Does this OptF seem correct?

Gamestudio Links

Zorro Links

Newest Posts

Zorro 2.70
by jcl. 09/29/25 09:24

Release 2.68 replacement of the .par format
by jcl. 09/29/25 09:21

optimize global parameters SOLVED
by dBc. 09/27/25 17:07

ZorroGPT
by TipmyPip. 09/27/25 10:05

best practice for futures historical data file organization
by purpledawn. 09/23/25 01:38

Discrepancy between Zorro and Visual Studio results
by dBc. 09/21/25 13:11

assetHistory one candle shift
by jcl. 09/21/25 11:36

Plugins update
by Grant. 09/17/25 16:28

AUM Magazine

Latest Screens

Who's Online Now

1 registered members (TipmyPip), 18,618 guests, and 5 spiders.

Key: Admin, Global Mod, Mod

Newest Members

krishna, DrissB, James168, Ed_Love, xtns
19168 Registered Users

Print Thread

Rate Thread

Does this OptF seem correct? #430259 09/26/13 10:05 09/26/13 10:05
Joined: Jul 2013 Posts: 522 D dusktrader OP User
dusktrader OP User D Joined: Jul 2013 Posts: 522	In a test strategy I'm working with I get the following OptF results after adding Margin and reinvestment: Quote: Annual growth rate 129% Profit factor 1.67 (PRR 1.34) Sharpe ratio 0.85 Kelly criterion 0.15 Ulcer index 14% Prediction error 42% Cycle performance 1.84 1.83 1.79 1.45 1.64 1.83 1.69 1.82 Portfolio analysis OptF ProF Win/Loss Wgt% Cycles EURUSD .505 1.82 53/229 134.9 X//X EURUSD:L .164 1.32 34/155 4.0 \/// EURUSD:S .999 1.86 19/74 130.9 ///\ I noticed in the testtrades.csv log, it is calculating lots for Shorts far bigger than Longs, sometimes 10x more. That may all be correct (if so, that's what I want). But it seems abnormal that OptF is .999 -- can you confirm that is correct? Here is the code I'm using for reinvestment. Please advise if you have any suggestions here. I see various code for this around the forum and help pages, but this seems to be one of the more recent calculations: Code: //reinvest a portion of profits int reinvestprofits = 1; //invoke margin setting during trade logic Capital = 1000; //simulate starting account with $1000 var MarginLong = OptimalFLong * Capital * sqrt(1 + (WinLong-LossLong)/Capital); var MarginShort = OptimalFShort * Capital * sqrt(1 + (WinShort-LossShort)/Capital); Thank you

Re: Does this OptF seem correct? [Re: dusktrader] #430264 09/26/13 10:25 09/26/13 10:25
Joined: Jul 2013 Posts: 522 D dusktrader OP User
dusktrader OP User D Joined: Jul 2013 Posts: 522	I guess the reason I'm confused on the OptF calculation is... If overall strategy is .505 with 1.82 ProF, how can short have a similar ProF of 1.86 but so much higher OptF?

Re: Does this OptF seem correct? [Re: dusktrader] #430268 09/26/13 10:43 09/26/13 10:43
Joined: Jul 2000 Posts: 28,024 Frankfurt jcl Chief Engineer
jcl Chief Engineer Joined: Jul 2000 Posts: 28,024 Frankfurt	OptF is not equivalent to ProF. If it were so, you could use the ProF instead of OptF, which is much harder to calculate. The .999 looks correct. OptimalF factors are calculated from the whole balance curve, which the ProF is only the win to loss ratio. Balance drawdowns affect OptimalF, but not ProF. We also used various formulae for calculating the margin from OptimalF, but what you're using above is the best method.

Re: Does this OptF seem correct? [Re: jcl] #430270 09/26/13 10:51 09/26/13 10:51
Joined: Jul 2013 Posts: 522 D dusktrader OP User
dusktrader OP User D Joined: Jul 2013 Posts: 522	Ok, I understand what you are saying, that OptF and ProF are not comparable metrics. But can you confirm .999 is not a glitch? It seems unlikely that .999 would ever be assigned. (I think) that says that 99.9% of the margin should be devoted to this component.

