RRR vs. W% | Forex Factory

Options

Permalinks
Print Thread

Similar Threads

A 1,000 RRR system!? 14 replies

Need help with Reward Risk Ratio (RRR) 7 replies

trendline break/bounce trade System - 1 to 3 RRR 13 replies

Tool which shows RRR and Target/Stop Loss in Money 2 replies

RRR vs. W%

  • #1
  • First Post: May 30, 2010 5:52am May 30, 2010 5:52am
Hello, i have a simple trading system. I want to improve it. My values are:

W% = 57% and RRR=1,6

and i would like to know what is more worthy improving? And i shall also to know what ratio of this 2 index do you use? What number is the best? Some values used by profitable traders? :-)
  • #2
  • May 30, 2010 6:08am May 30, 2010 6:08am
Quoting petertrader
Hello, i have a simple trading system. I want to improve it. My values are:

W% = 57% and RRR=1,6

and i would like to know what is more worthy improving? And i shall also to know what ratio of this 2 index do you use? What number is the best? Some values used by profitable traders? :-)
Are you aware of the Kelly criterion? If not, see, for example, http://en.wikipedia.org/wiki/Kelly_criterion

The data you provide gives a Kelly fraction of 30%, using the formula f = W% - (100-W%)/RRR. This relates to the amount of risk to take per trade as well as the growth rate of your account.

And one answer to your question would be to try to improve the Kelly fraction. Thereby the growth rate of your account will increase regardless of what risk level you are comfortable with.

Hope that helps.
  • #3
  • Jun 1, 2010 7:41pm Jun 1, 2010 7:41pm
Your profit factor is 0.57/(1–.0.57) x 1.6 = ~2.12, which (assuming a valid sample size, i.e. several hundred trades) is extraordinarily good. Congrats!

IMO the right answer is always to continue to improve your knowledge of market dynamics and behavior, and hence the accuracy of your entries and exits. The numbers you quote simply measure the end outcome, and if the proficiency of your trading improves, then the numbers will follow as a matter of course.

The Kelly criterion is a formula used to optimize sizing. Increasing either W% or RRR will increase the Kelly value.
  • #4
  • Jun 2, 2010 8:44am Jun 2, 2010 8:44am
Quoting hanover
Your profit factor is 0.57/(1–.0.57) x 1.6 = ~2.12, which (assuming a valid sample size, i.e. several hundred trades) is extraordinarily good. Congrats!

IMO the right answer is always to continue to improve your knowledge of market dynamics and behavior, and hence the accuracy of your entries and exits. The numbers you quote simply measure the end outcome, and if the proficiency of your trading improves, then the numbers will follow as a matter of course.

[i]The Kelly criterion is a formula used to optimize sizing. Increasing...
Sorry to be a nitpicker, but the Kelly criterion is used to optimise the average percentage gain per trade.
  • #5
  • Edited 11:07am Jun 2, 2010 10:51am | Edited 11:07am
Quoting medici
Sorry to be a nitpicker, but the Kelly criterion is used to optimise the average percentage gain per trade.
If you are going to nitpick, you might want to get it right.
It is used to optimize position size and terminal wealth, which are intimately linked. In order to optimize terminal wealth in terms of the the gain curve, you need to scale the bet size for each trial as some optimal fraction of reinvested principle, which directly corresponds as the abscissa of the optimal terminal wealth ordinate value.

For other viewers, there are two often cited problems with kelly:
1) It is very dependent on actual distribution, and less ideal with normal distributions.
2) Even worse, markets being non-normal, throw fat tails once in a while; which massively alter the curve outcome (which is why many, if not most, traders use a more heuristic based sizing value vs. theoretical optimal f).
  • #6
  • Jun 2, 2010 11:07am Jun 2, 2010 11:07am
Quoting jamjamjam
If you are going to nitpick, you might want to get it right.
It is used to optimize position size and terminal wealth, which are intimately linked. In order to optimize terminal wealth in terms of the the gain curve, you need to scale the bet size for each trial as some optimal fraction of reinvested principle, which directly corresponds as the abscissa of the optimal terminal wealth ordinate value.

For other viewers, there are two often cited problems with kelly:
1) It is very dependent on actual distribution, and less ideal with normal distributions.
2)...
Thanks for adding some detail. Nothing of what you are saying contradicts what I said earlier. Albeit brief, what I said remains correct - by maximising the Kelly fraction you maximise your account growth rate. As you say, and what I briefly hinted at, proportional position sizing needs to be considered in connection with this.
  • #7
  • Last Post: Jun 2, 2010 1:16pm Jun 2, 2010 1:16pm
Quoting medici
Sorry to be a nitpicker, but the Kelly criterion is used to optimise the average percentage gain per trade.
A>C
B>C

is A>B?

might A equal B ? sometimes? always?
0 traders viewing now