Pages:
Author

Topic: Haasonline Simple Trade Bot For BTCe and Bitstamp[Main topic] - page 82. (Read 416703 times)

hero member
Activity: 952
Merit: 502
SAPG Pre-Sale Live on Uniswap!

Please: Have plans to make avaliable a interface so I can write my own functions-like mt4?
Also: Maybe all this is my fault but i have read almost all of this thread and haven't found someone posting a working configuration for this BOT, only "educational" examples. Is secret or someone can share a good working configuration?

I think this bot is really simple and have many good functions and potential but lack flexibility. One example: possibility to  Buy/sell be dependent from one function result OR other function result. Or something as bot buy at x ( x is last buy value ) do normal operation but if anytime current value is below 95% of x bot should sell all values imediatly.

Regards
 
  

same thing.
member
Activity: 80
Merit: 10
"Can you explain how two indicators are combined to determine when a "Buy" and "Sell" are made?"
Hi,
There is only a buy or sell action when: ALL indicators that you have active say buy or sell & the price is enough above the last buy price or below the last sell price.

Bot bought at  1$ - now indicators say "BUY" but the price is only 1.001$ therefore you would make a loss on exchange fees, therefore there no action.


Having one indicator should cause more action and more indicators will have "stronger" signals, depending on how you set them up.

IMO you would have to organize the crossovers so they are "on beat" as if you where a dj. I have only tested this with 2x RSI but it works pertty good.
So if the crossover of indicator 1 is 10/20 then the next one would have to be the same, half, or double, then it is "on beat" otherwise you will have a lot less agreement and less signals. indicator one 10/20 and indicator two 20/40 ect. RSI works fine like that try default lenght 30 - 30/70 and then lenght 60 30/70 , then you will produce a singnal with wave forms that have shapes and not ones that "flange" if they flange then you wont have a clear signal. This is only one way to look at it, I suggest experimenting with really small ammounts of coin 0.01 btc or 0.1 ltc or 1 xpm ect ect.




Stephen: I lost almost 2 last weeks triyng every combination and benchmarking but i am unable to get bot do what i want: I think this metod to combine indicators very limitative ( for instance example above w/ Stochastic not in sync ) and for me is impossible in only one combination of functions put differents strategies for  up and down trends.

Until now bot only gave me loses and even drop loss mechanism is not reliable ( at least w/ Bitstamp ).

I suported this bot because i want to avoid time to learn metatrader 4 but lost time on this bot seems, for me, enough put MT4 to work ( and it is free for BTC-e clients ).

Please: Have plans to make avaliable a interface so I can write my own functions-like mt4?
Also: Maybe all this is my fault but i have read almost all of this thread and haven't found someone posting a working configuration for this BOT, only "educational" examples. Is secret or someone can share a good working configuration?

I think this bot is really simple and have many good functions and potential but lack flexibility. One example: possibility to  Buy/sell be dependent from one function result OR other function result. Or something as bot buy at x ( x is last buy value ) do normal operation but if anytime current value is below 95% of x bot should sell all values imediatly.

Regards
 
  
sr. member
Activity: 378
Merit: 250
Is there anything else with the new software release if you don't trade on cryptsy?
newbie
Activity: 30
Merit: 0
Krach,

Quote
"Can you explain how two indicators are combined to determine when a "Buy" and "Sell" are made?"
Hi,
There is only a buy or sell action when: ALL indicators that you have active say buy or sell & the price is enough above the last buy price or below the last sell price.

Bot bought at  1$ - now indicators say "BUY" but the price is only 1.001$ therefore you would make a loss on exchange fees, therefore there no action.


Having one indicator should cause more action and more indicators will have "stronger" signals, depending on how you set them up.

IMO you would have to organize the crossovers so they are "on beat" as if you where a dj. I have only tested this with 2x RSI but it works pertty good.
So if the crossover of indicator 1 is 10/20 then the next one would have to be the same, half, or double, then it is "on beat" otherwise you will have a lot less agreement and less signals. indicator one 10/20 and indicator two 20/40 ect. RSI works fine like that try default lenght 30 - 30/70 and then lenght 60 30/70 , then you will produce a singnal with wave forms that have shapes and not ones that "flange" if they flange then you wont have a clear signal. This is only one way to look at it, I suggest experimenting with really small ammounts of coin 0.01 btc or 0.1 ltc or 1 xpm ect ect.

Thanks for the suggestions, I will give them a try.

After trying out the PPO and Stochastic Indicators together, I finally realized that the PPO indicator doesn't just Sell on the crossing but at the crossing and after, so it makes more sense that the two indicators could work together.  In my tests, the combination of the two sometimes made the trades more profitable, but often caused the bot not to make a trade when it was obvious that it should have.

Overall, I'm not convinced that a PPO Indicator and a Stochastic Indicator will work well together.  The PPO Indicator usually signals a "Sell" just after a peak.  Since a Stochastic Indicator usually reacts faster than the PPO Indicator, I would expect that the Stochastic Indicator has already gotten out of the "Sell" condition and is probably headed for a "Buy" condition.   

