Backtest - whats the point?

[JesseOnt]

I created a simple strategy and I back tested through last month's data (from IB). It was a huge profit on daily basis, every day. Hard to believe.

I then re-played a few days real time data and tested upon that re-play...it was a loss on daily basis. However, it's more 'real' and I think thats how the strategy will work in live trading.

So what's the point of 'backtesting' while it gives you fake result? I knew the back test data only comes with HLOC type data...I guess the main purpose of back testing is for optimzing?

Jesse

[ctrlbrk]

I feel your frustration.

On your backtest results, go to the trades page and sort by # of bars. If there are many results on '1' bar, that could be the source of your trouble. Buying and selling on the same bar in a non-real-time strategy won't show accurate results, because it can't know the order of the ticks, just a static OHLC for the entire bar.

Now if it's more than 1 bar, and the results are still major different than real time... are you using market or limit orders? Market orders in backtesting always fill at the Close[0] value (in my experience). In real time, that is not always the case of course.

You might try adding 1 or 2 ticks of slippage to better approximate real-world results. You can do this in the options menu when running the backtest.

Backtesting in general still very much serves as an important function in developing and testing a strategy. I would use backtest results over the last 3 years, then run as many days in 'market replay' mode or in a sim as I can possibly stand, before going to real money.

GL and HTH!

Mike

[NinjaTrader_JoshP]

Jesse,

Please see my response over here: http://www.ninjatrader-support2.com/...1&postcount=10

[NinjaTrader_Ben]

Hello,

This link will have some good information on why the results differ:

http://www.ninjatrader-support.com/HelpGuideV6/DiscrepanciesReal-TimeVsBacktest.html