Haasonline Simple Trade Bot For BTCe and Bitstamp[Main topic]

[silverston]

kalross
it was reply, yes. I have same issue.
 I dont like unprofitable trades  

[1715265634]

1715265634

[1715265634]

1715265634

[1715265634]

1715265634

[kalross]

kalross
it was reply, yes. I have same issue.
 I dont like unprofitable trades  

Oh ok.  Looks like a bug then...easily fixed I would assume.

[krach]

"So...set a well benchmarked set of indicators including an MACD-MA Short 11 Long 24 for NMC/BTC.  Update timer at 600s.  Ticked Reset Last Buy and Sell.

Last sell price 0.00  Last Buy Price 0.00

Had sold some BTC to buy NMC/USD and PPC/USD expecting a strong upward trend (I was right on both counts).

Activated Bot and it sold NMC at 0.00801 at Interval 3 with two (of two) indicators on Yellow.  Was unable to buy back in quick enough as immediately afterwards, NMC/BTC went to 0.0085 then 0.00886.  Kinda bad luck combined with mental bot logic (or 'features' as some on here call them).

Missed out on a 10% profit.  Baaaa!

I take it the bot should not trade until the indicators have collected sufficient data according to their settings?

Stephan:  Can you reproduce/fix this?

K
"

-Was unable to buy back in quick enough-
Hi,
You have it set on 600 seconds so the bot will wait at least 600 seconds before the next action, I think.  I think Stephan has said many times, the bot is not faster than manual trading, the main advantage is that the bot does not have any emotions. If you want quick trades put it on 20 seconds, imo.

[kalross]

"So...set a well benchmarked set of indicators including an MACD-MA Short 11 Long 24 for NMC/BTC.  Update timer at 600s.  Ticked Reset Last Buy and Sell.

Last sell price 0.00  Last Buy Price 0.00

Had sold some BTC to buy NMC/USD and PPC/USD expecting a strong upward trend (I was right on both counts).

Activated Bot and it sold NMC at 0.00801 at Interval 3 with two (of two) indicators on Yellow.  Was unable to buy back in quick enough as immediately afterwards, NMC/BTC went to 0.0085 then 0.00886.  Kinda bad luck combined with mental bot logic (or 'features' as some on here call them).

Missed out on a 10% profit.  Baaaa!

I take it the bot should not trade until the indicators have collected sufficient data according to their settings?

Stephan:  Can you reproduce/fix this?

K
"

-Was unable to buy back in quick enough-
Hi,
You have it set on 600 seconds so the bot will wait at least 600 seconds before the next action, I think.  I think Stephan has said many times, the bot is not faster than manual trading, the main advantage is that the bot does not have any emotions. If you want quick trades put it on 20 seconds, imo.

Thanks for your contribution, but not sure you have fully understood the issue I (and one other) are reporting.  I've been studying the bot for a few months, so not unfamiliar with its function.  The bot SHOULD NOT have sold whilst one of the indicators (let alone both) were not sending a BUY/SELL signal.  However, it did.  If this is indeed a bug...somebody might suffer a great loss if large amounts are committed to trading.  Luckily...I am a relatively 'poor' trader, so damage was quantitatively limited, though not really as a proportion of my total coin.  Life goes on however.

If my word is not sufficient, then I can show this with screen capture as soon as I find time to host the image.  An issue tracker would be helpful for a piece of development software like STB.

I couldn't buy back in manually around the same price (using BTC-E web ui) because the price jumped immediately after the bot sold (which is why I set the bot up to begin trading there) and didn't return which was a particularly annoying effect of the issue I am reporting (supported by one other user)...but not directly related to the bot.  I had to adapt a well thought out strategy to limit my losses...the outcome of which was not losing 10% of my investment with a panic buy...but by sacrificing the profit I would have made.

Stephan doesn't recommend the 20s Update Timer and has removed it in the coming release, so it would not be wise to recommend its use to other traders.  Longer intervals work better as has been reported.  The 600s interval I used was benchmarked over several days on PPC and worked well without issues.  It is however experimental, but has not simulated unprofitable results.

Yes the bot doesn't have any emotions and neither does it know when its coding has errors.  Only the developer can investigate this.

Cheers.

K

[krach]

sure, but if speed is the issue then the only way I see to make it faster is to lower your time, I might be wrong there though. There will be some changes that happen so fast that it will be hard for a bot or for a human to get the buy or sell order in fast enough.
Could be an issue, what did it look like on the bot graph? Did the colour reflect the same indication as in the bot signal graph?
I had ok results with 20 seconds but yea everyone should test it out for themselves.

