- This topic has 191 replies, 1 voice, and was last updated 2 years, 8 months ago by
.
-
Posts
-
-
I use :-)
Maximum equity drawdown %: 10-15% (Depends on FX 10% or Crypto 15%)
Minumum Trades: 50
and Sometimes:
Maximum stagnation %: 30%
Minimum return / drawdown: 3
-
@dommech-Hey there! Good question. My acceptance criteria is:
Min count of trades: 10
Max equity drawdown%: 20
Min net profit: 10
Min backtest quality: 98
Min win/loss ratio: 70
Max stagnation%: 30
and sometimes Min profit per day: 1 *Of course this depends on the lot size being used for generation of strategies.
If I can’t get the generator to produce strategies with all of these requirements, then I just run it without acceptance criteria and use these as the performance filters. I use the criteria for all assets that I’m trading. After the the strategies are all created, I then do some additional analyzing to pick out the best strategies to use such as using the optimizer, turning on the stop loss or take profit, etc.
-
Hello Dommech,
something important from me here:
Min count of trades:50 is too small number. Specially in Forex you should go with min 300-400 count of trades.
Basically when the generator/reactor works it looks for combinations between different entry and exit conditions, indicators with different parameters etc., so when you have strategy with only 50 count of trades this might be not a stable strategy. We want to see bigger count of trades so we know that this combination of indicators, parameters, SL, TP and so on, works on more trades(longer period of time).
-
I really like the win/loss ratio, but it is hard for me to get results on all pairs when I keep it on 70.
In the acceptance criteria for min count of trades, Petko is right, I even try to keep it above 600 trades.
-
Thx for the answers guys!
Some strategies with especially extremely low DD (0.5-2%) I’ve found that they are extremely rare.
My Backtesting data is over 3 years, always when using FX I’m thinking of adding it to 5 years, my thoughts are that market behavior changes and 3-5 years is enough? Or how do you guys think?
Usually I get 200 trades/strategy but if I set it higher I might miss the rare ones with the low DD
I understand that this affects the robustness of the strategy if it’s only 50 trades but the low DD makes me want to take the risk
Off Topic, but:
How far back do you guys go with FX, how many years or bars do you use to get enough statistic data?
-
-
@Petko and @Haliffa-Does the chart timeframe that we are using affect the count of trades? For instance, I can generate a lot more count of trades on the M15 chart vs. the H1 chart. But then again, I get into the issue of wanting that 70% winning performance and only getting a smaller count of trades. I wonder if this is the trade-off (higher winning percentage=smaller count of trades vs. lower winning percentage=larger count of trades).
Since I don’t want to give up on that winning percentage, I may be working with short-term strategies for quite a while.
-
@dommech-Yeah, I leave the generator more wide open and once it’s done or as I filter is when I get really picky about the strategies, just so I don’t miss those nice and very low DD strategies.
For FX, I am going back 5 years. Now that we are in 2019, I changed my data horizon to start at 1/1/2014. I do this because it gives me enough bars as I am trading H1 and M15 charts. For the lower timeframes, 200k bars should be good enough to get some good strategies.
-
@jacpin thx sounds good, I go usually for M15 H1 and sometimes H4 but not lower
Will change to 5 years as well :-)
I recently heard in one of Petko’s courses that instead of demo testing the files you could (optional!) Change the data horizon and leave 1 month out of the generator and later recalculate. this gives me even more confidence that the strategy is robust. I still put many EA’s on demo so exchange the bad ones
-
-
Hello Jacpin,
You are getting more count of trades on M15 probably because you have more data on it ( I guess you have the max for EA Stduo 200k there?).
So basically the more bars we have the more count of trades we can expect to see. Of course it depends how you set the generator/reactor, acceptance criteria, etc.
Just be careful if you use small count of trades and higher win/loss ratio, not to go in over-optimized strategies. Make sure that you do the Monte Carlo test and Demo testing (I know that you do demo).
-
Hello Dommech,
Glad to see you are working in more details. Yes, this method of removing the last one month is really great.
This way we see if the strategy was going to be profitable if we really traded with it. It saves a lot of time, and as you said, it gives confidence.
Both strategies look fine.
-
@Petko and Dommech-Can you let me know which course has this method to leave off the last month in the Data Horizon? I want to understand the steps on how to do this and see if it will really save me some time or provide me with more confidence in the strategies that I generate.
-
Hello Jacpin,
Glad to hear from you. I show that in the The Ethereum trading in 2018 + 99 Robots every month course, but I have it in my “to do” list where I want to dedicate a whole course about it. I will show different methods on how to avoid the Demo trading.
-
@Petko-You are awesome! If you can please do that I would be so grateful. Sometimes, the waiting for my EAs to pass demo is frustrating to me. But I don’t want to take unnecessary risks. :) As soon as you have anything that shows how to avoid demo trading, then sign me up!!!
-
I will use Demo as well but I will not Demo test my strategies before going live, I will still continue to demo test “backup strategies” if I need to switch out some of the ones. Testing this way on “Out-Data” work as well as demo testing. Markets change anyway so it’s not a guarantee that the strategy will work live if it did on demo.
I go with enough history data to feel confident + 80-90% Monte Carlo + the last step is the one month of “Out data” I don’t feel that I need more confirmations than that, this way really saves Weeks or even months so I’m willing to take the risks.
Hey, I’m compounding profits so lets start to compound time ;-)
-
Jacpin,
this is my job to test, trade, and find new methods which are faster. It is great that there are such softwares nowadays which allows us to be innovative.
-
Hello Dommech,
your method is just fine I think, very similar to what I do. Also, you are right that there is no guarantee, and I think I make this clear in all of my courses. We never know if a profitable EA will continue trade well. It is just the same as to know where EURUSD will go tomorrow.
