(Closed) Butter Bot!: Premier Bitstamp, and BTC-E EMA Trading Platform (Closed)

[fible1]

My bots all stopped getting data from BitStamp @ 0700 this morning (running on multiple machines, so each exchange has tuned settings).

Anybody else seeing this?

Sean,
  We will track this down and get back to you ASAP, it may have to do with some back end work we are doing.

Please drop me a message if you don't hear from me by tomorrow AM.

Pablo.

My BitStamp has apparently come back online...thank you sir!!

*Haven't really seen any sort of profitability from my limited use of the trial. Any plans on supporting other currencies as well as adding additional exchanges (mcxnow of particular interest to me atm)?

Hi Nwfella ,
   We have plans to add other exchanges in the mid term. Altcoins on BTC-E will most likely be a part of V3.

Right now we are devoting all our resources to developing a hosting service for the bot; once that's done we will push out V3 as soon as possible.

Please let me know if you have any questions .

Pablo.

[Grinny]

Sorry, if this has been asked before, but is it possible to run two instances of butterbot with different settings using the same key?
Or to be more precise: Will it be possible to run two instances in v3, using one bot for BTC and one for LTC (with the same key)?

[fible1]

Sorry, if this has been asked before, but is it possible to run two instances of butterbot with different settings using the same key?
Or to be more precise: Will it be possible to run two instances in v3, using one bot for BTC and one for LTC (with the same key)?

Hi ,

You can run two instances of the bot with your same license key on either the same system or different computers. Running on your same system is a tad involved but on different computers it is as easy as install and run, You must use different exchange API keys for each bot however-

Please let me know if you need anything at all .

Pablo.

[seanrarey]

Sorry, if this has been asked before, but is it possible to run two instances of butterbot with different settings using the same key?
Or to be more precise: Will it be possible to run two instances in v3, using one bot for BTC and one for LTC (with the same key)?

Hi ,

You can run two instances of the bot with your same license key on either the same system or different computers. Running on your same system is a tad involved but on different computers it is as easy as install and run, You must use different exchange API keys for each bot however-

Please let me know if you need anything at all .

Pablo.

Hi Pablo.

I would be interested to know how to run multiple instances on the same machine.  I am running multiple instances on multiple machines right now, as each it tuned for the specific exchange.

[fible1]

Sorry, if this has been asked before, but is it possible to run two instances of butterbot with different settings using the same key?
Or to be more precise: Will it be possible to run two instances in v3, using one bot for BTC and one for LTC (with the same key)?

Hi ,

You can run two instances of the bot with your same license key on either the same system or different computers. Running on your same system is a tad involved but on different computers it is as easy as install and run, You must use different exchange API keys for each bot however-

Please let me know if you need anything at all .

Pablo.

Hi Pablo.

I would be interested to know how to run multiple instances on the same machine.  I am running multiple instances on multiple machines right now, as each it tuned for the specific exchange.

You can run it by running multiple profiles:

It's pretty easy, here's a description on how to set it up: http://cybernetnews.com/chrome-multiple-profiles/

Please let me know if you need anything at all .

Pablo.

[seanrarey]

Sorry, if this has been asked before, but is it possible to run two instances of butterbot with different settings using the same key?
Or to be more precise: Will it be possible to run two instances in v3, using one bot for BTC and one for LTC (with the same key)?

Hi ,

You can run two instances of the bot with your same license key on either the same system or different computers. Running on your same system is a tad involved but on different computers it is as easy as install and run, You must use different exchange API keys for each bot however-

Please let me know if you need anything at all .

Pablo.

Hi Pablo.

I would be interested to know how to run multiple instances on the same machine.  I am running multiple instances on multiple machines right now, as each it tuned for the specific exchange.

You can run it by running multiple profiles:

It's pretty easy, here's a description on how to set it up: http://cybernetnews.com/chrome-multiple-profiles/

Please let me know if you need anything at all .

Pablo.

Thanks Pablo, perfect!

[PuertoLibre]

Really you should not be using the BOT if you do not understand that it is not 'willy nilly', the whole point is that over time you will see an increase over a 'normal' human trader, but that may include  taking it in the ass once in a while....
There is a 'run test' function and whilst it is a bit basic, it will give you an idea.

For example the recommended settings would have made a significant loss on Mtgox over the past few days.

It gave me enough of an idea to understand the  functionality is not suitable for trading in such a market at the moment, as it stands the BOT is not quite polished enough, and yep I know it makes money, but is also takes rather more 'hits' than it should.

Personally I'm of the mind that the product is in need of significantly more 'safety mechanisms', but I also do not think that a simple 'stop loss' is the answer, specifically because it can be exploited by other traders and in some cases may well be a significant contributing factor to  a 'flash crash'.

That said and despite the BOT going mad and clearing out all my FIAT, I have still managed to manually trade the account to a significantly better position that the BOT over the last two days.
There are some rather 'shitty' patterns emerging from gox at the moment indicating some very rough BOT trading algorithms.

I agree with your assessment.

It takes more hits than it really should. Which like it or not, deducts from future gains.