I get "attempted to divide by zero" not sure what I did to get this notification in the debug log

[Soak]

The 600s interval I used was benchmarked over several days on PPC and worked well without issues.  It is however experimental, but has not simulated unprofitable results.

Is it possible for you to share your strategy or do you keep it secret?

[kalross]

sure, but if speed is the issue then the only way I see to make it faster is to lower your time, I might be wrong there though. There will be some changes that happen so fast that it will be hard for a bot or for a human to get the buy or sell order in fast enough.
Could be an issue, what did it look like on the bot graph? Did the colour reflect the same indication as in the bot signal graph?
I had ok results with 20 seconds but yea everyone should test it out for themselves.

I get "attempted to divide by zero" not sure what I did to get this notification in the debug log

I don't need it faster.  It Sold on interval 3 before the indicators had received all the data to generate any signal at all.  The bot signal was at the SELL position from interval 1-3 and the MACD didn't chart until interval 24 (as the Long was set to 24) which was correct and as programmed.  The RSI interval didn't chart until interval 13 as programmed.  The bot indicators shouldn't generate a signal until all the required ticker data has been received.

[kalross]

The 600s interval I used was benchmarked over several days on PPC and worked well without issues.  It is however experimental, but has not simulated unprofitable results.

Is it possible for you to share your strategy or do you keep it secret?

Ha ha ;-)

Stephan's guidelines a few posts back are a good start.  To be honest....no one strategy fits all circumstances but I found than I learned rather quickly as the bot was haemorrhaging thousands of dollars due (mainly) to my inexperience in the beginning.  :-)

If you're serious and need some advice, then I am always happy to share what I have learned, but via PM please :-)  If too many people traded with the same settings then we'd never get any trades :-)


More generally.  Early findings and mistakes were:

1.  Using the default 20s timer (which will be removed).  It is widely documented in this thread that longer intervals are best.  3-10 mins.  20s is next to useless and the indicators do not respond well or very predictably for me.

2.  Watching the bot in the early days and intervening out of blind panic and inexperience/incompetence often making the wrong decision.  Let the bot roll!

3.  Bugs.  Stephan has done an excellent job in resolving them and the bot is approaching perfection.  The new releases might have a few teething problems, but we'll have to wait until the 20th to see.  Maybe the logic is largely untouched, though I doubt it somehow.

4.  I found using 'All In' was not as effective as spreading across a few viable and moving currencies.  Three in my case.  I trade on BTC-E as I had issues with Cryptsy which we thought was down to their server lag and API.

5.  Trading in fiat currency pairs was not good for me as like to have my reward in Bitcoin then wait to sell some at a good price to cash out some profit.   Occasionally I manually swing out to fiat to get out of a jam and then get back into to a BTC pair

6.  Long intervals and update timer can mean you are waiting a long time before the bot has sufficient data to trade.  This is annoying, but there is a planned online database of ticker data so that trading can be instant even with long duration settings.

7.  I found the drop loss function a little too primitive as it is based on fixed intervals of time and tend to not use it.  Same for the duration to reset last buy and sell.

8.  The reset Last Buy and Sell check box is a must for me.

9.  I don't disable minimum drop raize safety as I have a policy of never making an unprofitable trade.

Hope that helps a bit anyway.

K

[dldavis007]

I'm happy to say that my bot has finally made a profit.  About $150 (usiing around 33 LTC, About $1000) in the last 3 days.  I finally learned to leave it alone and let it trade as Kalross has mentioned above.

I also downloaded MetaTrader 4 and learned how to use it.  I haven't really got it trading yet, but here is a screenshot of a Strategy test using the Sample MACD Expert Advisor that I highly modified.  One of the main things I did was make it so it could Sell Short while waiting for a Long Buy to close out. 

http://i373.photobucket.com/albums/oo175/dldavis007/MetaTraderStrategyTesterDec26-27_zps49a2d8bd.png

I'm turning it on tonight.  I hope all goes well

Darrell

[krach]

"8.  The reset Last Buy and Sell check box is a must for me.

9.  I don't disable minimum drop raize safety as I have a policy of never making an unprofitable trade."

I agree with 9 and do the same, but 8, do you mean that you allow a reset or you do not allow reset? Normaly I dont touch the advanced settings at all and just wait.