But the thing is we wan to put more things on our side before trading, anyway we would be gambling, and you will never see me in the casino :) No way!
-
Hello,
I want to ask a question.
In one course I saw that in the acceptance criteria Petko uses the ambiguous bars, and I do not see those in the acceptance criteria.
-
@Chingi-EA Studio went through about 2 updates from the time I have been using it in the last 4 months. Petko is probably going to be updating that course or making a note about the software changes between the time the course was created and now. Instead of ambiguous bars, I use Minimum Backtest Quality and leave it at the default of 98%. Here is the information from the Help tab on EA Studio about this criteria:
-
Hello Jacpin,
thank you very much for making that clear.
Actually Petko explains the same in the course about the ambiguous bars, but I was just thinking that this option is no more available.
So it is good to know it is there, just with different name.
-
Thanks for answering the question, Jacpin!
Yes, Chingi, it is the same filter.
Cheers
-
I like to use it too! And if one is developing the EA by coding, will never know this option exists :)
-
I am used to place the max consecutive losses in the acceptance criteria.
I think it is great filter because this way we know how many negative trades to expect during the trading.
also if there are more during the demo trading, than I know it is over optimized Expert Advisor
-
I’m currently testing from 2005 on the H1 for a few of my pairs (roughly 90k bars). I’ve been looking into using 20% OOS and then keeping the last 6 months of data in my back pocket as a final test after Monte Carlo. Today I just threw some GBPJPY strats on the demo account for testing, will let you know how it goes!
P.S. only 2 made it onto my demo account after 20 hours of scanning. Only the strategies that passed the 20% OOS and the 6 months visual check.
-
Hi Roman,
It is good that you test with the OOS and 6 months at the end. However, you are using huge period, so you can afford it.
Keep in mind that the more Historical data you have, the profit per day becomes smaller. That is why to look for this “golden middle” that will bring you you robust strategies with a nice profit.
-
Hey guys,
do you know that now it is available to use up to 500k bars in the Data in EA Studio. That is only for the licensed traders by the way.
Have someone tried that?
-
@Chris-I did not know that. That is a lot of bars. For H1, I think going back 5 years hasn’t even hit 200k bars yet for me. So, 500k…. that is a lot of bars. I wonder if that is more for the lower timeframes like M1 and M5 where 200k bars will only let you go back a couple of months. 500k might be able to give a full year on M1 or M5…or at least close to it. I’m gonna check it out.
-
Hello,
this increase of the Historical bars is still in test, and I am not sure if it will last.
Anyway, if you want to test it, you need to change the number from the Data Horizon.
There are many updates now in EA Studio, and more are coming up.
So what I will do is to open a new topic in the forum, where I will post videos with every update. This way everyone will keep it up with the updates, and I will explain them in details, because in the help of EA Studio they are written but not really explained.
-
Hi all,
I wish to know if someone has a strict system when generating strategies with EA Studio or FSB pro about the acceptance criteria?
What I have learned from Petko is to reduce the acceptance criteria if I can not get any strategies and to increase it if I am generating too many.
The thing is that with every asset it is different. I mean if I found good acceptance criteria for EURUSD, this does not work on GBPJPY for example. So I had to lower the acceptance criteria there.
Any suggestions on that?
-
@Maria-Hi there! I know what you are talking about. I did have to change the acceptance criteria in order to generate strategies, but then I would use the performance filers in the collection with the same acceptance criteria I would have normally use to get the strategies I want. That was the only around that issue that I found.
-
Hi Maria,
better to generate more strategies and than filter them in the Collection.
The generator shows different and unique strategies all the time, so you really can not predict what it will show in 100%.
That is why you better give it a little bit more space, and than you just filter the strategies you want to have using the filters in the collection.
-
Hello Maria,
one thing to keep in mind here is that the brokers offer a different number of Historical data bars with different assets.
Have a look at your data, and you will see that for EURUSD you will have more bars that for AUDUSD for example.
This difference could change the performance of the generator when using the same acceptance criteria which is logical.
Now when you get used with the EA Studio and you generate daily strategies you will know what acceptance criteria to place for any asset quickly.
Not a bad idea to make an excel table where you place the currency pairs with the acceptance criteria and the values that worked best the last time.
So when you run a generator for any asset you can have a quick look at it, and see the latest acceptance criteria that you used. It will save you time.
-
Hello Jacpin,
I get your point. So you just generate strategies with fewer criteria and than you make it strict with the filter in Collections.
Now, I am concentrated a lot on the new feature R – squared and I think I am trying to make my strategies with too good equity lines which is not exactly what I should be doing :)
-
Hey Andi,
this is exactly what I am planning to do. I think jacpin said the same. Just generate with not so much strict criteria and then make it strict from the filters in the collection.
Petko is doing the same in his courses, so I do not know why I lost a couple of days to struggle with that.
-
Hey Petko,
Your idea is just great. I am getting on work to start building such a table.
You are right that it will save me time just because of generating strategies for the same currency pairs will be easier.
With the time I will just make the acceptance criteria stricter.
-
Hello traders,
I think it will be great if in this topic everyone shares what acceptance criteria one uses. This was we will have an idea what to test as acceptance criteria because obviously running a generator without good strategies at the end is what we do not want to have.
So here I will start:
Currently, I use:
Profit Factor 1.2
R – squared – 70
Min count of trades 500
Stagnation in % – 15%
-
Here are my Acceptance criteria at the moment for the Forex pairs:
R – squared – 80
Min count of trades 450
Max consecutive losses 10
Backtest quality – 98
-
Here is what I use in the acceptance criteria:
Minimum bars in trade – 30%
Min profit factor – 1.2
Min months on profit – 80%
May I ask what is the R – squared? I see that you both use it?
