Quote from mvalley:
The reason is my backtesting only tests things that dont gap down & fill. I'm using stockfetcher to backtest and it just avoids those situations.
I guess in theory I could try it in amibroker, but I dont think it would backtest very well.
-Mike
(I know I'm going to get a lot of flak about saying this, but....)
For the most part, backtesting is meaningless to making money. You can't possibly take into consideration all the variables involved to attempt to "prove" much of anything. Why backtest without taking into consideration such basics as VIX for each time frame, $$ valuations at each time frame, "Peer groups" RS, mergers, political climate, and on and on.
I have seen too many smart people spend a lot of time "testing" - only to find that when they attempt to trade with the results, they don't make a nickel. The market is a living-breathing thing with so many factors to consider, that even if you find a workable pattern to follow, by the time you try to trade it, it's just too late.
We encourage our traders to learn "what works" and then spend time to maximize profits via order entry techniques and automation where applicable. If something works, and is making me money, I would rather spend time making money than backtesting to see "why" it's making money.
Once again, there are always exceptions to what I'm trying to say here.
Also, we "do" a lot of reviewing historicals for correlated pairs trading, which gives us a clear analytical view for current trading.
BTW is MVALLEY = Mill Valley CA? Just curious.
All the best,
Don