It seems that the only way to satisfy both indicators would be a rally with a small dip followed by a continuation of the rally.  In this case the trade would probably be more profitable than using the PPO Indicator alone.

Darrell
sr. member
Activity: 449
Merit: 250
Software developer
Version 0.9.0.9 beta has just been released.

This release will affect the Cryptsy users the most. The following coins are added: Asc, Bet, Col, Doge, Dem, Xjo, Lk7, Pts, Tag, Tek, Tgc & Uno. Hopefully this will solve a lot of reported issues of Cryptsy. If not, please keep me posted about it.

I also fixed a bug inside the drop/raise safety within this release. If you normally disable this safety then there is no need to get this update.

Edit:
All existing users can uninstall the old version, download the new version at the My Account -page and install it again.
sr. member
Activity: 449
Merit: 250
Software developer
I have collected the past days the comments about Cryptsy and i did found a problem with some new coins. Today i will make the needed changes and i will release a new version. If you do not have any problems with Cryptsy you do not have to update. This will be something like a quickfix.

@krach: Thanks for giving the answer. Wink

@dldavis007: To most easy way it to look at the indicators as they are a sum. If they both agree on something a signal will be produced. (of course the safeties can stop this signal back for your safety)
legendary
Activity: 1851
Merit: 1020
Get Rekt
"Can you explain how two indicators are combined to determine when a "Buy" and "Sell" are made?"
Hi,
There is only a buy or sell action when: ALL indicators that you have active say buy or sell & the price is enough above the last buy price or below the last sell price.

Bot bought at  1$ - now indicators say "BUY" but the price is only 1.001$ therefore you would make a loss on exchange fees, therefore there no action.


Having one indicator should cause more action and more indicators will have "stronger" signals, depending on how you set them up.

IMO you would have to organize the crossovers so they are "on beat" as if you where a dj. I have only tested this with 2x RSI but it works pertty good.
So if the crossover of indicator 1 is 10/20 then the next one would have to be the same, half, or double, then it is "on beat" otherwise you will have a lot less agreement and less signals. indicator one 10/20 and indicator two 20/40 ect. RSI works fine like that try default lenght 30 - 30/70 and then lenght 60 30/70 , then you will produce a singnal with wave forms that have shapes and not ones that "flange" if they flange then you wont have a clear signal. This is only one way to look at it, I suggest experimenting with really small ammounts of coin 0.01 btc or 0.1 ltc or 1 xpm ect ect.


newbie
Activity: 30
Merit: 0
Stephan,

Quote
@dldavis007:
Let me help you make better trades; First the targetprices, they will guard you will overcome the fee-costs. Hopefully this part is clear already. But now you have added a PPO indicator. This is a "long range" indicator, one that will benefit from long update intervals. This strategy is working properly when setup with a good interval. However you can improve your trade-momentums even more. The PPO "long range" indicator does not know anything about the current price on the short range. So to solve this matter you can add a Stochastic indicator or a RSI indicator too.

Give this a try in the benchmark to see what i am suggesting here. You will see the last profit you have posted will go up even more. Smiley

Can you explain how two indicators are combined to determine when a "Buy" and "Sell" are made?

For instance, the PPO Indicator "Buys" and "Sells" on the crossing (assuming that the trade is profitable).  I assume that the Stochastic Indicator "Buys" and "Sells" when it is above %80 and below %20 (and when the trade is profitable).  Since the PPO Indicator buys at a specific point (the crossing), I don't really see how adding a stochastic indicator would help to make the trades more profitable (I assume that the Stochastic Indicator would somehow make the trade occur nearer to the peak than just after the peak where the PPO itself would make the trade).

Could you please give a little more detail on how these indicators actually interact to Buy and Sell.

Thanks,

Darrell
newbie
Activity: 23
Merit: 0
API For Cryptsy is not working.
sr. member
Activity: 350
Merit: 253
there is a strategy to increase btc and not $?

thanks

Wouldn't that be mutually inclusive? Getting your bot to buy low/sell high would probably be a good starting point.
sr. member
Activity: 292
Merit: 250
there is a strategy to increase btc and not $?

thanks
sr. member
Activity: 449
Merit: 250
Software developer
@kireinaha:
I can be related to the signal line. Of course i am not sure, but i have had some people asking the same and there problem was most of the time the length of the signal line. I hope the same is going on in your case.
If you do find something wrong i am very happy to have a look and solve it.

@afia:
The release of January will get a new interface, scaling and thing like that will be problems of the past. At the current moment i test at the resolution on a screen of 1280x1024 and up. Which is not, i agree, perfect of course.