-
Hey Viktor,
The R-squared is an update we have with EA Studio. It helps us generate strategies with better equity lines.
I have opened a topic in the Forum called EA Studio Updates – have a look at it, there is video and more details about R-squared.
-
R – squared is great improvement, I think it improved a lot the generator because it chooses more stable strategies
-
Min Trades:200
R-Squared:75
OOS Stability:70% -
After the video of Petko, I am planning to start using the OOS. It seems like really nice and useful. SImply, we do not use the whole period that we have but we use chosen % of Historical data. After that, we see the results that would have happened in the last months. And with the update int he acceptance criteria, we get only the good strategies.
-
Anyway here is what I use:
Min count of trades 450
Profit factor 1.2
Max consecutive losses 7
Backtesting quality 99%
-
Hey Andi, you go it right, with the update of EA Studio acceptance criteria we can filter the separate sectors as In Sample and Out of Sample. That allows us to generate strategies that perform robustness tests as Multi Market and Monte Carlo in the In Sample part and in the same time those that pass the validation criteria in Out Of Sample part.
-
Hola Andi! I’ve always used OOS, I was testing another software at the same time as FSB in 2017. With the search style they use OOS is very important. With it I still keep some of my data in my back pocket (about 3-6 months) just to really simulate and confirm simulated trading. This ensures at least for me the strategy has some strength I can depend on. Doing my method I may only find 1-5 strategies a day, but it helps me feel more secure with them. Even though we are testing on demo first, this keeps my demo free of clutter. Saves me time, space, and emotional excitement lol.
-
Hey Roman,
you are very right here, OOS is very important, but once again the update now changed it a lot. We have the chance to choose acceptance criteria for the In Sample and for the Out Of Sample. Before it was for the complete backtest.
Personally, before the update, I did not use a lot the OOS. Simply because we were choosing the acceptance criteria for the complete backtest. So no matter what was OOS showing we were able to decide which strategy to use based on the whole test. Now we have it separately. This way we can choose what criteria to have in the generated part(In Sample), and what Training part(Out of Sample) to choose.
With simple words, now we really have the chance to simulate Demo trading with OOS.
-
As said I tested the OOS, and it worked just really great for me. The strategies are stable, they pass the robustness tests, and now I am on a nice profit on the Demo account, so I am sure that most of the EAs will go to my live trading portfolio.
Thanks, Petko! You really improve the systems and share with us! You are the man!
-
Hey Andi,
I am glad to hear that my videos improved your Experts trading.
I follow along all updated with EA Studio and test them when I find the method that works the best for me, I will always share it.
Anyway, OOS is great to be combined with the acceptance criteria.
-
Is there any correlation between the Minimum count of trades and Monte Carlo?
-
Hello Stephen, I can say there is. The more count of trades a strategy has, the more chances you have to pass the Monte Carlo.
Simply, if you have a small count of trades, you are getting closer to over-optimized strategy. The more trades a strategy has, the more robust it is. The strategy was tested more times in the past, and there is a higher chance to continue profit.
-
Hey,Pekto!
Thanks for the answer. I was guessing that the more count of trades we have, the better.
I saw that the strategies having more count of trades pass the Monte Carlo.
However, I think that EA Studio is capable to find great strategies, and especially if we succeed to filter them well.
Thanks for your help as always!
-
@Petko – Sorry for the delayed response. Yes I really do treasure the OOS, especially the acceptance criteria control of both in and out of sample is extremely useful! Finally got my new computer yesterday which is why I’ve been radio silent. Looking forward to cranking out more strats!
-
Hey Roman,
Glad to hear from you again. Yes, the new acceptance criteria really changed the usage of OOS. Before that, I was not using it…
-
Guys, when you use the win/loss ration what value are you looking for in the acceptance criteria as a minimum?
-
@Thapelo-I’m using 0.70 as my minimum in the acceptance criteria and only under the complete test part.
-
Hey Jacpin,
you do not use the OOS at all? How is the win loss ratio calculaed? Isn’t it the profit divided by loss?
-
I think it is calculated when you sum the Winner trades and the losers trades, and then divide by the winners…not 100% actually.
-
Hey guys, it does not really mean how it is calculated. I know that it is a ratio between winning trades and losing trades :)
But if your SL and TP are different(and they are) this will mean that these criteria could be misleading.
For example, if you have SL 2 times bigger then the TP, it will mean that one negative trade will equal to 2 positive trades. And it will be different for any strategy…
-
Good point Andi, I observed that my win/loss ratio is only high (>0.75) when SL is very close to or bigger than TP. Sometimes, those strategies with SL 2 times bigger than TP can give very nice equity line but it will require a lot of wins to recover one negative trade.
-
Hey guys, that is why I use more the profit factor. It is not connected to the number of profitable trades and losing trades but to the results itself. At the end of the day, the important is the results, not the number of trades.
-
Hey Ossaio,
yes, this is true as Petko confirms. It is nice to use the win/loss ratio but it is better if you combine it with some other criteria.
-
For sure, Andi. We always need to combine different criteria.
Now, when trading on the live account, you can use one strict rule/criteria to follow the trading on the live.
I think Jacpin was using exactly the Win/loss to monitor the EAs on the live which makes a lot of sense.
-
@Desita-Sorry, it took so long for me to reply. I was having issues logging in and for some reason, the forum wasn’t allowing me to respond. But now I can. I use the win/loss ratio as 0.70 (meaning 7 out of 10 trades are winners) I only will accept 3 losses out of 10 trades. This is just my conservative way of trading. But after watching Petko’s courses on the updates, I now use the win/loss ratio of 0.70 for the complete backtest as well as a minimum count of trades. I use the OOS just the way Petko shows us as well as the R-squared for optimization. After doing this, my strategies have become so much better.
