Topic: Stagnation as Generator and Optimizer Criteria

Hi Popov,

do you have a plan to add stagnation as generator and optimizer criteria this year?

I still believe that stagnation and max drawdown are the ideal combination to find good strategies.

Best regards.

do or do not there is no try

Re: Stagnation as Generator and Optimizer Criteria

Hello Yonkuro,

This may not work as you expect. If we use Stagnation or Drawdown for a search goal, the Generator will find strategy that makes Stagnation = 0 and Drawdown =0, which unfortunately will not trade.

The best solution is to search highest  profit when DD and Stagnation are set  as Acceptance Criteria.

Re: Stagnation as Generator and Optimizer Criteria

Thank you for your reply, is it possible to set the minimum allowed stagnation to 0.01% so the generator and optimizer won't find any strategies with 0 stagnation.

As for DD I prefer to keep it as acceptance criteria not as optimizer criteria, because it's quite easy to find a strategy with low drawdown.

do or do not there is no try

Re: Stagnation as Generator and Optimizer Criteria

Hi,
a different approach that might be of help could be something like automatically pruning the collection of strategies that don't meet some criteria. I mean, in the collection tab right now we have an option to automatically prune the strategies that are high correlated. But although we have an option to only show the strategies that pass some criteria we do not have the option to automatically prune the collection of strategies if they do not pass the criteria. The difference is that after some time the collection is filled with 100 strategies (no room for more) ordered by one simple criteria (SQN or drawdown or whatever) but only a few of them pass all the criteria that whe have selected. Once we push the button to prune them, only a few remain.
If we had a check to select to automatically prune the strategies that don't feet all the criteria (SQN and DD and stagnation, for example) then after some hours the 100 strategies should meet all these criteria.
It is not the same as being able to search/generate strategies using stagnation or DD, but in the end the results will be similar I guess.
Best regards
RJ

Re: Stagnation as Generator and Optimizer Criteria

Hi RJ,

Thanks for your post, that's what I'm currently doing, I set the maximum DD to 10%, minimum count of trades to 0.25% of bars number (e.g. 150 trades per 60000 bars) and stagnation as pruning criteria on collection tab. But when you're dealing with large data (11 years of H1 data) with special condition (I set 11 pips of spread and 20 bucks commision per trade as stress test), it takes quite long time to get strategies with short stagnation without an optimizer. My record is 6% stagnation using EA Studio portfolio (100 strategies).

Since the first month of using FSB Pro, my goal is to get strategies with less than 30 days stagnation, that means, no losing months, of course the strategies should be robust enough to survive on, at least 10 years market condition. This kind of stratgies may give us more confident and peace in mind, and off course less maintenance.

That's why I believe that stagnation optimizer will really simplify the process.

Well that's what I think, but I maybe wrong, so I'm open minded to any opinion.

Best regards.

do or do not there is no try

6 (edited by rjectweb 2018-03-02 16:20:26)

Re: Stagnation as Generator and Optimizer Criteria

Hi yonkuro,
please forgive my last post. In my post I was somehow asking for a new feature for EA Studio: being able to select if we want to automatically prune the collection and automatically removing the strategies that don't meet all the criteria. We can do this right now but not automatically. Being able to do this automatically (not having to push the prune button) would help us to collect only the strategies meeting exactly our criteria. 

Although I think it would be a nice feature for EA Studio I understand this is may not be what you need. I can visualize those strategies you say with vey low stagnation and drawdown and of course I would like to have a few of them. Having the stagnation in the optimizer could be a nice feature.

Thanks for your comments about how you build your strategies. Those are very instructive. I don't usually use stagnation in my builds but now I think I will start using it.

Lets see if Popov have something to add to these posts.

Regards

Re: Stagnation as Generator and Optimizer Criteria

Hi RJ, sorry for the little misunderstanding, I don't read your post carefully.

Anyway thanks for your contribution.

do or do not there is no try

Re: Stagnation as Generator and Optimizer Criteria

Popov wrote:

Hello Yonkuro,

This may not work as you expect. If we use Stagnation or Drawdown for a search goal, the Generator will find strategy that makes Stagnation = 0 and Drawdown =0, which unfortunately will not trade.

The best solution is to search highest  profit when DD and Stagnation are set  as Acceptance Criteria.

Hi Popov,

just realized an easy solution for this, if we set minimum count of trades criteria to at least 1, I think this issue won't occur right?

do or do not there is no try

Re: Stagnation as Generator and Optimizer Criteria

You are right. We have to add multiple filters in order to make only wanted strategies passing. It is also important to set the search goal. In my opinion Net Balance and System Quality Number work's best, but it is a matter of subjective preferences.