Austin, TX
Experience: Advanced
Platform: TradeStation
Trading: Futures
Posts: 839 since Mar 2011
Thanks Given: 124
Thanks Received: 704
|
Not sure if I'm understanding correctly, but I trade tick/tick....
I can use setstopcontract/setprofittarget and setstopcontract/setdollartrailing and it does a fairly good job of assuming the OHLC...obviously it has better precision/resolution with smaller timeframe charts (or the larger your profit-stop bracket is with respect to average bars).
However, as you pointed out, the part where the backtest fails miserably is that without looking inside the bar for entry criteria, I'm stuck using previous bar confirmation of an event and waiting till the next bar to initiate the trade obviously has undesirable consquences on the result of the backtest.
If you wait till the next bar on a 60 minute for an MA triggered entry, many times you've already missed the bus when backtesting.
The only thing I've been able to do is backtest, record the results and then forward test in simulation and do a quantitiative comparison between the results to give a sort of "interpolation factor."
i.e. if I backtest a strategy bar to bar and it yields a 1.05 profit factor and a 56% win rate, then forward test and it yields a 1.5 profit factor and 68%, I can draw linear conclusions (rough) as the the relative conversion rate for similar type strategies.
The problem as you might suspect is that beyond being a "crude" comparison, the relevance goes down with the ratio of period evaluated in backtest compared to period evaluated in forward test.
I can hedge out a little bit of that by simply backtesting the same period I forward tested, getting that ratio and then applying it further back (and further forward).
|