@Petko-You’re absolutely right. I still do. It is a lot easier for me to analyze an EA results in FX Blue Live with just monitoring the win/loss vs, anything else.
Hope everyone else is doing well and happy trading! My commodities are on a roll lately and my forex pairs with GBP and JPY are doing fantastic this month!
-
Hello Jacpin,
Glad to hear that the recent videos that I recorded improved your strategies with the OOS and the R-Squared.
I think your system to manage the live account with win/loss ratio of 0.70 is just great. Keep up the good work!
-
Hey Jacpin,
thank you very much for the replay. Your system is really nice. I will give it a try during the next month. It makes a lot of sense really.
And what is your system to test the EAs on Demo, or you do not? Do you only use the system of Petko for OOS?
-
@Desita-Hey there! I honestly tried to get out testing on Demo and only using Petko’s system of the OOS, but I just couldn’t confidently let it go. I guess this may be considered overly cautious, but I really don’t like waste time or money. So I do a combination of both.
I generate the strategies using my broker’s live historical data and include the updated criteria that Petko showed using the R-squared and 30% OOS. I then take these strategies and run them through the MT strategy tester (since I am using MT5, it might be a bit different for those on MT4). I analyze the last 10 trades to see if the win/loss is still at 70% or above.
For example..strategy generated and ran through tester shows a total count of trades to be 125. I go back to trade 110 (not 115) to analyze and see if 70% win/loss performance is maintained. From there I check the win/loss performance of trades 120-125. Lets say it is 4-1 in winning trades at this time. I then put this EA in demo for the next 5 trades to complete another 10 trade analysis. If it maintains losing more than 2 trades out of the next 5 trades, then this has maintained the 70% winning performance and passed demo testing. Potentially this is an EA that I could start live trading. There are other factors I look at as well before immediately putting them on, like profit factor, # of pips moved, etc.
Let’s take this same scenario and have it start off as a losing strategy, for instance it is sitting at 2-3 in winning trades. This means that it cannot have any more losses out of the next 5 trades to pass my demo criteria. Sometimes they pass and sometimes they don’t. But I have found that being this strict has helped me to avoid some major catastrophes.
Hope this helps.
-
Hello Desita!
Thank you so much for making this so clear. I see that you are very strict in your testing which is absolutely respectful.
I like that you look at the recent trades. I also do. I avoid placing EAs which equity line goes down at the end after the generation. But it is not really strict and precise to look at the chart. Probably the number of trades is the better solution.
Anyway, I watched many times the OOS update from Petko and it is just great. I am sorry that I did not use it so far.
-
Jacpin, so nice of you to share your experience here!
My acceptance criteria on live account are 1.2 profit factor, but I see that you are doing deeper analysis which is more professional.
-
@Desita-I did the same thing..not looking at strategies where the equity line is going down after generation. But since I am generating fewer strategies, I start to look at these and keep them in mind because it may be just going into its losing phase and will come back in winning later on. I agree, I am really mad at myself for not using OOS sooner.
@Haliffa-Thanks for the compliment! I would say that this is me being very conservative, but not overly cautious. If I became overly cautious as well, then none of my EAs would ever go live! LOL.
-
Hello Jacpin, yes Petko mentions that at one of the recent courses that it is important to place the EA on live when it is on its winning phase. It makes so much sense.
I always feel like: “OK, that’s it! I know what I have to do now!” and then something new comes up, something I did not think about or I missed :)
-
Hey Jacpin,
Nice to see you writing in the Forum, it is great to learn from your experience. I have built a similar system to yours but I want to test well before I share. :)
-
Hello Jacpin, you are right! If we go too cautious, we better stop trading Forex. :)
But as Petko says, we should be confident and not brave in trading.
-
Hello guys,
Quite interesting topic here. The thing is that really everyone is different. Different risk tolerance, different capital to trade, and different targets.
The best is to test it all on a Demo, and see where you will feel comfortable.
-
Hello,
I am Laila, a new trader to the forum. I want to ask if someone succeeded to get strategies from the generator if using all Acceptance criteria?
-
Hello Laila,
Welcome to the Forum! If you mean using all Acceptance Criteria at the same time, you won’t probably get any strategies. Normally we use 2 to 5 different criteria and it is already too strict.
Also, it is important to have good Historical data.
-
Hi Laila,
I use only 2 acceptance criteria at the moment in “training” and OOS. My reactor settings are pretty strict though, i like to see a very strong Monte Carlo. This test for me is the most important. But again this is my personal style. You’ll find what you like to see with time. It’s taken me 4 months of live trading to gain 30% in my account. September of last year is when i finally got serious about developing my strategies the way i’m comfortable. If you see Jac’s criteria he likes to see 70% win ratio, this for me means large stops with tight TP in my experience so far. Not something I’m comfortable with. I’d rather see at least 1-1 ratio between SL and TP, and as long as the win rate is above 40% I’m happy. Give it time, and patience. You’ll want to throw your computer from time to time when you get back to the generators and see nothing in the collection. If I knew what i know now in December 2017, I would have focused on learning my reactor style sooner.
-
Hello Petko,
After a couple of your courses, I realized that it’s not the idea to use all of them but few. Also, it depends on what value we will set.
I spend reading the forum(great and very positive), watching the courses and practicing. I have a much better idea of how to use the two software.
-
Hello Roman,
thank you very much for the information and for sharing your experience. I am sure it will take me time, and I am quite serious in trading for the last few years so I hope I will be in tune with the software in the next 1-2 months.
I am just exploring the OOS function(Petko’s videos about it are awesome). It is very hard for me to twist my trading style but I see that this is the right path and I am confidents I am in the right place.
It is very motivating to hear you made this profit. It is a very good profit for 4 months!!! Some professional traders I know, fellows of mine, aim at 20% per year. Of course, they trade with huge accounts.
