I think it would be a good thing if the devs were to make a training video that detailed each & every setting & how they affect trading. Explaining like I'm 5, lol.
They could teach things like why & when you'd want to have negative values set, , etc.
Not a bad idea... They should totally make a techsmith based type of summary. http://www.techsmith.com/camtasia.html=======[Discount Code for Butter (10% off)]========= https://butter-bot.com/?r=f5a17f5bUse my referal code cause' im sooo cute!    You'll get an extra month free, it's all on me! |
|
|
|
It would have to keep track constantly of how much of its working capital is from seed fund, & how much is earned, so that when you want to add more it knows how much its allowed to use.
 Like so? |
|
|
|
All I can think of now is for there to be a way for it to seperate seed funds from capital gains. I'll chime in with more ideas as they come to me.
I mentioned something like that to Pablo in private a few days ago. A sort of Butter BookKeeper. (Simple Accounting) I think if Pablo's team could work out a way to employ entirely different strategies within one account, then the losses in one strategy can be offset by the win's of another. A crop of the best, settings, working side by side to make some profit. It's an alternative to a "one strategy fits all". I don't know what came out of it or if he found anything useful in the ideas I presented. But hopefully it inspires some future version. (Considering they are moving to a hosted solution, it just puts a little more money in their pocket if they can sell a wider array of techniques.) I just hope they consider that most accounts on exchanges need to be KYC verified so it makes it impractical to run multiple accounts with each one having different strategies. With a little simple book keeping you could probably eliminate that problem by keeping numerical tabs on trades being conducted. Even if there is one pool of "seed" funds, it can conceivably be sub-divided. Each subdivision of seed funds inside one account can then be accounted for. Since any USD from a single account pool is perfectly interchangeable with any other USD from anothe rinstance there wouldn't be a problem as long as the book keeper keeps numerical tabs so it doesn't overdraw seed funds between instances. Same is true for BTC within one account. The only thing distinguishing one BTC from another is just the book keeper. =================== If you analyze the market carefully you can pick out people using bots to trade with and these bots run on fixed settings. (Heck, you can even use the bots trades to your advantage! (I made a few K today and yesterday just identifying them) If butter could be retooled for various parallel strategies (even if it is EMA based) it might prove to be a bit more profitable. I am currently trying it out myself to see how it handles. |
|
|
|
@ Pablo I took notice yesterday when trying to build a B-Bot simulator in excel, that your bot actually takes 3 "update cycles" to actually execute an action. I have to ask why? Shouldn't it execute (ideally) when the second update cycle takes place? Why wait until the third cycle? I used to think two things about B-Bot which were dispelled when trying to build my own version in Excel. A) The idea that B-bot will act immediately if you set the first pair of thresholds to 0. Not the case. It actually still uses it, and there doesn't seem to be a way to disregard the first pair of threshold values. (Not a problem, but it is something worth knowing if you are using very long update periods (i.e. 1 hour 2 hour etc..) B) This means that even if B-Bot noticed a correction in the market at 1 hour intervals it won't actually execute anything until 3 update periods. (3 hours) At 2 hours it should take 6 hours and this is only 1 way. So between a buy and sell order, if you have your settings at 1 hour update intervals, then B-bot has a Buy and Sell period of about 6 hours. At 2 hours it should be double that. That was kind of alarming to figure that out. It seems like a pretty long turn around period... (1 hour, Buy threshold 1 check, YES/NO) (1 hour, Buy threshold 2 check, YES/NO) (1 hour, Buy Execute, YES) (1 hour, Sell threshold 1 check, YES/NO) (1 hour, Sell threshold 2 check, YES/NO) (1 hour, Sell Execute, YES) =======[Discount Code for Butter (10% off)]========= https://butter-bot.com/?r=f5a17f5bUse my referal code cause' im sooo cute! You'll get an extra month free, it's all on me! Hey  , That's not quite accurate: Butter will execute immediately when threshold 1 and threshold 2 are met (so after 2 hours in your 1 hour interval example). As for disabling the first threshold (which you set to 0) you need to set it to a negative value, so that it will always be lower than the actual difference. That should void the first threshold for all intents and purposes. I hope that helps . You are definitely right, I accidentally mixed two different data sets with different EMA values when comparing how butter does things. One data set reads one way, the other has slightly different values. (Thank God, will save me a ton of time!) |
|
|
|
