Forum Replies Created
-
Posts
-
Hey Steve,
I’ll take a second look at your post tomorrow as there is a lot to digest in what you are trying to do.
A couple of initial points though.
One is that when Petko suggests checking with 200k bars if have generated using 100k bars this allows a check of OOS BEFORE the insample data as opposed to after the OOS which is done by setting OOS % in generator.
I agree also to stick with Monte Carlo only for robustness in the Reactor. You could increase it’s effectiveness further by increasing the number of tests, or afterwards running it through the validator a second time perhaps but for me it’s the best robustness tool.
I also never really using the optimiser but in this case I can see how it probably is useful for what you are doing. As you have so few indicators and no exit indicator along with not allowing optimization of stop loss and take profit I think it’s okay and has a much lower risk of over optimization than normal but generally I would avoid.
In this specific case I would personally either use only the optimiser in the reactor as you have the setting OR use the normalizer one the reactor as Petko suggests and then for each strategy optimiser only the time entry indicator with reasonable size steps (The more time consuming option but the better of the two in my opinion).
So ideally in a nutshell: Reactor with normalizer and Monte Carlo and then optimize only the time setting on an individual basis with hourly or two hourly steps for example.
Also check out the time setting in tools for either control of when your strategy trades.
Hope that helps
Hi Alan,
Yes that is what Petko is saying and it is definitely the case.
Make sure that you close other unneeded applications and browsers when running the reactor. Also take note of the number of strategies generated and not just the number making it into the collection if you’s want to gauge how optimally EA studio is working.
My rough rule of thumb is to run a separate window running a single instance of EA studio per processor core.
Also make sure that your browsers are active, you cannot minimize them. This is the main reason I want to see a screenshot actually, I want to see how many strategies have been generated rather than added to the collection. I suspect the window may be going inactive.
After an overnight run the number of generated strategies should be many hundreds of thousands and maybe even over a million depending on computing power.
Hopefully that’s the issue as I really don’t think it’s to do with the account settings like I originally suspected.
Oh and always use chrome of course :-)
Hi Alan,
I noticed that after I had sent you the settings so change to 20% for normalizer.
Also my reactor using your settings generated plenty of strategies throughout the day? Can you confirm that you are now generating plenty strategies with these settings (About 40-50 over night?). Hopefully the issue is resolved now? If not then please send me a screenshot of your reactor page after running all day. Cheers Sam
Hey Alan,
I checked my reactor this morning and had 42 strategies in collection. That’s from running the same settings as you but I changed the account settings to default.
I have set it running again with your account settings so will update you tonight.
meantime I would suggest using the default account settings as this will be absolutely fine as long as you are trading a portfolio of say around 10 EAs (which you should be).
Also you setting look good btw, except that I would change the normalizer to 20% OOS to match your generator.
Optimal pack is Avery good choice btw Alan. I’d focus on EA studio for now but it’s excellent value to get FSB pro for the extra cost as you develop as a trader you will be glad to have both options :-)
Hmm, I’ll take another look in the morning Alan. That was the thing that stood out but perhaps I was too hasty in thinking that was the issue.
As to how those changes would affect the simulated trading, shouldn’t be a problem as long as a semi sensible portfolio.
I’ll actually quickly set it running on my computer and see how it looks in the morning also okay ;-)
Also please keep yours running overnight and I I’ll likely want a couple of screenshots from you in the morning to compare. I just loaded that file to Ea studio and hit run on reactors and in 3 mins have 2 added to the collection, will see how it looks in the morning though.
Hey Alan,
I checked out your settings file and I see the issue :-). You are using a 0.1 lot size on an account that is too small. Presumably this is the account settings you intend to trade so you should definitely be setting the lot size in the strategy properties to 0.01. I would also increase the range of the take profit and stop loss a bit.
I will email you back the modified settings file that added a couple of strategies to the collection in about 5 mins.
I’d probably recommend that you increase the account size/ leverage a bit during the generation and then can check on your account size/leverage using the validator. This will help you get a feel for the threshold values that are sensible.
I personally just use a 10k account with 100:1 leverage using a 0.1 lot size (pretty much the default) during generation.
Also in the data horizon uncheck the start date box and just have the bars set and the end date set. That’s just my preference though. Hope you checked out that thread and video also ;-)
Let me know how you get on.