As soon as I see what works for me, I will share it in details.
Regards,
Laila
-
Hi Laila,
I began as a manual trader as well, but as Popov once said in a post on the main forum for FSB “This is a different way to trade”. When I first started using the software I rushed into throwing EA’s live without really testing them on demo first. Now i know what to look for in order to do that, but when i started i blew an account just as fast as i funded it. OOS is something I pay Very close attention to, and I keep a little bit of data in my back pocket to truly verify the OOS in EA Studio. I’m currently working on learning FSB so you’ll see me asking a lot of questions in the FSB thread as well. To make that profit I’ve also had to endure two major drawdowns with my errors (i started a thread on this). I rushed strategies onto the live account at full demo size and a single loss on two occasions (the last one was a legit accident due to me being half asleep) put me in a 50% dd the first time, and the second was about 3 weeks ago and it took out all my previous gains and then some. Luckily last week was a REALLY good trading week and i made it all back with some EA’s i had put live on GJ and EU.
-
In thinking about it, my account would be up around 60% since the beginning of 2019 if not for my last error. Was about to hit the 30% mark before coming home from a 12 hour day and putting a strategy full blast on the account. It’s still a profitable strat, but now it’s at about 3% account risk instead of the “Full Nelly” that took out my previous profits. I’m enjoying algo trading much more than i did as a discretionary trader. I still stalk the markets on the daily like i’m used to, but enjoying my trades getting triggered while chasing people around for my day job lol
-
Hey Roman, may I ask what was your error you did, because I did not understand, sorry fr that. Just knowing each other mistakes would be so useful. Thanks!
-
Hola Andi,
I’ve been trading using a smaller live account to really get a grasp on how to track the demo EA’s (and practice doing this consistently every day) then make sure the live account continues to grow. What I ended up doing twice is putting my full demo size on my live account dragging myself into a 50% drawdown the first time. And the second time I had come home from a long day of driving (I’m a territory sales rep) and placed an EA live without adjusting the position size. Erasing all my previous gains of about 32% bringing be back down to break even. This was about a month ago so I’ve just recently recovered it due to having more EA’s on the live account now than I did at that time.
Silly mistakes, but cost a lot of time and aggravation. So had to remind myself to slowdown again.
-
Hey Roman, you should be very careful when placing new EAs in the account. Make sure when you generate them in EA Studio or FSB pro to use the lots that you are planning to trade with. This way you will not need to change or adjust when placing on the live or demo account.
-
Hey everyone! Just wanted to post an update of the Acceptance Criteria that I am using after having implemented the OOS from watching Petko’s videos:
Complete backtest-Minimum win/loss ratio-0.7In Sample (training) part-Minimum net profit-100 and Minimum r-squared-65
Out of Sample (trading) part-Same as In Sample
Doing well with this. :)
-
Hey Jacpin, the Minimum net profit is if you trade with 0.1 or? It depends on the lots…
-
@Stephen-Yes, this is with trading at 0.1 lots. I always generate my strategies with trading at 0.1 lots. I will then just increase the lot size during live trading when I feel my account balance is high enough to handle the risk.
-
Jacpin thanks for the update, are you using 30% OOS? I’ve been mainly using 20%.
-
@Roman-Yup, I’m using the 30% and it has made such a great, huge difference in the strategies that I am generating for the better. I just copied what Petko had in his videos and went with that. Thank goodness for Petko.
-
I stay with 20% OOS, I am not sure if there is a huge difference between 20% and 30%…maybe if the data is smaller it is good to use 20%. Anyway, we would have too small data to generate the EAs.
-
And yes, good we have Petko! He changes my trading it total. First I realized from the Forum that I was scammed last year, then I learned how to trade by myself and not to depend on anyone, and now he keeps improving my trading with all updates and new courses.
-
@Haliffa – Agreed, Petko is a huge asset!
I personally like the 20% OOS, it simulates the most recent market conditions (I mainly trade H1) so for me 20% is about a year and a half worth of data. Then i verify the OOS with the most current 6 months of data that I’ve kept with Data Horizon. It’s all about personal preference :)
-
Hello Roman,
yes indeed it is a personal choice but the great thing with OOS is exactly that we are able to test the EAs on the recent months. And it happens automatically with the Reactor which is really awesome!
-
Hello, guys, recently we had an update with the EA Studio which combines the OOS and the Walk forward. I will do my best to record a video about it until the end of the week.
-
@Petko-I can’t wait. Another way to make better strategies generate??? I’m here for it!!!
-
Me too, every time I am excited to see updates and new videos.
-
Hello Jacpin,
yes, today I will record the video and launch it till Sunday.
Kind regards,
-
Hello Jacpin, what else do you use when you generate the EAs or just win/loss 0.7?
-
@Violet-Hi there! Here is the updated Acceptance Criteria that I use based on Petko’s latest video:
Complete Backtest:
Min. profit factor-1.1
Minimum r-squared-70In Sample (training) part:
Minimum r-squared-50
Out of Sample (trading) part:
Minimum r-squared-50
After the strategies are generated, I then filter their performance based on the win/loss, # of trades, etc.
I hope this helps. Happy trading!
-
Hello Jacpin,
thank you very much for sharing. Yes, I saw that videos but I tried generating with the win/loss of 0.7 and I did not get any strategies. Now it makes sense to me.
I saw that you follow the results, and you keep the EA if it makes 7 of 10 profitable, right? What is the SL is higher than the TP. If we have 7 profitable with small TP and 3 negative with bigger SL, isn’t that going to be a losing EA?
-
@Violet-Yes, I keep the EA if it makes 7 out of 10 trades. I follow the performance using FX Blue Live and it has helped tremendously. I usually do a check twice a day on the performance of the robots-once in the morning and once in the evening.