TIP: Maybe this is usefull: but you can scale the charts, there is some hidden line between them which allows scaling up and down. Maybe this will help a bit getting better charts in your screen.
newbie
Activity: 21
Merit: 0
hmm, mb u will do it for normal resolution? omg... with 1366х768 can't tap on save setting cause it was out of my screen. can look only 1 graphic at a time. and if i try both graphic they are so small that its hard see smth. think right down corner menu must be smaller. and this buttons with "add indicator" must be drop down menu like. anyway thx for your work. will test it soon.
sr. member
Activity: 350
Merit: 253
There are differences between the charts of bitcoinwisdom. I can confirm this. But that does not say the Simple Trade Bot is wrong.  Smiley

First lets scale the posted image a little bit better, just a little bit will do (i am not good at this). But this is what you will get:



Alright, now lets a look at this with 3 things in mind:

1. The lines of bitcoinwisdom are more smooth because they apply a smoothed line while the Simple Trade Bot shows you raw data points.
2. The moment of reading the prices is different for sure. So both will get a bit different prices inside there calculations. In SMA and WMA this is not really visible, but in EMA it is.
3. You can not temper with mathematics.

Like i posted before, i am looking into this but i have not yet find anything wrong with it. If i did, i would have posted it.

BTW This far only dldavis007 posted something very interesting about the MACD-MA outcomes.


Thanks Stephan for the response. It's strange, because I've been playing around with the bot for the past few days, trying some news indicators out, and every time I use the MACD or MACD-MA, it's like my trend follows five intervals behind BitcoinWisdom's & Bitcoin Charts. So if I have my update timer on 3 minutes, I end up 15 minutes behind the trend. I'm sure that the results can vary a little depending on the polling, but this is a huge difference. I tried reinstalling the bot to see if it's a local problem on my end, but it's still doing it. And the same thing happens no matter which interval I use, short or long. So I have no idea what's going on with that, but at least the ROC indicator seems to be working fine.

Anyone else out there experiencing this, or just me?
sr. member
Activity: 449
Merit: 250
Software developer
There are differences between the charts of bitcoinwisdom. I can confirm this. But that does not say the Simple Trade Bot is wrong.  Smiley

First lets scale the posted image a little bit better, just a little bit will do (i am not good at this). But this is what you will get:



Alright, now lets a look at this with 3 things in mind:

1. The lines of bitcoinwisdom are more smooth because they apply a smoothed line while the Simple Trade Bot shows you raw data points.
2. The moment of reading the prices is different for sure. So both will get a bit different prices inside there calculations. In SMA and WMA this is not really visible, but in EMA it is.
3. You can not temper with mathematics.

Like i posted before, i am looking into this but i have not yet find anything wrong with it. If i did, i would have posted it.

BTW This far only dldavis007 posted something very interesting about the MACD-MA outcomes.
newbie
Activity: 42
Merit: 0
Does it work under Linux ? Where can I look at user manual ?
newbie
Activity: 8
Merit: 0
and this is why there should be open source.. if not for the entire software, at least for the mathematical calculations and the main algorithms and indicators. since pretty clearly there are some errors, that could've been fixed a long time ago, had other people had a chance to look at and test the program :\
sr. member
Activity: 350
Merit: 253
Yes, I have been running the MACD-EMA indicator on 3 minute intervals for awhile, and I must say I'm not very impressed with the results. When comparing it with the MACD on BitcoinWidsom -- both on 3 minute intervals with 12/26/12 settings -- I'm getting completely different values. It almost seems that the STB values are lagging several minutes behind BitcoinWisdom.



Note that BitcoinWisdom's MACD signal line crossed but STB's MACD signal line -- on the exact same settings -- is several minutes away from crossing still. What can explain such a marked difference in output?
member
Activity: 80
Merit: 10
Are you totally sure the setting are fully equal?
I have tested this too and compared the output to the charts bitcoinwisdom and bitcoincharts.  Undecided

I will have a look, there is no difference in the math we both use.

@Stephan
Which indicator in STB is most similar to the MACD perimeter on BitcoinWisdom?

I've tried the 12/26/9 setting on the MACD-MA indicator with EMA and SMA values, but I seem to get different results. Perhaps the time offset between my bot and their site can attribute to some of that, but often I get some very different results, ie. BitcoinWisdom's MACD will show buy, but my bot's MACD-MA will show sell and vice-versa.  Huh

Edit: I think I may have figured it out... the "difference" seems to do with the way the histogram is drawn up in the STB. I think the MACD-MA with SMA trend type is most similar to BitcoinWidsom. Seems about right... but would be happy if you can confirm for me Smiley

I have also one question related w/ long interval (3m,5m and 15m) : On Bitcoinwisdom they do a polling and calculate values with many samples on interval ( well is whats appears to me..) on STB you just do one sample on interval (and have a possibility to have a bad one) or do also some average and processing - I think this can help to explain some differences.

Regards
member
Activity: 85
Merit: 10
I'm still having issues with stuck open orders (and vice versa...bot continuing trading when order is still open).

Bot bought more NMC today before sell order was complete on BTC-E.

K
Pages:
Jump to: