http://cryptotrader.org/backtests/UnL49AiFOY8y2yiHpBalance: 368.06042 BTC (296501.75 USD) medji takes first place with Ichimoku strategy. Congratulations! medji, shoot me a PM with your btc address to get the prize. Though his result - $1,140,904 from $5,000 in one year is exciting, during the last month, Cryptotrader users developed even more profitable strategies. Check out these threads https://cryptotrader.org/topics/349497/the-best-strategy-for-2013 and https://cryptotrader.org/topics/367450/best-strategy-for-2013-continued to see algorithms that potentially could earn more than $4,000,000 in one year!! Thanks for everyone who took part in our competition! I hope in the new year we will have more great contests here!
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
January 01, 2014, 03:08:22 PM |
|
medji takes first place with Ichimoku strategy. Congratulations! medji, shoot me a PM with your btc address to get the prize.
Congrats medji, well deserved! Thanks for everyone who took part in our competition! I hope in the new year we will have more great contests here!
Great, we hope that too!
|
|
|
|
weaknesswaran
|
|
January 01, 2014, 05:11:40 PM |
|
What would be the result if they were backtested for 2011 or 2012?
|
|
|
|
medji
Newbie
Offline
Activity: 31
Merit: 0
|
|
January 02, 2014, 07:40:38 AM |
|
Thanks, glad to finally take this one down, also really nice to see that the strategy I wrote a while ago has sparked some interest into ichimoku and got the algo rewrote and made quite better. Edit: prize received, thanks pulsecat! hopefully there will be other rounds and with them more interesting and profitable strategies.
|
|
|
|
Grinny
|
|
January 02, 2014, 10:09:42 AM |
|
gratz medji! and i would love to see a LTC/USD contest here
|
|
|
|
LoneMiner
Newbie
Offline
Activity: 16
Merit: 0
|
|
January 02, 2014, 09:06:48 PM |
|
Congratulations medji!
For the next round I personally think the period should start from November 2013, when the market started to show movement. Anything before then is largely pointless in my opinion.
|
|
|
|
shadyz
Member
Offline
Activity: 66
Merit: 10
|
|
January 05, 2014, 04:13:14 AM |
|
Congratulations medji!
For the next round I personally think the period should start from November 2013, when the market started to show movement. Anything before then is largely pointless in my opinion.
Yeah ,or there needs to be different contests depending on when we think the best representation of current market conditions are/were and another one for the whole year as well
|
|
|
|
tubbyjr
|
|
January 05, 2014, 01:18:49 PM |
|
When will the next round be
|
|
|
|
armodilloben
|
|
January 07, 2014, 02:08:34 PM |
|
They really made this much from 5,000$ starting capital. Impressive, well played.
|
|
|
|
nfigueir
|
|
January 08, 2014, 12:13:06 AM |
|
Hi guys,
I accidentally bumped into this thread and "played" with the simulation engine a bit to see if I could make some sense of it and now I have a question that I hope someone can help.
The simulations assume that you are able to sell/buy any amount you would like at any particular time. Is this correct? If it is, isn't too far off from the truth?
I was looking for at the bots which netted 1million from 5000US over one year and the last trades were stuff like "buy 2400BTC", "sell 3000BTC", etc... Is this a realistic model? I guess that in reality you will never be able to buy such a large amount of coins at a given price.
Thanks in advance.
|
Gulden NLG: GdQgmEN1ptPzKpnMmRw7pwAuPGiJZCZjHi Europecoin ERC: Edg1HCFSsVweehu35YeHXfURKXgEi7qnLm
|
|
|
Sephera
|
|
January 08, 2014, 03:19:55 PM Last edit: January 08, 2014, 03:32:32 PM by Sephera |
|
Is it possible someone can help me setup and run a bot I really need to make money... I can give 5-10% of earnings. My strategy is to buy and sell but ultimately hold for the long term. I need an algorithm to increase my BTC holding in the long term.
|
|
|
|
Sephera
|
|
January 09, 2014, 09:06:14 AM |
|
Bump.
|
|
|
|
FireBTC
Newbie
Offline
Activity: 14
Merit: 0
|
|
January 10, 2014, 04:47:05 AM Last edit: March 05, 2014, 10:28:20 AM by malevolent |
|
This is a great contest. I don't know how to make something automated, but I can read into BTC trends pretty well. I have been doing it for a little over 2 years now. I just use this site: [malware url removed] to find the trends.
|
|
|
|
blo8i
|
|
January 10, 2014, 10:13:34 AM |
|
this is looking great, i just wish i was good at finding trends. looking forward to see how good the winner script doo. and if it becomes public, i will not hesitate donating if it helps the community earn some ekstra cash.