Right now I am almost finished writing my own arbitrage bot and testing it through it's paces. I too would like Butter Bot to add certain real time simulation features. I think a simulation feature is better than turning on the bot and letting it do trading. This will help you get an idea of the decisions the bot makes in different situations without risking BTC or USD.

------------------------

I would love it if they made a "secure-bot" to "get you out" of the market in the event of a flash crash. Just that feature alone is very useful if you don't intend to do actual EMA based trading. If you do want a bot to watch the current market conditions on a continual basis butter bot would be a good start if you understand the settings.

I think they should package this separate from the normal butter bot set of features.

-------------------------

There are also other types of already coded software based on ADX that don't take anywhere near "the number of hits" that butter bot takes. (not butter bots fault as ADX is just a better indicator IMO)

There is one publicly available but someone decided the neuter the public code so that it didn't do anything other than prevent losses. I back tested the public version and while it wasn't written to make a profit, it back tested to the beginning of the year and retained 99% value of the currency traded. Had someone with experience redone it to actually "make a profit" it would have no doubt been better than the EMA based bots.

Future version of butter bot will come with different strategies so that people can choose whatever type they prefer.

[fible1]

Really you should not be using the BOT if you do not understand that it is not 'willy nilly', the whole point is that over time you will see an increase over a 'normal' human trader, but that may include  taking it in the ass once in a while....
There is a 'run test' function and whilst it is a bit basic, it will give you an idea.

For example the recommended settings would have made a significant loss on Mtgox over the past few days.

It gave me enough of an idea to understand the  functionality is not suitable for trading in such a market at the moment, as it stands the BOT is not quite polished enough, and yep I know it makes money, but is also takes rather more 'hits' than it should.

Personally I'm of the mind that the product is in need of significantly more 'safety mechanisms', but I also do not think that a simple 'stop loss' is the answer, specifically because it can be exploited by other traders and in some cases may well be a significant contributing factor to  a 'flash crash'.

That said and despite the BOT going mad and clearing out all my FIAT, I have still managed to manually trade the account to a significantly better position that the BOT over the last two days.
There are some rather 'shitty' patterns emerging from gox at the moment indicating some very rough BOT trading algorithms.

I agree with your assessment.

It takes more hits than it really should. Which like it or not, deducts from future gains.

Right now I am almost finished writing my own arbitrage bot and testing it through it's paces. I too would like Butter Bot to add certain real time simulation features. I think a simulation feature is better than turning on the bot and letting it do trading. This will help you get an idea of the decisions the bot makes in different situations without risking BTC or USD.

------------------------

I would love it if they made a "secure-bot" to "get you out" of the market in the event of a flash crash. Just that feature alone is very useful if you don't intend to do actual EMA based trading. If you do want a bot to watch the current market conditions on a continual basis butter bot would be a good start if you understand the settings.

I think they should package this separate from the normal butter bot set of features.

-------------------------

There are also other types of already coded software based on ADX that don't take anywhere near "the number of hits" that butter bot takes. (not butter bots fault as ADX is just a better indicator IMO)

There is one publicly available but someone decided the neuter the public code so that it didn't do anything other than prevent losses. I back tested the public version and while it wasn't written to make a profit, it back tested to the beginning of the year and retained 99% value of the currency traded. Had someone with experience redone it to actually "make a profit" it would have no doubt been better than the EMA based bots.

Future version of butter bot will come with different strategies so that people can choose whatever type they prefer.

Hey ,
  All good ideas!

V4 should come with alternate strategy modules, and we are looking at releasing a new and improved trading engine for Butter in the near future to make it even more profitable. Stay tuned.

Pablo.

[luicon]

i guess the bot can win money just if the price crash, right?
i notice the price went up very quickly those days,

[fible1]

i guess the bot can win money just if the price crash, right?
i notice the price went up very quickly those days,

Hi Luicon ,
   Price is not really the main factor in bot profitability but rather volatility is more important. Volatile trends manifesting themselves will increase bot profitability.

Please let me know if you have any other questions.

Pablo.

[Grinny]

Hey Pablo,

I think the bot's Profit Analysis shown in the backtest is a little bit misleading, because it ignores the actual value of your BTC after the chosen period is over IF you didn't sell them.
Or in short: The bot doesn't calculate the profit for buy&hold right.

Example:

1 week / 100 USD (EMA/threshold, etc. doesn't matter)

       Time               Type    Price      Value    Profit
16.10.2013 21:45:00   buy   141.260    $99.50   -0.50%

As shown in the example, the bot bought approx. 0.7 BTC @ $141.26 one week ago for a total of $99.50.
Today, the BTC price is at $198, so our 0.7 BTC would be worth $138.

This is not a loss of 0.5% as stated by the bot, it's a profit of 38.7% within one week!

This issue appears very often and kills backtesting over short periods. So I suggest to add a hypothetical sell at the end of all calculations.

Best,
Grinny

[fible1]

Hey Pablo,

I think the bot's Profit Analysis shown in the backtest is a little bit misleading, because it ignores the actual value of your BTC after the chosen period is over IF you didn't sell them.
Or in short: The bot doesn't calculate the profit for buy&hold right.

