読み取り専用

フォーラムは現在、読み取り専用のアーカイブとなっています。.

バグ報告やプラットフォームに関する質問はこちら → [email protected]

私たちのコミュニティはDiscordとYouTubeで活動しています。ぜひ参加してください!

ご参加ください Discord YouTube

weighted fitness problem

6件の返信

チャールズ・ウマニ

購読者、bbp_participant、36件の返信.

プロフィールを表示

6年前 #258373

こんにちは,
I have a big problem with fitness.
Here is my case, I check weighted fitness and use only 3 criteria.
But during the generation I find strategies with a much higher fitness while all the criteria are worse than others …

where is the problem ? i don’t understand…

thanks by advance !

添付ファイル:
に違いない ログインしました 添付ファイルを表示するには.

0

Tomas Matejka

管理者、sq-ultimate、2件の返信.

プロフィールを表示

6年前 #258378

Thanks for the report, we will need to verify this. Also do not forget to set those data columns to IS (In-Sample). The fitness value is calculated based on in-sample results not out-of-sample or full-sample

添付ファイル:
に違いない ログインしました 添付ファイルを表示するには.

0

チャールズ・ウマニ

購読者、bbp_participant、36件の返信.

プロフィールを表示

6年前 #258473

thanks, but i’ve made my test without OOS.
on build 128, whatever  i use in “compute from” the result is the same…

 

0

チャールズ・ウマニ

購読者、bbp_participant、36件の返信.

プロフィールを表示

6年前 #258474

my fitnes is ALWAYS calculate on calmar ratio. have you the same problem ?

0

DarkHope

定期購読者、bbp_participant、コミュニティ、14件の返信。.

プロフィールを表示

6年前 #258489

I had this exactly debate with Mark many years ago, and the problem is that, the fitness function is weighted on the trades (if i’m not wrong, is weighted over 100 trades).
You can actually see it, strategies with high number of trades have bigger fitness function under the same evaluation criteria. Once he said me: “If you need i’ll make a custom version for you”, i think that, right now is more proper make a feature to enable or disable this limitation, open to everyone. IMHO.

0

チャールズ・ウマニ

購読者、bbp_participant、36件の返信.

プロフィールを表示

6年前 #258490

hi, i understand but , for me, number of trade have never change anything on my fitness :/

Not in this case, but, under 100 or over 100 trades i’ve never noticed any difference in the fitness calculation.

Look at my problem, whatever i use as parameters , le ranking is always exactly the same of calmar ratio… 🙁

0

Tomas Matejka

管理者、sq-ultimate、2件の返信.

プロフィールを表示

6年前 #258798

There has a been a modification done in the way the fitness value value is being stored with a strategy file. That should be available with the next SQ update. Then sorting will work as expected

Just to let you know our forum is not an ideal place to request features and report bugs. That’s what we have our roadmap for. If you have an idea for a new feature or find an issue using StrategyQuant please use our tool to report bugs and requests: https://roadmap.strategyquant.com When reporting a bug, please, provide also your strategy .SQX and the build configuration used as a saved CFX file – if it is related to build.

 

0

6件の返信を表示中 - 1 - 6件目 (全6件中)