Yes, you’re right. It would be a losing EA with a higher SL than TP. Those EAs that I have that are like that are trading at over 90% accuracy. So, I am usually a good 15-20 wins in before a SL occurs. But as soon as it does look like it is just wiping profits, then it is time to take them off.
-
@jacpin2002 – thanks for sharing all of that. It sounds you already know your system and you know what you are doing and what you need to do. That is lovely. One last question :) How long time normally one EA stays on a live account with this management of 7 out of 10?
-
@Violet-I only started live trading in October 2018, so I am just 6 months in. However, I can say that my current EAs are lasting about 3-4 months staying in profit for right now. But I am constantly running EA Studio as with updates and Petko’s videos, I am generating better EAs all the time that are lasting longer and with tighter performance, meaning the SL is closer to the TP.
-
@Jacpin – I’m very happy to hear your results! You always give me hope LOL. Trying to get my stats to improve, so far the best strategy i have is one Petko released in one of his courses back in September last year :/. Very curious though, with such performance you’ve been getting what’s kind of tp and sl range are you using in the generators?
-
@Roman-Hey there! I’m glad my experience is helping you out. I wanted to make sure I share both my ups and downs, so that other traders know that they are not alone in what they experience. I keep the TP and SL range for the generator as follows: Forex (SL 10-100 pips Fixed or Trailing and TP 10-100-both are set to Always use) Commodities (SL 1-10 pips Fixed or Trailing and TP 1-10 pips-both are set to Always use). Hope this helps.
*The only thing I am struggling with is to have a range of SL and TP for the cryptocurrencies. I want to be able to keep it to Always use, but I’m having trouble finding the range. @Petko-Could you please help me with this? Maybe you can provide some guidance on what range could work for Bitcoin, Ethereum, and Ripple? I did notice that range may differ amongst these three as they all move differently in the relative amount of pips.
-
Hello Jacpin,
I succeed to create EAs so far that last for no more than 5 weeks. So I guess this is the thing we need to look for – longer successful performance. This way the longer the EAs last, the less work we will do. At the end of the day, this is why we are at Expert Advisors :) And what do you mean by tighter performance? Why are you looking for SL closer to the TP?
-
@Violet-Having the SL closer to the TP is what I mean by tighter performance. I like the SL being close to the TP because with a 70% winning performance, my EA will only take 1 or 2 trades to make up the loss instead of 4 or 5. And yes, always be on the lookout for longer successful performance. My two EAs that have been the most successful and longest running performance are for GBPUSD and gold. :) Hope this helps.
-
Jacpin, you are so kind to share all of that. Thanks a lot! I am sure you are in help to many that read but afraid to ask :)
I have started using gold too and I am happy with the results so far :)
-
Hello Jacpin,
It is indeed hard to set the range for the cryptocurrencies. This is because the recent months the Bitcoin and the others were staying in a range without huge volatility. During this time when I was generating EAs, I was looking for smaller TP because if I place a huge range, the generator will look at the whole Historical Data. This will include the data from 2017 and 2018 when the bigger TP worked better and Take Profit of 200$ was taken a few times during the day. This was the time when I recorded the two courses for the Bitcoin Never Losing Forumala.
Anyway, now there is some volatility again, so I have increased again the ranges. Also, I am testing now the cryptocurrencies with the update with the Walk Forward Optimization. I have described it in the recent update called EA Studio updates: Walk Forward optimization.
This way we take the last parameters which are best for the recent period of data. And this is very suitable for the cryptocurrencies which have different volatility all the time.
Hope that helps.
-
Hey Petko,
Thank you for the video, I have missed it somehow. I am on the halfway now but it looks like a really interesting one.
-
Hey Leon, just keep an eye on the EA Studio Updates topic. Whenever there is a major update in EA Studio I record a video about it.
Actually, there are updates in EA Studio all the time but most are inside the code of the software which improves the speed of the generator and other features.
-
Alright, so there are updates all the time as far as I get it. But not on the interface. And I guess, you make updates only on the things that we see, or some extra features? Thanks for your work, Petko! I really enjoy watching the courses and using the EA Studio and FSB Pro.
-
Hello all, do you use different acceptance criteria for the different time frames or not? At least with the different time frames, we have a different number of bars, right? Does this mean that we need to require a different count of trades?
-
Hello guys, it also happens to you that the Acceptance Criteria is not accepted by the Reactor all the time until it has led you to despair? Everyone is responsible for generating their own EAs. I wanted to know how you behave. Do you have an example of Exel tables where you have the whole in overview? Can you take a few screenshots to show your Acceptance Criteria values? This would make it easier for me and for many of you who are still at the start.
-
Hello Centcel,
the best thing to do is to start with small criteria, not strict. With the time when you collect more bars, you will be able to increase it. For example, try to make stricter the criteria every new month. Increase the Number of Trades from 300 to 320, 340..every new month. Because with every new month you will have more bars. It worked for me before…
What I do now is to create many EAs, and after that filter them in the collection. Not looking so much to make it strict.
-
Hello Centcel,
I think you have bad Historical data. I had this issue at the beginning. You won’t get any decent strategies before you collect many bars. Focus on the small time frames because there the data collect faster.
-
Hello Centcel,
the guys have you good answers. The problem is with the data obviously.
What you can do is to use the MetaTrader-Demo data that is in EA Studio. It has many bars. You will be able to create good strategies. As you see in most of my courses I use exactly this data. Also, it updates automatically.
-
Petko, do you use the same criteria in the Acceptance criteria when you generate the EA and the additional filter in the collection? Does it make sense to filter the EAs in the collection with different criteria or better use the same?
Thanks!
-
Hello friends,
Are there any Risk/reward ratio in the Acceptance criteria on Ea Studio
Thanks
-
As far as I know, there are the Win/loss and Return/drawdown.
