Wasn't sure which forum to post this in as it applies to *many* but figured I would give it a go here. Have read many posts looking for my exact answer to no avail, so here it goes.
Backstory: Developed an FX algo that was trading on ninjatrader with a reasonable amount of success. Went live when there only option to trade was thru was Gain capital and the re-quoting that went on screwed up my results as I wasn't being executed in the trades that the system thought it was. Decided I couldn't go live on ninja and looked for a different platform. checked multicharts, meta trader, etc and decided that all of them had some short comings so decided to go platform agnostic.
Had programs black-boxed and now run on my server via FIX api to my forex FCM. Have been modelling the algo for a hedge fund and it has performed over the last 6 months very well and did what it is supposed to do.
Now they want to see back-tested results, as well as out of sample, random walk etc. Obtained 2 years worth of back tick data for close to 20 different forex pairs and had the data cleaned the best that it could be. there are still some significant gaps, but those are manageable.
The question is HOW do I backtest the algo without a platform and what metrics are used to review that performance? Right now we are strictly going off of P&L and drawdown.
This seems like such a basic question yet I was afraid to ask the investor for fear of showing my lack of knowledge. Is there a platform for just bactesting or a template I can follow to achieve what I am looking for? My programmer likes to write everything from scratch but I am afraid that A) this may take too long and B) a mistake is bound to happen one of these days and I would hate for it to be this!
I don't mind paying for something that is comprehensive and secure, though free is always better if given the choice.
So I'm basically looking for something that I can upload my data to, do the backtest, and then have the various metrics spit back. Sharpe, sortino, max DD, etc. In fact, I'm not even sure all of what would be needed in the eyes on an investor. I suppose in this instance more is better, but I want to know what the "must haves" are to make sure I do not miss something important.
If you have experience in this area, please help!
Thx.
Backstory: Developed an FX algo that was trading on ninjatrader with a reasonable amount of success. Went live when there only option to trade was thru was Gain capital and the re-quoting that went on screwed up my results as I wasn't being executed in the trades that the system thought it was. Decided I couldn't go live on ninja and looked for a different platform. checked multicharts, meta trader, etc and decided that all of them had some short comings so decided to go platform agnostic.
Had programs black-boxed and now run on my server via FIX api to my forex FCM. Have been modelling the algo for a hedge fund and it has performed over the last 6 months very well and did what it is supposed to do.
Now they want to see back-tested results, as well as out of sample, random walk etc. Obtained 2 years worth of back tick data for close to 20 different forex pairs and had the data cleaned the best that it could be. there are still some significant gaps, but those are manageable.
The question is HOW do I backtest the algo without a platform and what metrics are used to review that performance? Right now we are strictly going off of P&L and drawdown.
This seems like such a basic question yet I was afraid to ask the investor for fear of showing my lack of knowledge. Is there a platform for just bactesting or a template I can follow to achieve what I am looking for? My programmer likes to write everything from scratch but I am afraid that A) this may take too long and B) a mistake is bound to happen one of these days and I would hate for it to be this!
I don't mind paying for something that is comprehensive and secure, though free is always better if given the choice.
So I'm basically looking for something that I can upload my data to, do the backtest, and then have the various metrics spit back. Sharpe, sortino, max DD, etc. In fact, I'm not even sure all of what would be needed in the eyes on an investor. I suppose in this instance more is better, but I want to know what the "must haves" are to make sure I do not miss something important.
If you have experience in this area, please help!
Thx.