Ill be able to take a look in about an hours time
Cool, I’ll need you to send me your settings file first though? Have you downloaded it yet? Once you have sent me the file definitely check out those links I sent you also. Both are really very useful when starting with EA studio ;-)
Feel free to either attach the file here or just email it to me. I’ll get it quicker with email though ;-)
Also I would strongly recommend you check out this video https://youtu.be/z8FF1Uefdpo as a further check of how well EA studio is doing it’s job :-)
I know there is just something off with this particular issue in the settings. Can be a bit slow resolving this way but as a long term user I’m sure I will be able to get to the bottom of it quickly once I upload your settings and take a look.
Hope the above forum thread and videos are helpful.
Hi Alan,
100% EA studio is working correctly. But there is definitely some input that is off somehow.
I think the quickest and easiest way for me to get to the bottom of it at this point is for you to send me your EA studio settings .json file?
Are you able to attach it? Just go to tools and save settings and you will get a json file.
Having been through this process several times myself and also helped others, there is often some doubt with EA studio at the start and then later zero doubt and shown to be user error. It’s remarkably clever and highly quality assured software without doubt :-)
Take a look at the thread below of someone having total dispair about Ea studio not working properly and then working through with support to find that all is well:
Send the settings through and I will take a look at what’s going on tonight and get back to you ;-)
Hold on, lowered minimum stop loss and take profit from 10 to 5?? Set your ranges for minimum stop loss and take profit from 10 – 200.
Also for acceptance criteria just start with a simple minimum 300 trades and minimum profit factor of 1.2 for the complete sample. If using OOS then set the IS and OSS min profit factor to 1.1.
Also this seems to be a generation issue so don’t worry about running Monte Carlo or anything in the reactor. Just use the validator after you have a good collection to do this.
let me know how you go?
Interesting. All I can think of is that for M15 it’s not managing to find strategies because it is simply more difficult for this data, perhaps it’s just more noisy on this timeframe over the last 5 years (which it will be) and harder to predict.
I understand it’s a little unintuitive but it doesn’t mean that there is a mistake really. Worth asking questions but if after checks everything is as it should be then we can only assume it’s just harder to find strategies passing the acceptance criteria on this timeframe.
On the one hand it takes longer to generate strategies for M15 as there are 4x as many bars but then on the other I would have expected that it would have been easier to pass the minimum trade acceptance criteria but it must be struggling with the others.
What other acceptance criteria do you have set? I would leave the minimum count of trades at 300 but relax the other acceptance criteria a little bit so that you get some strategies in the collection and then filter from there.
Much better after an overnight generation to have a couple of hundred strategies in the collection that you can trim down using the performance filter than just have a handful in the collection.
Hi Alan, yes I would have expected it to be easier to pass the minimum trade count acceptance criteria on M15.
First are you absolutely sure of these three things:
1. The EA studio settings are identical except for the timeframe
2. This is for the same symbol
3. That you have all the bars for the full 5 year days range for M15 and there are no gaps (use the check in the data statistics to confirm this)
Once you have confirmed these I will be able to help more.
Hey Alan, not sure what you mean by accurate but I’m assuming you mean reliable and robust.
To be able to have any kind of faith in the strategy it’s important that it has been backtest over a significant period and has made a significant number of trades in that period.
Personally I use about 5 years data for both M15 and H1 and set the minimum trades to 300. Obviously this will be 4x more bars for M15 so will take longer to generate strategies, but it is also easier to pass the minimum 300 trade criteria for M15.
Hope that answers your question? Using less is fine as is using more, that’s just what I use ;-)
Also to get this amount of data use the Premium data or historical data app
Hey Matthew, you are welcome re my suggestions. Glad they helped :-)
Requiring 300 trades over 2 years data will definitely be much easier to pass for M15 than H1 so that makes sense. If you want to get strategies for H1 also then I would be increasing the number of bars used in backtesting to be closer to what worked well on M15.
Not sure re NZDUSD, maybe try initially relaxing the acceptance criteria to get 100 or so eas into the collection and then investigate in the collection what seems to be the criteria that it is struggling the most to pass and make a decision about how best to proceed from there.
Well done on the great results so far. Good to see :-)
