Bitcoin Forum
December 04, 2016, 10:27:42 AM *
News: Latest stable version of Bitcoin Core: 0.13.1  [Torrent].
 
   Home   Help Search Donate Login Register  
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 [13] 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 »
  Print  
Author Topic: Goomboo's Journal  (Read 250716 times)
Goomboo
Sr. Member
****
Offline Offline

Activity: 409



View Profile
January 27, 2012, 01:18:44 AM
 #241

Alright, got it!

Now let's do a simple backtest of a 10 / 21 daily SMA crossover.  I entered it as a stock, so this assumes you trade in 100 BTC blocks.  See why I'm totally skeptical of this and why I don't think it represents the future? Tongue




1480847262
Hero Member
*
Offline Offline

Posts: 1480847262

View Profile Personal Message (Offline)

Ignore
1480847262
Reply with quote  #2

1480847262
Report to moderator
1480847262
Hero Member
*
Offline Offline

Posts: 1480847262

View Profile Personal Message (Offline)

Ignore
1480847262
Reply with quote  #2

1480847262
Report to moderator
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction. Advertise here.
1480847262
Hero Member
*
Offline Offline

Posts: 1480847262

View Profile Personal Message (Offline)

Ignore
1480847262
Reply with quote  #2

1480847262
Report to moderator
1480847262
Hero Member
*
Offline Offline

Posts: 1480847262

View Profile Personal Message (Offline)

Ignore
1480847262
Reply with quote  #2

1480847262
Report to moderator
1480847262
Hero Member
*
Offline Offline

Posts: 1480847262

View Profile Personal Message (Offline)

Ignore
1480847262
Reply with quote  #2

1480847262
Report to moderator
Goomboo
Sr. Member
****
Offline Offline

Activity: 409



View Profile
January 27, 2012, 01:20:52 AM
 #242

Blah :p


M4v3R
Hero Member
*****
Offline Offline

Activity: 607



View Profile
January 27, 2012, 01:33:00 AM
 #243

Blah :p




Nice one. One could actually earn a nice pile of cash if he used it back then.
Goomboo
Sr. Member
****
Offline Offline

Activity: 409



View Profile
January 27, 2012, 01:46:42 AM
 #244

Looking over these backtest results does give me something interesting to talk about.


Look at the MAE (Max adverse excursion).  We've said this in a previous post, but it's basically how far the price moved on average against you before the trade ended for a profit or a loss.
-MAE of 10%.  This means BTC price moved 10% against you on average in each trade.
-Imaging using 10:1 leverage :p
-You're guaranteed to be wiped out...mathematically!
StewartJ
Sr. Member
****
Offline Offline

Activity: 392



View Profile
January 27, 2012, 02:05:46 AM
 #245

Looking over these backtest results does give me something interesting to talk about.


Look at the MAE (Max adverse excursion).  We've said this in a previous post, but it's basically how far the price moved on average against you before the trade ended for a profit or a loss.
-MAE of 10%.  This means BTC price moved 10% against you on average in each trade.
-Imaging using 10:1 leverage :p
-You're guaranteed to be wiped out...mathematically!

Looks like the proof is in the pudding with the performance table.

If your start/end dates were for the last 6 months instead, would that provide better numbers?
elux
Legendary
*
Offline Offline

Activity: 1454



View Profile
January 27, 2012, 02:19:51 AM
 #246

Now let's do a simple backtest of a 10 / 21 daily SMA crossover.




Monthly profit, ~30%. Cumulative profit ~4700%. That's uh.... Completely nuts.  Smiley

Is that the first or best strategy you backtested?
How did the 10 / 21 hourly EMA fare in comparison?

Goomboo
Sr. Member
****
Offline Offline

Activity: 409



View Profile
January 27, 2012, 02:26:32 AM
 #247

Looking over these backtest results does give me something interesting to talk about.


Look at the MAE (Max adverse excursion).  We've said this in a previous post, but it's basically how far the price moved on average against you before the trade ended for a profit or a loss.
-MAE of 10%.  This means BTC price moved 10% against you on average in each trade.
-Imaging using 10:1 leverage :p
-You're guaranteed to be wiped out...mathematically!

