Is this strategy good enough?

February 10th, 2011, 02:02 PM

dutchbookmaker

I'm surprised we don't see more discussion about this. The metrics on the last 50 trades of a system that takes multiple trades a day would be far more valuable than backtest going back to 1999 that includes the dotcom bust, the crisis..stuff that you can't reasonably assume going forward. If anything if the system performed great over the crash you should be biased to NOT trade it as it is almost a certainty we will not see that level of volatility any time soon.
On the original system buying 25k worth of SPY on sept 1st and holding until now would have given practically the same thing as mentioned. That is why it is not enough information to base things off.
Sharpe ratio is totally overblown IMO too. Viewing historic data as if regimes don't change is just bizarre and not very logical.
If you have two systems that both made an equal profit the last 6 months, however one took 80% shorts and one took 80% longs..the one that took 80% shorts is VASTLY more impressive and has a much higher chance that it is actually exploiting some kind of market inefficiency. That would not be reflected in common metrics though.

Walk forward testing is more reliable. Start at the beginning of your historical test period, backtest 6 months, optimize on those 6 months, take those optimized values, and use them to backtest the next 2 months, and repeat in that fashion moving forward in time..

February 10th, 2011, 02:20 PM

baruchs

Well its a wrong answer, sorry. This result is not possible!..
I'm not playing, its basic understanding about trading. I don't understand how people can try being traders without understanding some basics.

Respectfully, I can only speak for myself when I say that my efforts at being a trader leads me to ask the questions that I do, and offer what I feel that I know to assist.

Jon

February 10th, 2011, 02:23 PM

monpere

Walk forward testing is more reliable. Start at the beginning of your historical test period, backtest 6 months, optimize on those 6 months, take those optimized values, and use them to backtest the next 2 months, and repeat in that fashion moving forward in time..

monpere,

Each for their own ... I prefer to find a generally bullish time period with some irrational bearish activity and optimize the SHORT side of the strategy on that information: keep those parameters and see how they worked before and after that, test/in -sample, time period.

This is an example (chart of weekly PnL, optimized on 7 weeks for Shorts only, all is backtesting on replay data): can you decipher which 7 weeks were used for optimization? I cant.. maybe you can .. anyone? Is it goodenuf? Fees were not deducted, and no allowance for slippage.

YMMV and I hope it does,
Jon

February 10th, 2011, 03:37 PM

I need everyone's input ; If I explain my trading system and everyone back tests it and tells me it won't produce a profit... do I have to give the money back that I made trading with it? I don't want to create an embarrassing faux pas.

Regards,
TMFT

February 10th, 2011, 03:39 PM

ThatManFromTexas

I need everyone's input ; If I explain my trading system and everyone back tests it and tells me it won't produce a profit... do I have to give the money back that I made trading with it? I don't want to create an embarrassing faux pas.

Regards,
TMFT

No you have to give to your favorite charity...ME!!!

February 10th, 2011, 03:45 PM

MWinfrey

No you have to give to your favorite charity...ME!!!

@Mike Winfrey

If you trade your system ... you will be taking all the money you want from all the other traders... you won't need my help

February 10th, 2011, 03:46 PM

ThatManFromTexas

@Mike Winfrey

If you trade your system ... you will be taking all the money you want from all the other traders... you won't need my help

oh ok...I'll do it on my own. I don't want your money.

February 10th, 2011, 03:49 PM

ThatManFromTexas

I need everyone's input ; If I explain my trading system and everyone back tests it and tells me it won't produce a profit... do I have to give the money back that I made trading with it? I don't want to create an embarrassing faux pas.

Regards,
TMFT

yes you do. Now would you is one other question!

TickVix/Gregory

February 10th, 2011, 05:04 PM

tickvix

yes you do. Now would you is one other question!

TickVix/Gregory

Well no.. I have too many of the girls who work at the Eager Beaver on scholarship ... they say I remind them of J Howard Marshall ... but just not as good looking...

February 10th, 2011, 07:28 PM

ThatManFromTexas

I need everyone's input ; If I explain my trading system and everyone back tests it and tells me it won't produce a profit... do I have to give the money back that I made trading with it? I don't want to create an embarrassing faux pas.

Regards,
TMFT

Only that amount that came out of my account!

I am not greedy ... only bad traders are greedy lol

Is this strategy good enough?

Discussion in Currencies

Is this strategy good enough?