@ Pablo I took notice yesterday when trying to build a B-Bot simulator in excel, that your bot actually takes 3 "update cycles" to actually execute an action. I have to ask why? Shouldn't it execute (ideally) when the second update cycle takes place? Why wait until the third cycle? I used to think two things about B-Bot which were dispelled when trying to build my own version in Excel. A) The idea that B-bot will act immediately if you set the first pair of thresholds to 0. Not the case. It actually still uses it, and there doesn't seem to be a way to disregard the first pair of threshold values. (Not a problem, but it is something worth knowing if you are using very long update periods (i.e. 1 hour 2 hour etc..) B) This means that even if B-Bot noticed a correction in the market at 1 hour intervals it won't actually execute anything until 3 update periods. (3 hours) At 2 hours it should take 6 hours and this is only 1 way. So between a buy and sell order, if you have your settings at 1 hour update intervals, then B-bot has a Buy and Sell period of about 6 hours. At 2 hours it should be double that. That was kind of alarming to figure that out. It seems like a pretty long turn around period... (1 hour, Buy threshold 1 check, YES/NO) (1 hour, Buy threshold 2 check, YES/NO) (1 hour, Buy Execute, YES) (1 hour, Sell threshold 1 check, YES/NO) (1 hour, Sell threshold 2 check, YES/NO) (1 hour, Sell Execute, YES) =======[Discount Code for Butter (10% off)]========= https://butter-bot.com/?r=f5a17f5bUse my referal code cause' im sooo cute! You'll get an extra month free, it's all on me! |
|
|
|
Hi all, I'm a new user to Butter Bot, I've been running the demo for the last few days and reading the various conceptual arguments about bot profitability vs B&H etc... I'm now in the bizarre position of actually hoping for a BTC price crash so I can buy back low, assuming that the bot gets me out in time! All in all, I think the bot is for me and I'm looking forward to being able to use it with LTC/USD also. Just wanted to pipe in though that I too would love to see a 'don't sell at a loss' feature. It seems to me that if you're able to specify that the bot shouldn't sell below it's last purchase price, then you're basically on 'B&H plus extra profits during market volatility'. Sure, the price could tank. If you're lucky, the EMA settings you have got you out of BTC in time and you can buy back at a lower price and profit as the price climbs back up - but if not, you've basically switched to B&H on those coins until the market recovers. Since people are arguing over whether B&H is a better strategy, and the suggestion of a split between the two seems sensible, this looks like an ideal scenario. We're assuming of course that the overall trend is up. But if you're B&H, then you're assuming that already... and if the price never goes back up, well you'd have lost with B&H anyway! Of course, if the price has really tanked and takes months to recover, you'll be losing out on all the profit the bot would be making on a lower priced, but still volatile market over that time - but if the feature's available then you ticks yer box and you takes yer chances.  If the feature is planned, I'd be interested to know how it'll work - If I find myself locked in to BTC at a higher price because the bot didn't spot the downtrend in time to get me out at a profit, my thinking is that I'll probably want to pump some more USD into the account and have the bot pick up some more BTC when it spots the next uptrend. I'd want the bot to remember that X amount of BTC can't be sold below $Y, but the rest is available for it to trade with. From a development point of view, I can see how that would get complicated. You'd need to track the amount of BTC sold and at what price. Presumably that means the trades would have to be recorded to a database and have some kind of open / close position. A trade of X BTC opened at $Y. Current price below $Y so trades can't be used to 'close' X BTC until Y > the price of BTC at the point of the X trade. Trades instead 'open' new records and close themselves until the $Y is reached. The point being you wouldn't be able to just look at the last trade....... I'm digressing - just wanted to say hi. Hi Simiane , We're really on the fence about this one; we have been back testing the "do not sell for loss" scenario and it has pretty bad results compared to EMA. Then again, if you guys want it we should accomodate your requests; but I kindda feel like a lot of people are gonna loose some cash if we put that feature into production. Let me percolate on this for a bit, I'll get back to you . Pablo. I have actually taken about two days of raw feed data and have tried to recreate B-Bot in excel. It is surprisingly very simple from what I can tell. I am still working on simulating the exact behavior of B-Bot but it is coming out very close to the same. I was attempting to "bolt-on" the "do not sell on a loss". I am still having some issues with the fee deduction. |
|
|
|
Right now Im using butterbot demo on this settings and it doing strange things
Sell for 400
Buy for 405
Sell for 406
Buy for 419
the dip last night got the better of the bot on a lot of settings. So is that normal? I am talking about last (3rd) settings It is normal for butter to do that. =======[Discount Code for Butter (10% off)]========= https://butter-bot.com/?r=f5a17f5bUse my referal code cause' im sooo cute! You'll get an extra month free, it's all on me! |
|
|
|
I was going to build trading packages in VM's for my little cousin's for their college funds, but after talking to their parents, who would rather see me spend that money on myself, I won't be doing so anymore. Oh well. Its that thought that counts right?
It's a very nice idea. |
|
|
|
Hey lets break down what this bots settings actually mean.  The way I read the settings, it says to me:  |
|
|
|
After losing some money daytrading I would suggest to activate the bot only when we are in bubble mode with high volatility. In normal period I would stick to buy and hold (long term trend is clearly up).