Re: Does this OptF seem correct? [Re: dusktrader] #430275 09/26/13 11:31 09/26/13 11:31
Joined: Jul 2000 Posts: 28,024 Frankfurt jcl Chief Engineer
jcl Chief Engineer Joined: Jul 2000 Posts: 28,024 Frankfurt	Yes, it is not a glitch. But it is indeed a special value. It sometimes happens that the backtest balance curve has very little drawdown. Then the maximum gain can be achieved when assigning the full capital to that component. OptimalF is then at 0.999. I would here not use separate OptimalF for long and short, but one OptimalF for both, even if this reduces backtest profit.

Re: Does this OptF seem correct? [Re: jcl] #430294 09/26/13 13:56 09/26/13 13:56
Joined: Jul 2013 Posts: 522 D dusktrader OP User
dusktrader OP User D Joined: Jul 2013 Posts: 522	Perhaps I could use separate OptimalF factors for long and short, but limit them with the strategy's overall OptimalF value. Like this: Code: //reinvest a portion of profits int reinvestprofits = 1; //invoke margin setting during trade logic Capital = 1000; //simulate starting account with $1000 var MarginLong = OptimalF * OptimalFLong * Capital * sqrt(1 + (WinLong-LossLong)/Capital); var MarginShort = OptimalF * OptimalFShort * Capital * sqrt(1 + (WinShort-LossShort)/Capital); I'd like to be able to use separate figures because in some cases one of the components is really bad, as seen in this example.

Re: Does this OptF seem correct? [Re: jcl] #433697 12/04/13 16:27 12/04/13 16:27
Joined: Jul 2013 Posts: 522 D dusktrader OP User
dusktrader OP User D Joined: Jul 2013 Posts: 522	Hi jcl, I came across this OptimalF value of 0.999 again. I don't fully understand why this happens. I am concerned that it is not normal, and that it could be caused from something I'm doing incorrectly. See below on USDJPY which shows .999 for longs and .000 for shorts. I'll post here also the USDJPY chart for this strategy. Thanks for taking a look at this. Code: Portfolio analysis OptF ProF Win/Loss Wgt% Cycles AUDUSD avg .167 1.35 1148/3094146.4 //X/ EURAUD avg .726 1.63 537/1673 381.9 //// EURCAD avg .328 1.27 1557/2572146.9 /\// NZDJPY avg .098 1.39 2588/2829376.2 \/// USDJPY avg .499 2.72 652/752 448.6 /\// AUDUSD .139 1.35 1148/3094146.4 //X/ AUDUSD:L .116 1.37 823/2071 94.9 //// AUDUSD:S .218 1.33 325/1023 51.5 //\/ EURAUD .612 1.63 537/1673 381.9 //// EURAUD:L .459 1.35 334/1119 108.9 //// EURAUD:S .994 1.91 203/554 273.0 //// EURCAD .339 1.27 1557/2572146.9 /\// EURCAD:L .374 1.35 1104/1584122.8 /\// EURCAD:S .283 1.13 453/988 24.2 /\// NZDJPY .117 1.39 2588/2829376.2 \/// NZDJPY:L .196 1.39 2588/2829376.2 \/// NZDJPY:S .000 ---- 0/0 0.0 .... USDJPY .999 2.72 652/752 448.6 /\// USDJPY:L .999 2.72 652/752 448.6 /\// USDJPY:S .000 ---- 0/0 0.0 .... EDIT: I should note also that this may be normal/correct. In looking at the testtrades log, I see that Zorro took the maximum Lot size on every USDJPY long trade as indicated above. However, even the worst losing trade was only a 6% loss (and far less than the MAE %). Attached Files Last edited by dusktrader; 12/04/13 16:45.

Re: Does this OptF seem correct? [Re: dusktrader] #433713 12/05/13 07:45 12/05/13 07:45
Joined: Jul 2000 Posts: 28,024 Frankfurt jcl Chief Engineer
jcl Chief Engineer Joined: Jul 2000 Posts: 28,024 Frankfurt	Factors close to 1, as for USDJPY:L and EURAUD:S are unusual, but can happen. Aside of future peeking or overfitting, I cannot think of a method to do something incorrectly for getting too high OptF factors.

Moderated by Petra