GLHF to all who is doing this
|
|
|
|
LoneMiner
Newbie
Offline
Activity: 16
Merit: 0
|
|
January 18, 2014, 10:21:38 PM |
|
Looks like you need to pay for backtesting now. Fair enough if it was getting smashed and affecting paying customers, but harder to show off its potential to new people now.
|
|
|
|
bit-fxtrader
|
|
January 19, 2014, 07:02:53 AM |
|
anyone trading the automated strategies on cryptotrader with real cash? Is it stable and working as intended?
|
|
|
|
Coldstuff
Newbie
Offline
Activity: 1
Merit: 0
|
|
January 26, 2014, 04:34:41 AM |
|
Sephra,
If you really need help setting up a BTC bot PM me and I will walk you through how to do it. You don't have to give me a percentage of your earnings, but if you make a lot you can send me something.
Remember this is HIGH RISK, you are trading with a mixture of unpredictable individuals as well as professional traders who will cut you to ribbons if you give them the chance. Also, some of these strategies that produced millions would have realistically pushed the market around toward the end making huge trades that could have pushed the market $20 of $50 each way. Depending on the conditions that could have made the real results better or worse. And one final note of caution: past performance doesn't predict the future although that is all we really have. The last week of this month could get dicey as I have already seen some panic selling and such related to peoples worries over what is happening in China and elsewhere. If you doubt the confusion just read the troll boxes on the exchanges.
That said it is pretty easy to set up: 1) Open an account at one of the major exchanges, I prefer BTC-e but Bitstamp is also good. Prices on Mt. Gox look attractive and it is more stable but stability is your enemy on a lot of these trading algorithms. 2) I also have a wallet at Coinbase but other wallets are good too. This will facilitate you moving money in and out of the brokerage. I also mine various crypto currencies and have them deposited directly to BTC-e as Coinbase doesn't handle them but that is digressing. 3) Fund your brokerage account. The easiest, fastest and safest way is to obtain bitcoins and then send them to them. Wire transfers to these exchanges seem to have a lot of problems. Beware, if you verify enough information you can get "some" (like $100) worth instantly but if you want any significant amount you are going to have to wait a few days. You are going to pay a bit of a premium for them plus a small fee, and in the meantime you are stuck where you can't sell them if they go down. I don't like this but haven't found a better way other than mining. Maybe someone else can enlighten us. Ebay and other places you can buy them seem to want too much of a premium for me. 4) Every bot needs a Key and a Secret. On BTC-e you go to finances and click "Generate Key". 5) Decide what bot you want to use. I have tried several and for the most part they are very primitive trading algorithms. Most of them don't tell you how they trade or give you enough control. The worst part is other people already have figured out how these will trade (assuming the people who wrote them haven't beat them to it) and they can use this information to leave you holding the bag. I just recently found cryptotrader.org from this forum and have been pretty happy with it. The willingness of people to share their algorithms is great and the fact you can see the actual code you are running and make your own tweaks make it the best I have found. Do NOT buy a bot that will not let you change parameters and time frames and backtest it with historical data before you put it live. 5) Once you have settled on a trading algorithm (cryptotrader has many you can choose from and even modify yourself) and backtested it to your satisfaction you just tell it to start trading it. If you haven't already it will ask for your exchange and your key and secret. Click Save and hold on!
I would really suggest you start with a small amount (that you can afford to lose) and see how it goes. You really do need at least a few hundred dollars to give most of these algorithms room to work though.
|
|
|
|
pulsecat (OP)
|
|
February 11, 2014, 03:58:46 PM |
|
Looks like you need to pay for backtesting now. Fair enough if it was getting smashed and affecting paying customers, but harder to show off its potential to new people now.
Free backtesting is available again. Though some limits apply to it.
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
February 11, 2014, 04:02:57 PM |
|
Looks like you need to pay for backtesting now. Fair enough if it was getting smashed and affecting paying customers, but harder to show off its potential to new people now.