Anyway it would be great to have a BTC profit calculator and a testing time range selector in order to test such hypothesis.
Agreed, I would love to backtest several schemes against a down market trend and not just against a bubble like we have been for the last 2 months. |
|
|
|
 Or  =======[Discount Code for Butter (10% off)]========= https://butter-bot.com/?r=f5a17f5bUse my referal code cause' im sooo cute! You'll get an extra month free, it's all on me! |
|
|
|
@ Pablo
I think I might have found a strange inconsistency.
I was installing B-Bot on a ESXi server, on a virtual desktop to test out a few settings. I set B-Bot to standard settings (2 hours, 0.25 buy/sell 10/21 ema settings) and refreshed the popup.html page and I noticed that the desktop time sync was off by about 3 or 4 hours.
When I went to refresh B-Bots status page I noticed it was reading in sync with the wrong hour.
I corrected the clock on the virtual desktop and only then did B-Bot actually use the current stats up to the current hour.
I know this is a small issue but if someone has their clock off by a significant portion of time they might be trading with very old data. You might want to look into it and see if B-Bot is synchronized by the clock on the users computer rather than the api/feed servers it is getting data from.
As soon as I corrected the time, B-Bot started pulling several extra hours of data that weren't there before.
(By the way, are there any current Beta Tester spots available?)
Thank you PuertoLibre, on it. About Beta Tester Spots: If you want in the next Beta I will guarantee you a spot, you are one of the top bug spotters in the community, if not number one . Pablo. I am writing the email right now. I do want in to the next Beta. I am kinda surprised that I am supposedly number one at bug reports.... I only try to turn butter into margarine through plenty of abuse...haha |
|
|
|
Thanks for the reply Pablo...
I got another Question...
Is there an ETA for alt-coins support? Like LTC and etc... ?
Hey , We will push that out ASAP. Right now the immediate priority schedule looks like this: 1. BTCChina 2. Hosting - Strategy Improvements 3. Version 3 (Altcoins) The bottleneck right now is Hosting which is consuming more of our development resources than we thought it would. We are looking into ways to speed this up; rest assured we want to get out as many features as we can as fast as we can safely build them and test them . Pablo. When BTCChina is added...there is gonna be an epic flood of sales. http://www.coindesk.com/bitcoin-exchange-btc-china-experiences-time-price-high/“This is similar to what happened in April, where we saw trading well above $265 for over 12 hours. This time, it seems to be lasting longer. We hope that, this time, there won’t be such a severe price correction as in April,” he said. Lee recently told Forbes that BTC China’s recent elimination of trading fees helped to boost interest in bitcoin. “It was a strategic move to increase trading in China and draw people to our site,” he said, and it appears to have worked." |
|
|
|
|
@ Pablo
Any chance we can buy extra licenses at a significant discount? (5 - 10?)
I know you are going to be offering virtualized butter-bots in the near future...so I assume our current [or concurrent] instances of butter won't be there in the near future...
Or any chance we can convert subscription months into extra butter bot licenses?
I have devised a strategy that requires more than one instance...and I want to make sure this won't flop over the long term.
|
|
|
|
|
I don't get it, if you want to seriously increase your BTC stash as compared to your USD stash...then simply buy BTC when it is low. Over time you'll make plenty of BTC and increase your standing. With each subsequent rise, due to extra BTC your USD also increases.
Free Point:
Just remember, that you shouldn't buy in (manually) as you approach round numbers. 10's and 5's, and definitely not on quarters of a hundred. (x00, x75, x50, x25)
People usually have bot settings that rely on these points to start a sell. So it is sort of easy if you know these rules (and have ample chart) to make the call to buy near the bottom of a drop.
Another free point:
People (mostly the "whale" buyer and seller) will try to offload their earnings near what they think is the top. As long as you recognize this, you won't be enticed to buy in by a few thousand BTC changing hands near the top. Only the people who don't know better buy in near the end of a significant run. Don't be a dummy!
|
|
|
|
Further...
Why are there FCC, EC and ETL marks on the Monarch prototype when they have not been submitted to those agencies for testing?
Why is this a concern? I would think they would pre-print the FCC, etc. logos on the board to save time. And even if that's not allowed, this is just a prototype board. It's not like they're selling them to the public yet. Voted the most retarded comment of today... |
|
|
|
|
@ Pablo
Is the backtesting feature hardcoded to execute a sell at -16.50% ?
I am noticing a pattern that the bot uses on down trends. If it is the case, is there any possibility we could choose our own number during backtesting rather than a rigid one?
|
|
|
|
600GH makes 0.59 BTC per day right now. By Feb these things will be so worthless.