How should be risk/reward calculated? How much you are risking vs how much you profit. This means the loss (this is what we risk) vs the profit (this is what we benefit). If this is the case this is the win/loss or the profit factor which is the gross loss/gross profit.
-
Hello Haliffa,
Reason why i was asking about risk/reward ratio is because many of the strategys i get, often has bigger SL than TP but has great Profit Factor according to EA studio. But i wanted to aim more for example: risk 10$ to make 20$, which criteria should i focus on then?
-
Hello Riuzk,
I understand your question. EA Studio taught me not to look for a smaller SL than the TP. This is an old trading rule that was never proved to me.
Anyway. if you want to have that, just put in the range for SL and TP when you are generating the EAs.
For example, place a range for the SL 10-15 and for TP 30-45.
Kind regards,
-
“EA Studio taught me not to look for a smaller SL than the TP. This is an old trading rule that was never proved to me.”
Petko, would you please explain abit more on this?
-
-
One of the old trading rules says that your Take Profit should be 3 times bigger than the Stop Loss. After I tested 100s and even 1000s of strategies with EAs I realized that this is not the reality. I have so many profitable EAs where the SL is bigger than the TP.
Also, with EA Studio and the Experts, we test with precise statistics, something not available at manual trading at all.
Of course, there are strategies that work with SL smaller than the TP but not the majority.
And as I mentioned in some of the courses, if you ignore everything but think just about the price. When you open a trade and you place a SL smaller than the TP, you have a bigger chance of hitting the SL first. Higher possibility.
-
And I just did not see prove along the years that the SL should be 3 times smaller than the TP… :)
-
Thanks Petko for you explanation, im guessing you focus more on profit factor with counts of trades.
Many of my running EA’s has bigger SL than TP, but i see sometimes my EA studio find a strategy that have SL on 150 pip and TP 10 pip, these are some strategy i tend to ignore. Because sometimes i have 15 winning trades in a row and 1 loss that takes you back to breakeven. What kind of experience do you have with these strategy Pekto?
-
I normally avoid those as well. There were traders sharing in the forum that they profit with such strategies. I think it was for the Gold…
Anyway, if the strategy passes the Monte Carlo and it is not over-optimized for such SL or TP I think it worths trying it.
Ye, recently I use a bit more the profit factor, but give it a try, use different things, experiment. This is the way that you will find your trading style.
-
Just an update to this topic: I started using more the consecutive losses acceptance criteria. I went down to 5 and the strategies I get are really awesome. I use it as well to remove the EAs from my live account. If they make more than 5 consecutive losses, I remove them.
And it makes a lot of sense because if we create an EA following one thing when this thing is broken or comes not true, it means the strategy was over-optimized for that criteria. Correct me if I am wrong Petko.
-
Hey Andi,
Glad to hear from you. I would not correct you because I use the very same thing with some of my accounts.
Let’s make it clearer to the beginner traders what is consecutive losses in Forex trading. This is how many losing trades in a row we have when generating a strategy or backtesting one.
Now, when Andi says that he uses 5 consecutive losses, it means that he aims at strategies that have no more than 5 trades losing in a row.
So yes, it makes sense to keep the same rules when generating strategies and when trading on a Demo to filter the EAs for live. The consecutive losses in Forex strategies are quite interesting and easy to follow when trading as well.
-
How to follow the consecutive losses in Forex trading with the EAs? I am not sure if I see that in FX Blue?
-
Hey Mery,
yes, you can not see exactly the consecutive losses in FX Blue, you can see the winner and the losers there. So one of the things you can do is to keep an eye when you have 10 trades and see how many of those were the losers. Let’s say you want to have win/loss ration of 0.7, this will mean that you can allow just 3 losers. After that, you can restart it when the strategy goes to 10-20 trades. Or you can restart at every 5 trades.
The other thing you can do is to bring back the strategy in EA Studio and from Data horizon you can set as a start the date when you started trading with the EA. And you will have a detailed statistic for the strategy. Basically, you can import 10 EAs, update the Historical data and see how they go, what stats you have, but just make sure to set the start date.
The consecutive losses in Forex are just an example for that, you can do it with the profit factor, the r-squared, max stagnation and so on…
-
Hey Petko,
thanks for making this clearer. I got it now. I am surprised that in FX blue there are no so many statistics as within EA Studio. It’s a statistical website and EA Studio shows you more about the strategies… :)
Anyway, I wish there was such a feature in EA Studio where we can track the performance of the Experts and we would not need to use the external website.
-
Hi Meryjonestexas,
I see some columns in FX Blue about consecutive losses and winners. They’re all the way to the right on the stats page, so they’re kind of hidden.
Is this what you’re looking for?
-
Hey Traderaw,
thanks indeed! Yes, I see now that there are consecutive losses and profits.
I am thinking of a new system to place the EAs trading on the live account. If I generate EAs with consecutive losses 7, for example, I would then keep the EAs in the live account until there are 7 consecutive losses. If this happens I would remove the EAs from the live account.
How does this sound?
-
Very good point here! In my mind was that Mary was asking about the win/loss ratio! Thanks for the clarification traderaw19!
Mary, I am not really sure about that system…
let’s say you have selected to use 7 consecutive losses. In your real trading you might have the case to hit on 4-5 consecutive losses than 1 winner, and then again 4-5 consecutive losses and 1 winner. And this could continue for some time and it will be within your criteria but you will have losses…
-
Hi everyone. I’m new to algo trading and to this forum. Just wanted to know your thoughts regarding this subject. Thanks
Since the common acceptance criteria can be found under the generator (generator settings), and also under optimization (whether full data, walk forward, and normalization), what is the best way to make use of the common acceptance criteria?
