Up to you, it depends on how many people you would like to get involved. I guess it will be enough to attach a file to this thread, so more eyes can comment on it.
Trading algorithm for EXANTE broker clients
- Thread starter arteml
- Start date
Up to you, it depends on how many people you would like to get involved. I guess it will be enough to attach a file to this thread, so more eyes can comment on it.
Okay, I’ll start with one. Can share others. This is backtest for ES.CME.U for July. I wasn’t trading the whole month but can also attach real P/L for most of the month trading with algo that made that backtest.
*note that there is a typo problem in backtest. When short pos is closed by TP the log still says “long closed”
Attachments
Why are the log timestamps always exactly in 15 minute intervals? Ie. 07:15:00, 07:30:00, 07:45:00 etc... This looks suspect to me, b/c one would expect variations of the seconds etc...
Is the program reacting only every 15 minutes to the actual market conditions?
And: what does "Benchmark volume" mean?
And: what exactly does this entry mean:
"Maximum draw down was -8.25 and it was 0.8048780487804879 of stop"
Is that absolute value or relative value, ie. percent?
I think you need a better documentation for your presentation, ie. a fully finished documentation, with a complete explanation of the used non-standard terms, as such snippets will not convince, I am afraid...
Your backtest results should use industry-standard methods of presentation, so that it can easily be verified and compared...
Is the program reacting only every 15 minutes to the actual market conditions?
And: what does "Benchmark volume" mean?
And: what exactly does this entry mean:
"Maximum draw down was -8.25 and it was 0.8048780487804879 of stop"
Is that absolute value or relative value, ie. percent?
I think you need a better documentation for your presentation, ie. a fully finished documentation, with a complete explanation of the used non-standard terms, as such snippets will not convince, I am afraid...
Your backtest results should use industry-standard methods of presentation, so that it can easily be verified and compared...
Last edited:
Sorry, I thought it’s quite obvious but I guess I should have clarified. I am trading 15 min timeframe or 1H timeframe. In this example it’s 15 min.
Benchmark volume is indicative volume, needed for the strategy to work properly. Of the current 15 min bar is okay to trade than it’s volume fixed as benchmark until the next one.
Actually I am not currently using MAXDD in the algo and this parameter left as indicative. I didn’t expect somebody to dig into this one, thought that people are interested in trades and stops. Not sure that this MAXDD calculated correctly, will check it. But -8.25 means -8.25/0.25 = 33 ticks.
Hi, thanks for the files but it is going to be quite difficult to analyze that pdf.
Can you produce an export similar to the one attached? It is an export from my live account.
Can you produce an export similar to the one attached? It is an export from my live account.
angrytrader, your method of logging is exemplary, ie. brilliant! kudos! 
Why are the log timestamps always exactly in 15 minute intervals? Ie. 07:15:00, 07:30:00, 07:45:00 etc... This looks suspect to me, b/c one would expect variations of the seconds etc...
Is the program reacting only every 15 minutes to the actual market conditions?
And: what does "Benchmark volume" mean?
And: what exactly does this entry mean:
"Maximum draw down was -8.25 and it was 0.8048780487804879 of stop"
Is that absolute value or relative value, ie. percent?
I think you need a better documentation for your presentation, ie. a fully finished documentation, with a complete explanation of the used non-standard terms, as such snippets will not convince, I am afraid...
Your backtest results should use industry-standard methods of presentation, so that it can easily be verified and compared...
About backtests.
The files I sent here is my log from test. It’s not some kind of presentation. Usually I am dealing with non-experienced people, thus P/L chart and detailed strategy explanation is enough. I suggest my clients to start testing with minimum lot so they can faster and cheaply understand what is that all about. That’s why I don’t make some detailed statistics.
Hmm. But here at EliteTrader the audience consists of mostly critical experts as also the name indicates: "elite traders"...
My .csv is not so clean as yours but hope you will understand.
Okay, got it.Hmm. But here at EliteTrader the audience consists of mostly critical experts as also the name indicates: "elite traders"...
Well, if anyone will be interested in the strategy we can discuss it in details. However, what I do here is trying to find any EXANTE clients so as discussion of the strategy wouldn't be a waste of time.
However, if I won't be able to find any EXANTE clients I guess I will be forced to relocate my algo to IB API and thus I will need to be more collaborative about my strategy.