Wow, are you kidding? (Seriously, has it dropped that much already?) Wow, what planet have you been hiding on? I stopped paying attention to the BitCoin tragedy/drama story. Geez! Haha, if true, wow....just wow. |
|
|
|
One must subtract the price appreciation of bitcoin itself from the numbers that back testing gives to get a more realistic view of potential profits. I agree that it's a very close call versus buy and hold.
Friends, We just released a fix via Mass Mail for the email issue labeled 2.4.1 I want to personally apologize for the inconveniences these past two days, we have taken strong measures to prevent this from recurring in the future. Please do not hesitate to let me know if anyone needs anything at all . Pablo. Oops, I just noticed the bug creeping up again and I sent you a PM (v2.4) with the console logs and relevant settings. No email was dispatched during the buy event. |
|
|
|
I've been trying Butterbot, and although it's an interesting concept, the truth is that you are clearly loosing money.
I tried a great amount of different settings (from the ones suggested by the creators to others) and in every case, you cannot find any settings that simply beat leaving the BTCs alone and capitalizing. You are always loosing money. If the creators of ButterBot put an option to see the profit in BTC instead of USD you could see that the number of BTCs using this algorithm - in every case - it's reducing.
Put the monthly amount you pay for the software on top of that and you are clearly loosing quite a big sum of money by using it.
I actually agree with you but I also disagree with you. I was discussing this very point with my family last night. If you think about the market in terms of a purely vertical increase. Then the market probably moves into the positive 3 or 4 times what it does when using an ema bot. For example, if you counted each price rise over a single day and added them all up commulatively you'd see it probably moved 200 dollars in equivalent value even if the price only increased $20 on that day. (as well as decreasing in value by a similar figure) An ema bot might require a sacrifice of 30% on an upward trend to be sacrificed in order to detect a trend. The same is true on the down turn. Lets say a (plucked out of thin air) figure of 60% potential profit is lost by waiting until a trend actually exists. About 30% is your only real profit. So even if the market only went up 20$ but vertically stacked clocked in at 200$ vertical cummulative climb at the days end. From the start of the day till the end, your ema bot might have actually made you: 60% or 120$ is lost in detecting a trend. While 80$ might be potential profit. ========================== The only difference with buy and hold strategies is that you have to assume the value will only go up. It may go down and you go bust. (While using a bot protects you from down turns, holding does not, unless you exit the market manually) Even if the commulative dollar amount for the total rise in a 24 hour period is 200$, in a buy and hold strategy it might only net you 20$ at the end of the day. It is technically static and doesn't take any opportunities of market corrections and repeated rises in those 24 hours that eventually end in a 20$ gain at the end of the day.So when someone argues that a straight forward buy and hold strategy is best, they are basically trying to debate that the bot will mess up enough [in detecting trends] that overall the (bot + trend) strategy will make enough bad calls that it does worse than a simple buy and hold. It's like saying: "So what if your bot made 80$ today V.S. my buy and hold strategy and profit of 20$? Your bot will eventually make enough "bad calls" that your 80$ will actually turn out to be 10$." (which is true, as most bots lose just as much as they gain. Pablo mentioned that 70% will be losing trades while 30% will be gaining with this ema bot.) The only perfect bot is the one that makes the right decision every time and has "zero loss" for trend detection. (it doesn't exist) A perfect bot, would earn 200$ on a single day on a 20$ average rise. Not even human traders are that good. No where even close or that lucky. Think about how long it takes butter (or how many dollars each way) it loses as it waits to make sure a trend is building. Butter loses quite a bit getting you out of a crash dive as the ema signals take some time after the drop. Quite a bit is lost getting you into an upward trend as well. But a buy and hold strategy only nets you the average rise of that day. =========================== It should be a toss up as to whom will do better. If the market went from 100$ to 400$ over a week, then the buy and hold speculator will only ever make (at best) 300$ with a simple buy and hold. If the bot user using ema has extremely efficient settings, then they might make well over 300$. Or if they have inappropriate settings they might make a hell of lot less than 300$. The bot user also has to contend with the notion that not all upward trends will be clear enough when using an ema system. So often you won't even make the 300$. If anyone tallies up all the commulative price rises over the last month you'd probably find it has gone up more than 10,000$ vertically. Though most of our bots are at best doing a fraction of 300$ in profits. We should all share our settings and work out what is the most efficient path (not the settings themselves) to figuring out what settings are optimum in different scenarios. We need to pry open EMA for example and really understand deep down what makes it tick. Why certain setting in one particular week give spectacular returns and why it doesn't work as well in another week with a different scenario. At the moment we are all mostly using the "Jack of all Trades" strategy. Which is finding one setting that does "okay" in most scenarios. I think it would be "profitable" to find settings that detect down trends quickly and adjust our settings as necessary throughout the day. Like changing settings based on the market conditions. (I can hear Pablo just telling us not to do that, that a jack of all trades is best) |
|
|
|
|
Show Posts