Looks like the proof is in the pudding with the performance table.

If your start/end dates were for the last 6 months instead, would that provide better numbers?

Yeah, ignore the dates - I didn't feel like working with the correct dates so I just pulled my dates from another currency pair and pasted them over them...so the dates are absolutely meaningless :p...it doesn't influence the performance one bit though
Goomboo
Sr. Member
****
Offline Offline

Activity: 409



View Profile
January 27, 2012, 02:41:01 AM
 #248

Now let's do a simple backtest of a 10 / 21 daily SMA crossover.


Monthly profit, ~30%. Cumulative profit ~4700%. That's uh.... Completely nuts.  Smiley

Is that the first or best strategy you backtested?
How did the 10 / 21 hourly EMA fare in comparison?



Just backtested the EMA for validity's sake - earned a 5067% return with a MAE of 8.85%.

The actual return would be a lot more - would you only buy 100 BTC when the price was .86?  The correct answer to this question is: how much does my risk-management methodology tell me to trade.

But I think it's a really, really good idea to realize that these results DO NOT represent the future.  This is why I didn't want to backtest - it's meaningless!  Don't look at these results and go "ooooh" - it's going to happen again.  "But what if...".  No.

***Anything you backtest across this data will be silly - THIS IS NOT HOW MARKETS CONSISTENTLY BEHAVE***
Goomboo
Sr. Member
****
Offline Offline

Activity: 409



View Profile
January 27, 2012, 02:49:50 AM
 #249

Looking over these backtest results does give me something interesting to talk about.


Look at the MAE (Max adverse excursion).  We've said this in a previous post, but it's basically how far the price moved on average against you before the trade ended for a profit or a loss.
-MAE of 10%.  This means BTC price moved 10% against you on average in each trade.
-Imaging using 10:1 leverage :p
-You're guaranteed to be wiped out...mathematically!

Looks like the proof is in the pudding with the performance table.

If your start/end dates were for the last 6 months instead, would that provide better numbers?


Meh.  I was worried about posting backtested results.

"Goomboo has a system that made 5000%!!!"
-Goomboo didn't make it
-It's a concept that I'm trying to teach through a system, not a system
-Those results aren't happening again - tough for all of us!

I'm not going to tinker with it to find the best results - please understand this - they are meaningless.  This is backtested a trend-following system across a bubble.

Wouldn't you have just looooved to be long tulip bulbs in the 1600s?

http://en.wikipedia.org/wiki/Tulip_mania

Backtest that and use it as a justification for raiding the florist Tongue
Goomboo
Sr. Member
****
Offline Offline

Activity: 409



View Profile
January 27, 2012, 03:07:26 AM
 #250

I suppose now is a good time as ever to explain what "Profit Factor" is on the charts above.

Profit factor = gross profits / gross losses
-If the number is above 1, the system is profitable
-If the number is below 1, the system is not profitable

-A trader can do very, very well with a profit factor of around 1.5.  The idea here would be that you have a marginally profitability strategy that you trade across many, many markets - think Black Jack.  If you can beat the dealer consistently, you should be at every table in the world at once.

-An figured we quote in the industry for profit factor is 5.  If a trading strategy has a profit factor of 5 or more, we consider it to be curve fit.  Curve fitting is basically staring at the chart, seeing anomalies that probably won't repeat in the future, and creating a trading strategy that exploits them.

A practical example of curve fitting would be:
-If the year is 1995 -> buy the S&P 500
-If the year is 2000 -> sell your S&P 500 position
See how worthless this trading strategy is?

Remember our simple moving average crossover strategy on the EURUSD that we backtested across 10 years?  We took daily candles and took the trades no matter what?  That system had a profit factor of 1.4.  It outperformed the market, was robust, and I have full confidence trading it in the future because I know that it has proven an edge for decades and captures the essence of the concept of trend following.
Goomboo
Sr. Member
****
Offline Offline

Activity: 409



View Profile
January 27, 2012, 03:32:05 AM
 #251

For you who are desiring to become traders,


Do you see how hard this job can be?  I am asking you to turn a cautious and wary eye towards one of the most beautiful things you may have ever seen in the markets: The Holy Grail.  If you want to be a trader, you MUST be objective above all.  If you think of how much money you can gain or lose, you shouldn't even trade.  The whole idea of trading is that the only reason I'm not parking my money in a risk-free security is that I believe that I have the ability to generate a risk-adjusted return on my capital greater than my alternatives.  I'm not thinking dollars, I'm not thinking wealth, I'm thinking consistent growth which I can scale across multiple markets.  I'm not after emotional satisfaction, admiration from others, or a cool job title.  I am after consistent and scalable profit.


If you want to design a system which works in the BTC market, here is an idea how you should go about it:

-Start your studies in the "real markets".  By real, I mean a market in which the other side of your trade is a regimented professional who has done this for 30 years and is out to win.  Pure competition.  Learn from the school of hardknocks, not the school of wistful thinking.

-Begin with a general concept: what are you after?  Trading with the trend or against it.  Do you want to scalp or follow the trend.  Move to specific:  across what timeframe?  What is your patience level?  Can you really only trade 5 times a year?

-Test and prove this system in the "real" market - if it can stand the test of regimented professionals and trillions of traded dollars, then it can surely stand in the BTC market of "weekend" and "couch" traders.

For more specific instructions on testing, please see one of the earlier posts.
Goomboo
Sr. Member
****
Offline Offline

Activity: 409



View Profile
January 27, 2012, 03:43:31 AM
 #252

Still short from $6.28

+17% unrealized PNL on the trade

Prices could go to 0 or they could reverse and stop me out at a loss - it doesn't matter, I have a robust system that I have tested in the most competitive financial markets in the world and I will follow this system.

Do you?
nrd525
Legendary
*
Offline Offline

Activity: 1182


View Profile
January 27, 2012, 03:52:50 AM
 #253

Would the results of the historical test be affected by the fact that you couldn't reliably short BTC until recently?

Would the market have been affected by this? 
(If so, how?)

Don't day trade.
Goomboo
Sr. Member
****
Offline Offline

Activity: 409



View Profile
January 27, 2012, 04:05:10 AM
 #254

Would the results of the historical test be affected by the fact that you couldn't reliably short BTC until recently?

Would the market have been affected by this? 
(If so, how?)

If you examine the results, you will see that only 103% of the profit came from short trades.  Look at the three columns and their headings.  The far right column is only short trades.  The long trades would have occurred as advertised.
stochastic
Hero Member
*****
Offline Offline

Activity: 532


View Profile
January 27, 2012, 04:08:59 AM
 #255

Would the results of the historical test be affected by the fact that you couldn't reliably short BTC until recently?

Would the market have been affected by this? 
(If so, how?)

If you examine the results, you will see that only 103% of the profit came from short trades.  Look at the three columns and their headings.  The far right column is only short trades.  The long trades would have occurred as advertised.

When do you have time to trade with all these posts...  Maybe you are a fast thinker.  When I make a post it takes me at least 30 minutes... including this one

Introducing constraints to the economy only serves to limit what can be economical.
Goomboo
Sr. Member
****
Offline Offline

Activity: 409



View Profile
January 27, 2012, 04:19:16 AM
 #256

Would the results of the historical test be affected by the fact that you couldn't reliably short BTC until recently?

Would the market have been affected by this? 
(If so, how?)

If you examine the results, you will see that only 103% of the profit came from short trades.  Look at the three columns and their headings.  The far right column is only short trades.  The long trades would have occurred as advertised.

When do you have time to trade with all these posts...  Maybe you are a fast thinker.  When I make a post it takes me at least 30 minutes... including this one

I'm a system trader - I don't sit there looking for opportunities, I wait for them to come to me.  When I have some free time to spend on the forum (tonight), it's mostly spent talking about trading concepts / systems / fielding questions.

So yeah, even though the markets are going crazy right now, I'm not trading - I have my position.


And right here let me say one thing:  After spending many years in Wall Street and after making and losing millions of dollars I want to tell you this:  It never was my thinking that made the big money for me. It always was my sitting.  Got that?  My sitting tight!  It is no trick at all to be right on the market.  You always find lots of early bulls in bull markets and early bears in bear markets.  I’ve known many men who were right at exactly the right time, and began buying and selling stocks when prices were at the very level which should show the greatest profit.  And their experience invariably matched mine – that is, they made no real money out of it.  Men who can both be right and sit tight are uncommon.  I found it one of the hardest things to learn.  But it is only after a stock operator has firmly grasped this that he can make big money.  It is literally true that millions come easier to a trader after he knows how to trade than hundreds did in the days of his ignorance.

-Jesse Livermore

http://www.amazon.com/Reminiscences-Stock-Operator-Investment-Classics/dp/0471770884/ref=sr_1_1?ie=UTF8&qid=1327637856&sr=8-1
Sargasm
Member
**
Offline Offline

Activity: 112


View Profile
January 27, 2012, 04:20:19 AM
 #257

How do you import the data into ninjatrader?  I might be retarded, but I cannot figure it out. 

I know the free version is a little neutered...
Goomboo
Sr. Member
****
Offline Offline

Activity: 409



View Profile
January 27, 2012, 04:27:32 AM
 #258

How do you import the data into ninjatrader?  I might be retarded, but I cannot figure it out.  

I know the free version is a little neutered...

No, the free version is the same as the full version.  They are identical except you can't place live trades with the free version.

I'll just go ahead and post this.  Here's the data to import.

http://tinyurl.com/7nhd9ql

-Create an instrument in the instrument manager called BTC
--Make it a stock, set the exchange to NASDAQ
--Add it to the default list by clicking the arrow key to push it over there

-Add historical data through the historical data manager
--Import the file from the zip

-Backtest is under strategy analyzer
-When charting, be sure to chart daily data

The dates are meaningless, I just copied some from another currency just to get a "quick fix".
stochastic
Hero Member
*****
Offline Offline

Activity: 532


View Profile
January 27, 2012, 04:33:05 AM
 #259

Did you produce that chart showing historical success of your plan?

If so, how?

Yes,

It is using NinjaTrader - an excellent automated and discretionary trading platform.

Basically you program a strategy (or use their easy-coding utility) and backtest the strategy on historical data.

If you're interested in testing your ideas and automated trading, I highly recommend you learn about the platform.

www.ninjatrader.com

Feel free to ask me any questions about it if you're interested.


Im trying to sign up but it says I need a broker... Can I not just try this out? I dont have a broker.  

Also less important / off topic: why do people need brokers?  Why cant they just connect to the exchange and make trades like we all do to Gox?

I tried to download but it says I need Windows 7 or whatever version they are at now.  I had to use a Windows computer for a presentation and had no idea what I was looking at.  I think I will stick with R for backtesting.  It is open-source and is a powerful statistical tool.  I don't have to worry about faulty statistical methods in R.  There is even a package that allows trading through Interactive Brokers.

Ill check this out too Thanks

EDIT WTF is that...

R is a programming language that is used by most statisticians today.  Even if you don't know how to program you can use the tools available to analyze data.  There are some versions with GUI but I prefer working in the global environment.

Introducing constraints to the economy only serves to limit what can be economical.
Sargasm
Member
**
Offline Offline

Activity: 112


View Profile
January 27, 2012, 04:34:21 AM
 #260

How do you import the data into ninjatrader?  I might be retarded, but I cannot figure it out.  

I know the free version is a little neutered...

No, the free version is the same as the full version.  They are identical except you can't place live trades with the free version.

I'll just go ahead and post this.  Here's the data to import.

http://tinyurl.com/7nhd9ql

-Create an instrument in the instrument manager called BTC
--Make it a stock, set the exchange to NASDAQ
--Add it to the default list by clicking the arrow key to push it over there

-Add historical data through the historical data manager
--Import the file from the zip

-Backtest is under strategy analyzer
-When charting, be sure to chart daily data

The dates are meaningless, I just copied some from another currency just to get a "quick fix".

Awesome.  PM sent.

How do you typically gather new data?
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 [13] 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 »
  Print  
 
Jump to:  

Sponsored by , a Bitcoin-accepting VPN.
Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!