Hi,
My strategy is written in IOG code, that is, calculation is performed on each incoming tick, is it possible or practical to optimize IOG-code based strategy like that? Any particular settings or considerations need to be taken care of? Thanks.
-kernel
IOG strategy optimization
- ABC
- Posts: 718
- Joined: 16 Dec 2006
- Location: www.abctradinggroup.com
- Has thanked: 125 times
- Been thanked: 408 times
- Contact:
Re: IOG strategy optimization
kernel,
in general it's possible and practical to optimize code that uses IOG (although there can be exceptions, if your code uses values that are
only available realtime for example). When you backtest/optimize code that uses IOG I'd suggest to use the Bar Magnifier with the highest possible
resolution (i.e. 1 Tick) to get results that are as close as the tick by tick live behavior as possible.
Regards,
ABC
in general it's possible and practical to optimize code that uses IOG (although there can be exceptions, if your code uses values that are
only available realtime for example). When you backtest/optimize code that uses IOG I'd suggest to use the Bar Magnifier with the highest possible
resolution (i.e. 1 Tick) to get results that are as close as the tick by tick live behavior as possible.
Regards,
ABC
- JoshM
- Posts: 2195
- Joined: 20 May 2011
- Location: The Netherlands
- Has thanked: 1544 times
- Been thanked: 1565 times
- Contact:
Re: IOG strategy optimization
Like ABC says, if you use the Bar Magnifier the backtested results should resemble real-live trading quite well.
I say 'should' because there are several things that cannot be addressed for with the Bar Magnifier, even when using a setting of 1 tick. One thing is the time between ticks (and the resulting liquidity in real trading); which is information that's lost with the Bar Magnifier. But if you don't trade/scalp very actively, this won't be an issue (or otherwise you can pessimistically adjust the backtest results).
I say 'should' because there are several things that cannot be addressed for with the Bar Magnifier, even when using a setting of 1 tick. One thing is the time between ticks (and the resulting liquidity in real trading); which is information that's lost with the Bar Magnifier. But if you don't trade/scalp very actively, this won't be an issue (or otherwise you can pessimistically adjust the backtest results).
- ABC
- Posts: 718
- Joined: 16 Dec 2006
- Location: www.abctradinggroup.com
- Has thanked: 125 times
- Been thanked: 408 times
- Contact:
Re: IOG strategy optimization
To clarify what I wrote above, when I say "to get results that are as close as the tick by tick behavior as possible", this means as close as you can get with MC. Historic testing will always be different to live trading as there is only so much you can simulate.
JoshM makes a good point about pessimistically adjusting the backtest result. In fact I'd even prefer to always be pessimistic in the testing (in terms or slippage, fills etc.) as I don't think it's likely that you'll discard a profitable strategy due to that, but it will help you avoid losing ones.
Regards,
ABC
JoshM makes a good point about pessimistically adjusting the backtest result. In fact I'd even prefer to always be pessimistic in the testing (in terms or slippage, fills etc.) as I don't think it's likely that you'll discard a profitable strategy due to that, but it will help you avoid losing ones.
Regards,
ABC