Respectable Profit Factor

Ubzen 2011.07.12 10:47 #1

Imo, the relative drawdown% is more important than pf. However, if you want a number then 2> is my answer. As-far-as-ideas, isn't that why we have all these forex forums. Tho, I must admit the only forum I can navigate easily is mql4.com, the others I run into interesting stuff searching google.

How , when close 5 points/day or 100 a trading strategy based

YYZ 2011.07.16 09:41 #2

ubzen: thanks for responding. And you would consider a decent drawdown to be what? Thanks!

Ubzen 2011.07.16 14:56 #3

A relative draw-down of 30% or less is acceptable for me on a 10+ years back-test.

YYZ 2011.07.16 20:29 #4

ubzen: Thanks :)

Elroch 2011.07.16 23:23 #5

LINE123:

Hi: This may appear to be a stupid question but could I hear some feedback on what are respectable profit factors. Obviously anything over 1.0 is an accomplishment but at what point would you say is a good (safe for trading) PF?

Thanks for any input.

Also does anyone have any interest in participating or creating an online workshop to share ideas?

You are right that if you know that your method gives a profit factor of over 1.0 it is useful. One issue is that there is always going to be some uncertainty in the profit factor of a trading method going forward. This is an argument for requiring a higher P.F. before risking money.

It is a fact that many traders who trade large volume at high frequency have rather low profit factors. There is a statistical argument why a lower profit factor is acceptable for higher frequencies of trading. This is that, roughly speaking, several trades with a low profit factor correspond to one trade with a higher profit factor. More precisely, if you define a quality statistic for a trade which is the ratio of it's mean return to its standard deviation, then you can do the same for several trades (the ratio of the total mean return to the standard deviation of the total). If the several trades have the same statistical properties (eg they have the same expected profit factor) then their combination has a better quality statistic. More precisely, the ratio of the mean return to the standard deviation of N trades is sqrt(N) times as high as for one trade (if they all have the same statistical properties). Hence high frequency low quality trades (but profitable on average) can be as good as low frequency high quality trades.

This doesn't answer your question, but it does give a useful way (which many people are probably not aware of) to compare the usefulness of two systems which provide different frequencies of trades.

Some Ideas to Select Optimise an EA and stop loss, profit target

Brian Dee - Random Trader 2011.07.17 08:02 #6

Elroch: ...Hence high frequency low quality trades (but profitable on average) can be as good as low frequency high quality trades...

Good comment thanks - I didn't know the maths but I have the experience of this!

Also, too many people expect a system to trade very frequently, with very high PF and very high win/loss ratio...

They also expect a system to work on any pair... I am preparing an article that should explain why that is unlikely!

-BB-

ASCTrend system ASCTrend system Raw Ideas