[silverston]

I'm happy to say that my bot has finally made a profit.  About $150 (usiing around 33 LTC, About $1000) in the last 3 days.  I finally learned to leave it alone and let it trade as Kalross has mentioned above.

I also downloaded MetaTrader 4 and learned how to use it.  I haven't really got it trading yet, but here is a screenshot of a Strategy test using the Sample MACD Expert Advisor that I highly modified.  One of the main things I did was make it so it could Sell Short while waiting for a Long Buy to close out.  



I'm turning it on tonight.  I hope all goes well

Darrell

if not a secret, what settings do you use?
bot settings I mean

[kalross]

I agree with 9 and do the same, but 8, do you mean that you allow a reset or you do not allow reset? Normaly I dont touch the advanced settings at all and just wait.

Yes I tick the lower check box in advanced settings.  It was a feature we requested a good few pages back and Stephan was kind enough to implement it for us.

[dldavis007]

silverston,

if not a secret, what settings do you use?
bot settings I mean

I didn't do so well early this morning and most of the day.

The simple bot only lost about $180 (out of around $1000), pretty much all that I had gained over the weekend.  I had the stop-loss timer enabled and it hit it about 3 times.

If you are still interested in my settings, I posted them about 1 page back when subhead ask about them.

The second bot had a feature (more like a problem) that I wasn't aware of, so it quit selling short which would have saved my a$$.  When the account balance gets low and the bid amount is higher than the margin balance, it basically quits trading completely (no selling, no buying, no closing orders).  Anyway, I fixed it now and it is selling short again while I wait to get out of a long position.

All in all it was a pretty rough day.  I'm probably down around $750.  Hopefully I can make it up in the next few days as long as the action continues.

Darrell

[silverston]

silverston,

if not a secret, what settings do you use?
bot settings I mean

I didn't do so well early this morning and most of the day.

The simple bot only lost about $180 (out of around $1000), pretty much all that I had gained over the weekend.  I had the stop-loss timer enabled and it hit it about 3 times.

If you are still interested in my settings, I posted them about 1 page back when subhead ask about them.

The second bot had a feature (more like a problem) that I wasn't aware of, so it quit selling short which would have saved my a$$.  When the account balance gets low and the bid amount is higher than the margin balance, it basically quits trading completely (no selling, no buying, no closing orders).  Anyway, I fixed it now and it is selling short again while I wait to get out of a long position.

All in all it was a pretty rough day.  I'm probably down around $750.  Hopefully I can make it up in the next few days as long as the action continues.

Darrell

if BTC you will need to calculate a different Sell profit price for the "Price difference Indicator")

To be honest, I do not quite understand how to calculate this indicator for BTC. maybe there's some sort of formula? 

[dldavis007]

To be honest, I do not quite understand how to calculate this indicator for BTC. maybe there's some sort of formula? 

The price change indicator should be set so that it covers the fee plus any profit you would like to make.  Since BTC is around $850, the fee will be about 3.40 (.4%), so the price has to raise at least $3.40 just to break even.  I usually try to make a couple of dollars, so I set it around $5.  Any more and you may miss a trade, but you could set it higher, you would make more per trade, but make less trades.

good luck

Darrell

[lmazzola]

pardon the interruption guys/girls. I just bought the software. Wondering where I can get a manual to learn how to use it? Can anyone point me in a direction?

[silverston]

pardon the interruption guys/girls. I just bought the software. Wondering where I can get a manual to learn how to use it? Can anyone point me in a direction?

http://www.haasonline.com/simple-trade-bot-help/

[silverston]

To be honest, I do not quite understand how to calculate this indicator for BTC. maybe there's some sort of formula? 

The price change indicator should be set so that it covers the fee plus any profit you would like to make.  Since BTC is around $850, the fee will be about 3.40 (.4%), so the price has to raise at least $3.40 just to break even.  I usually try to make a couple of dollars, so I set it around $5.  Any more and you may miss a trade, but you could set it higher, you would make more per trade, but make less trades.

good luck

Darrell

thanks, will try it

[krach]

anyone know what the error "divide by zero" means?

[mies_reed]

Hey there,

I bought the simple trade bot today. Now I'm trying to set it up that it will sell and buy when the EMA lines cross each other (short 7 long 30). timeframe one hour.

Is this possible, if so. how?

I would really appreciate you help.

Thx in advance