What Performance Metrics do you like to Track?

i would say if all cost included "like commissions and any slippage" if you can achieve 1.25 or better you can make money. a profit factor of 2.80 only indicates that it hasn't been run for years and years. many systems will have burst up to 4.00+ occasionally but that's not sustainable, i feel right at home with 1.25 to 1.75 over a decade or two at least of trades. i like to see and use 5000 trades to get a sense of how durable the model is. almost always the more data you add the more it degrades the results.

if i don't have enough data, i make a synthetic data stream that mimics the markets i'm trading. i love demolishing trading systems, very seldom i will find one that holds up. the most consistent models are those that exploit reaction of human emotion .
 
MarkBrown said:
i would say if all cost included "like commissions and any slippage" if you can achieve 1.25 or better you can make money. a profit factor of 2.80 only indicates that it hasn't been run for years and years. many systems will have burst up to 4.00+ occasionally but that's not sustainable, i feel right at home with 1.25 to 1.75 over a decade or two at least of trades. i like to see and use 5000 trades to get a sense of how durable the model is. almost always the more data you add the more it degrades the results.

if i don't have enough data, i make a synthetic data stream that mimics the markets i'm trading. i love demolishing trading systems, very seldom i will find one that holds up. the most consistent models are those that exploit reaction of human emotion .
More...
Good Morning MarkBrown,

I agree with you fully. Finding a trading system with edge tested on X-XX years of historical data with +X,XXX trades is not easy, and it should not be easy.
 
SimpleMeLike said:
Good Morning MarkBrown,
I agree with you fully. Finding a trading system with edge tested on X-XX years of historical data with +X,XXX trades is not easy, and it should not be easy.
More...

Finding profitable edges is easy. But finding really good profitable edges is hard.

Low edge, 40% to 60% profitable months. (eg. Sharpe ratio between 0.5 to 1.5)
Medium edge, 60 to 80% profitable months. (eg. Sharpe ratio between 1.5 to 2.5)
Really good edge, 80% to 100% profitable months. eg. Sharpe ratio above 2.5.

Above should not include blow up strategies like selling premium. Which win for many months but then lose it all...
 
Last edited:
Businessman said:
Finding profitable edges is easy. But finding really good profitable edges is hard.

Low edge, 40% to 60% profitable months. (eg. Sharpe ratio between 0.5 to 1.5)
Medium edge, 60 to 80% profitable months. (eg. Sharpe ratio between 1.5 to 2.5)
Really good edge, 80% to 100% profitable months. eg. Sharpe ratio above 2.5.
More...
Good Morning Businessman,

You right. Your right about that. You are so right. I agree. I am stuck in the Low edge range from my work and research. I personally have not obtained any credible edge above profit factor 1.0 with my out-of-sample testing. Finding credible edge is a tuff business.

But as Jim Simon quoted, "thank God, edges are not easy to find"
 
Last edited:
You must log in or register to reply here.
Back
Top