Example:

1 week / 100 USD (EMA/threshold, etc. doesn't matter)

       Time               Type    Price      Value    Profit
16.10.2013 21:45:00   buy   141.260    $99.50   -0.50%

As shown in the example, the bot bought approx. 0.7 BTC @ $141.26 one week ago for a total of $99.50.
Today, the BTC price is at $198, so our 0.7 BTC would be worth $138.

This is not a loss of 0.5% as stated by the bot, it's a profit of 38.7% within one week!

This issue appears very often and kills backtesting over short periods. So I suggest to add a hypothetical sell at the end of all calculations.

Best,
Grinny

Hey Grinny,
   I think I see your point. Basically the back testing takes into account the last trade, since the last action was a "buy" and that buy was at a small loss with respect to the previous trade that's what you see on your screen. The profit/loss analysis is made with respect to the last trade, hence, when the bot sells from this bull run you will see the massive profit you are talking about.

A hypothetical sell at spot is an interesting idea, but it sort of defeats the purpose of back testing as a long term indicator (as said ad nauseaum EMA is a long term strategy and best back tested over the long term) and may encourage people to trade by hand given the at hand profit.

Someone was discussing a few days ago how they would like to manually trade because they would make a substantial amount of money by selling, BTC was around 190 at this point if memory serves. If back testing had included the hypothetical spot profit they may well have missed out on the current 207 price.

I'm not shooting down the idea, I think it's worth looking into and I will float it internally, I'm just thinking of possible drawbacks of including unexecuted trades in the back testing.

Keep them ideas coming .

Pablo.

[Grinny]

Thanks for the quick reply.
I see your point with EMA as a long term analysis as well, so maybe the spot sell isn't the way we get this fixed.

Maybe the simple way is the way to go.

There is this row above the table which shows "Min. Profit / Max. Profit", right?
What about adding a line: "Final Balance: X BTC (X USD)" in our case: 0.7 btc (138 USD) | cryptotrader does it that way btw.
In general: The profit is shown twice, above the table and within the table, so maybe some space can be saved here...?

[fible1]

Thanks for the quick reply.
I see your point with EMA as a long term analysis as well, so maybe the spot sell isn't the way we get this fixed.

Maybe the simple way is the way to go.

There is this row above the table which shows "Min. Profit / Max. Profit", right?
What about adding a line: "Final Balance: X BTC (X USD)" in our case: 0.7 btc (138 USD) | cryptotrader does it that way btw.
In general: The profit is shown twice, above the table and within the table, so maybe some space can be saved here...?

Grinny,
 This is a great suggestion, I will add it to the board and see what the developers think.

Let me know if you have any other ideas .

Pablo.

[Pangia]

Hi Pablo,

Just an idea here. Please shut me down if it's not a worthwhile venture.

Would it be possible to create a bot that could perform arbitrage trades WITHIN an exchange such as BTC-E?

My reason for asking is that BTC-E has various Alt coins as well USD & Russian RUB.  Would it be possible for a bot to detect an arbitrage opportunity between the Alt coins as well the currencies (USD/RUB)?

Thanks Sir.

[fible1]

Hi Pablo,

Just an idea here. Please shut me down if it's not a worthwhile venture.

Would it be possible to create a bot that could perform arbitrage trades WITHIN an exchange such as BTC-E?

My reason for asking is that BTC-E has various Alt coins as well USD & Russian RUB.  Would it be possible for a bot to detect an arbitrage opportunity between the Alt coins as well the currencies (USD/RUB)?

Thanks Sir.

Hi Pangia ,
   I think this is a great idea and we had an internal discussion about this. There are certain cons to doing this type of bot that I won't go into here to save time, but basically we decided to build another type of bot (secret for now) that although it is not specifically designed for arbitrage, it should eventually let you do this type of thing. There will not be a specifically built arbitrage bot however, for the time being.

Please let me know if I may be of service in any way .

Pablo.

[Pangia]

Hi Pablo,

Just an idea here. Please shut me down if it's not a worthwhile venture.

Would it be possible to create a bot that could perform arbitrage trades WITHIN an exchange such as BTC-E?

My reason for asking is that BTC-E has various Alt coins as well USD & Russian RUB.  Would it be possible for a bot to detect an arbitrage opportunity between the Alt coins as well the currencies (USD/RUB)?

Thanks Sir.

Hi Pangia ,
   I think this is a great idea and we had an internal discussion about this. There are certain cons to doing this type of bot that I won't go into here to save time, but basically we decided to build another type of bot (secret for now) that although it is not specifically designed for arbitrage, it should eventually let you do this type of thing. There will not be a specifically built arbitrage bot however, for the time being.

Please let me know if I may be of service in any way .

Pablo.

AWESOME!!!!

And the "Secret for Now" has me honestly excited.

Thanks for all the GREAT work that you and your team do.

[seanrarey]

This Butter community is damn near as exciting as BTC itself. 

[testerx]

Is BTCChina planned to be added?  Having no fees definitely opens up more frequent trading strategies...

[darkry]

Is anyone else experiencing problem in fetching data from BTC-e?
My bot is unable to do so in the last hours (MtGox and Bitstamp still work correctly).