System MA filter suggestion

[CamKawa]

I'm just starting out building a system and have created a very basic moving average crossover expert for Metastock.

1. Do you think the whipsaw factor here is too high?
2. What filter(s) would you suggest?

Below is a system backtest result for 400 periods, stop loss @ 2%, going both long and short, over the 126 shortable stocks available from IB. The gain was 33.4%. Is that any good?

 

[Boggo]

First thought that comes to mind is have you delayed both the entry and exit by one bar, ie entries and exits are on the open of the day following the signal.


I stuffed up a formula in Tradesim when I started using that and got some great results until I corrected it to what would really happen.


Just my

 

[CamKawa]

First thought that comes to mind is have you delayed both the entry and exit by one bar, ie entries and exits are on the open of the day following the signal.

No I don't think I have, but I've been wondering out it. I'm preparing myself for that when I start paper trading the system... or maybe now.

I've got the MA to calculate on the close. I'll move it to the open. Will that help? As I said I'll have to paper trade it to discover little gems like that. I'll keep an eye out. Thanks.

 

[Boggo]

No I don't think I have, but I've been wondering out it. I'm preparing myself for that when I start paper trading the system... or maybe now.

I've got the MA to calculate on the close. I'll move it to the open. Will that help? As I said I'll have to paper trade it to discover little gems like that. I'll keep an eye out. Thanks.

Have a think about how you would trade this in reality, ie. if your signal is generated at the end of the day then you will not be able to enter until the open of the next trading day at the earliest.

Same applies to your exit, signal today after the market has closed, earliest exit is tomorrow on the open.

I am assuming that you are using end of day data, you can only really enter on the day that the signal is generated if you are using intraday data and charting such as eSignal data with NinjaTrader, MTPredictor and Tradeguider etc to mention a few.

Hope I am not confusing you here.

(I have had a couple of reds with dinner and I am rapidly coming to the conclusion that reality is an illusion created by alcohol deficiency )

 

[Boggo]

I've got the MA to calculate on the close. I'll move it to the open. Will that help?

Addition to my previous post... no it wont help, you generate your signal based on end of day close, now the reality is that you cannot enter until the next day (long or short), consequently you must backtest based on that fact.

Signal today, entry/exit tomorrow.

 

[tech/a]

Your wasting your time.
As with all M/A crossover systems you'll be whipsawed to death as they are lagging..
Brokerage will kill you off.
You only make a profit when your in a trend.
Stocks dont trend all the time.
A 33% win rate is acceptable providing you have very big wins and small losses.

Whats your R/R
Whats your return /$ invested
Whats your drawdown.
Blah blah