Simulator Issues???

[maninmoon]

Hi guys,

I started a thread about accuracy of simulator against Live trading results. But I have received no replies - does that mean there are some issues with how it fills trades in backtest? Or am I addressing the wrong forum?

[NinjaTrader_Bertrand]

maninmoon, I'll check into your other thread, but yes there are some expected discrepancies when comparing backtesting to Market Replay or Live trading results - http://www.ninjatrader.com/support/h...ime_vs_bac.htm

[maninmoon]

Hi Bertrand, thatnks for your reply.

But I was rather hoping to get some feedback from someone trading Live too!

[NinjaTrader_Bertrand]

maninmoon, you can generally use backtesting as a tool to judge if you have an edge worth spending more time testing on in market replay and live sim trading, then if you're still happy with the results you could move to testing on a live account / market with entry level size and move up if results are in line...

[maninmoon]

No offence Bertrand, but that's exactly what I am trying to ascertain from someone who has actually gone through this process...

[NinjaTrader_Bertrand]

No worries, I hope that other members share their experiences as well - I will not get into making personal trading recommendations here....

[neb1998]

maninmoon, What your asking for is the confidence to trade the system. That has to be developed by yourself using what you have learned and through experience. The "EDGE" of being successful or not is not something many successful traders are going to share openly. Backtesting is curve fitting, the trick is to find a way as to do it with as little curve fitting as possible so it still works for a while into the future with similar results. Expect results never to be the same as backtesting...sure it happens but its not realistic to aim for.

Form a hypothesis, test it, confirm it, trade it or burn it. and restart.

[maninmoon]

Hi neb1998,

Thanks very much for your reply.

I am not asking for secrets - I just want confirmation that someone has found merit in the simulations, judged by what they see trading Live.

For all I know, perhaps the simulation is too far from reality to add value...

I assume you are trading and have a positive experience of using the simulator? I am interested in the High Frequency end of the spectrum.

[neb1998]

A Simulator is only as good as the person running it. If you dont understand how it works yourself then i would not trade a penny on the live market. Run as many short simulations like with pop size of 10 and generations at 10 - as you need to so that you understand why your getting the fills you are. Then you can design systems accordingly.

Start with a very simple system, like the MA Xover that comes with ninja, and make sure you understand how the simulator works on it before you attempt on your own work.