Free backtesting is available again. Though some limits apply to it. Great! May we know the details about the limits?
|
|
|
|
erpbridge
Legendary
Offline
Activity: 954
Merit: 1000
|
|
February 19, 2014, 06:04:52 AM |
|
How well do these bots handle the current downturn? I have been trying to run the Mediji bot, and so far it has been dropping my profits like crazy in the past few weeks.
|
|
|
|
dddbtc
|
|
February 19, 2014, 07:12:13 PM |
|
This is a great contest. I don't know how to make something automated, but I can read into BTC trends pretty well. I have been doing it for a little over 2 years now. I just use this site: btctradingsignals.netau.net to find the trends.
I don't trust your blank webpage that instantly attempts to open a Java applet on my computer. Stay away from that link.
|
|
|
|
pulsecat (OP)
|
|
February 22, 2014, 04:54:38 PM |
|
How well do these bots handle the current downturn? I have been trying to run the Mediji bot, and so far it has been dropping my profits like crazy in the past few weeks.
Some algorithms are still profitable even when the market is moving down. For example check out Adaptive EMA results https://cryptotrader.org/backtests/akdD5nWNFMKS3pNS5
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
February 22, 2014, 07:21:22 PM |
|
How well do these bots handle the current downturn? I have been trying to run the Mediji bot, and so far it has been dropping my profits like crazy in the past few weeks.
Some algorithms are still profitable even when the market is moving down. For example check out Adaptive EMA results https://cryptotrader.org/backtests/akdD5nWNFMKS3pNS5I know the best way to prove some algos are still profitable: quick 2-weeks (or 4-weeks?) contest with a nice prize pool There's no proof like seeing it with your own eyes, and nothing is better to spark ideas than some healthy competition. What do other previous or future contestants think about this?
|
|
|
|
medji
Newbie
Offline
Activity: 31
Merit: 0
|
|
March 02, 2014, 11:16:56 AM |
|
How come that cryptotrader managed to fetch tick data even when btc-e was under maintenance? Here in bitcoin charts: 2014-02-28 07:00:00, please enlighten me if I'm missing something...
|
|
|
|
pulsecat (OP)
|
|
April 01, 2014, 04:43:10 PM |
|
How come that cryptotrader managed to fetch tick data even when btc-e was under maintenance? Here in bitcoin charts: 2014-02-28 07:00:00, please enlighten me if I'm missing something...
Even if API is down, Cryptotrader's backend flushes collected tick data to ensure it won't be lost.
|
|
|
|
needtowin
Newbie
Offline
Activity: 1
Merit: 0
|
|
April 05, 2014, 02:22:41 PM |
|
Hi guys how to implement the source code to meta4 thanks for your time
|
|
|
|
5thStreetResearch
|
|
April 11, 2014, 12:36:36 AM |
|
Hi guys how to implement the source code to meta4 thanks for your time
it's time to get out of the old MetaTrader paradigm, step-up to new cloud based trading platform designed for Bitcoin I have the perfect automated trading system. Write a script that just keeps shorting bitcoin until its back at $100
|
|
|
|
pulsecat (OP)
|
|
April 15, 2014, 08:20:41 PM |
|
Hi guys how to implement the source code to meta4 thanks for your time
it's time to get out of the old MetaTrader paradigm, step-up to new cloud based trading platform designed for Bitcoin I have the perfect automated trading system. Write a script that just keeps shorting bitcoin until its back at $100 This is certainly a great strategy if you want to miss all your trading opportunities.
|
|
|
|
gogle
Member
Offline
Activity: 104
Merit: 20
|
|
June 24, 2014, 06:22:53 AM |
|
Anyway, half arsed attempt at following a really short ema slope's rate of change, and direction change. Thought process was the rate of increase / decrease would have to slow down before it switches direction. Watch for the slow-down in the rise or fall and buy/sell when the direction change happens.
|
|
|
|
statdude
Legendary
Offline
Activity: 1498
Merit: 1000
|
|
July 07, 2014, 07:48:29 PM |
|
What are the best automated traders ATM?
|
|
|
|
itod
Legendary
Offline
Activity: 1974
Merit: 1076
^ Will code for Bitcoins
|
|
July 07, 2014, 08:49:09 PM |
|
What are the best automated traders ATM?
That would be interesting to find out, but since there is no contest you are left to judge by each bot-programmer bragging about his bot past performances
|
|
|
|
Sharky444
|
|
July 10, 2014, 06:52:12 PM |
|
Is this contest still active? The first post mentions 2013.
|
|
|
|
|