Should we use it while generating strategies, or during optimization? Or both?
What are the advantages or disadvantages in using one method over the other?
-
Hello Jay-R,
Very good question here, but it is from those questions that have no precise answers.
However, I will explain to you what is the logic.
If you use the common acceptance criteria in the generator, all the strategies will be filtered there(this is what I do).
If you want to use the optimizer, it is recommendable to use the common acceptance criteria again. Because the strategies that will come from the generator fitting the acceptance criteria but the optimizer can change that. Normally it should change them in a better way and improve Profit Factor, Consecutive losses and so on, but might lower the number of trades.
Anyway using the Monte Carlo is a must after the generator.
Have a look at the Normalizer, it is quite interesting. I will upload an update video about it.
Cheers,
-
Thanks for elaborating on the logic Petko.
For the Monte Carlo, I did notice in your courses that you use it frequently.
And yes, I’m looking forward to your updates on the Normalizer.
Cheers
-
-
Hey Jay-R,
I will do my best to record those updates tomorrow.
And yes, I use the Monte Carlo all the time. It is the strongest robustness tool for me.
-
Hi Petko,
I’ve been using Monte Carlo and I also think it’s a very nice robustness tool. Are the updates you spoke about ready yet?
-
Hey Stephen,
Glad to hear from you. Yes, it is recorded, just the team is editing it.
Tomorrow it will be uploaded on YouTube, and I will inform you in the EA Studio Updates topic.
-
Hi everyone. I see the term Acceptance criteria is a vast combination of features. When choosing the Acceptance criteria, how do you know what and what not to combine for you to get the expected outcome?
-
Hey John,
The best practice is to start with wide acceptance criteria. For example, use just Min count of trades 300 and a profit factor of 1.2.
Or you can use just consecutive losses in Forex which means that the strategies you want to trade must not have more than 7 consecutive losses, for example.
When you get many strategies in the collection, like over 100, you can add some more criteria
-
Something I wanted to share. Actually Petko said that long ago but I did not realize it. Now I see it is 100% true.
The more count of trades the strategy has, the Monte Carlo passes with more validated tests. This means that we have more robust strategies.
So basically the most important Acceptance criteria is a profitable strategy with many count of trades. With some currencies and time frames I get over 1000 without a problem, and with other 500 is the level. But at the end of the day, I get better strategies for any asset.
-
Yes, Andi!
Simply with the strategies with more count of trades the entry conditions happened many times in the past. This is what makes them robust.
-
My Acceptance Criteria
Balance line Stability 85 and minimum count of trades
It always gives me R-squared greater than 90 and very good profit factor.
It is a new update I believe, you should all give it a try.
By the way, this is my first post, you all have been helpful here. I will try to share one of my results.
-
sodekeaugustine3 do you use out of sample…and any robustness tests?
-
I use full data optimization and Monte Carlo validation(robustness test) in the reactor.
I use OOS Monitor for the Out of sample.
Balance Stability is a performance metric develop to be as a combination of R-Squared, correlation and better scaling for a strategy
Eastudio guide, under acceptable criteria
-
Is the oos monitor truly out of sample? Or us it just showing in sample?? Not sure how it works?
-
Yes jenialyinvest
I do use 30% OOS for my full data optimization but the green part won’t show in the collection, so I have to check OOS Monitor for the OOS.
This way it is the same out of sample you are talking about.
-
Oh,
I just check well,it isn’t the same thing, that is why I used 30% OOS in my full data optimization settings.
What the OOS MOnitor does is that ,it divides the result into 3parts: in sample,out sample and complete backtest in respective of whether you use OOS in your reactor or not.
-
Great discussion here, guys. Keep up the good work!
-
Hello guys,
something that I figured out recently and I wanted to share.
I use the OOS when generating the strategies with the reactor, and after that, I filter the strategies with the min count of trades.
This way, when generating strategies, I am already testing the EAs for the last 20% of the data(simulating the Demo account) and when I filter them with the min count of trades basically I get the robust strategies.
Plus the Monte Carlo is there of course. This way I feel much more comfortable…but it took me nearly 2 years to realize what should I be using for my self.
Yes, I know Petko showed that in some of the courses, but maybe I was not ready to accept it :)
-
Hello,
I’ve been using EA Studio since the beginning of the year, and I took most of the courses. The reactor here works most of the time daily and I have tested many combinations of acceptance criteria and timeframes and I played a lot with data horizons to simulate trading after the out of sample period.
I am not using normalization, it doesn’t seem to have much impact on the robustness of the strategies,
I am not using walk forward too, because 20% OOS with the minimum profit factor 1.2 or 1.1 seems to work better to generate robust strategies. Also this way I have the chance to delete strategies that have 1.1 OOS profit factor but are loosing recently or have huge stagnation.
I like to use minimum R Squared 70 and maximum consecutive losses 10, but what Andi said is really really important. There is a huge difference in miminum 300 trades or minimum 500 trades. The strategies with 300 to 500 trades are much less robust even if they passed Monte Carlo test.
-
Hello Luis,
You are doing an excellent job! It is all about testing and trading with the EAs.
EA Studio, in the beginning, was much simpler without Walk Forward, Normalizer, and other tools. With time, many were added, which makes it a bit more confusing and time-consuming for the traders to find their method of generating strategies.
About the number of trades, that is absolutely correct.
-
Hello guys,
I am new in the Forum and trading, so I want to ask you what acceptance criteria is?
Sorry if this question is stupid, but it is hard for me because I am a beginner trader!
Have a great day!
-
Hello Ivone,
the Acceptance criteria are the rules that we want our strategies to fulfill.
For example, when we run the generator to show us new strategies, we can set acceptance criteria with Profit Factor 1.2.
This means that we will see strategies that have higher PF.
Cheers.
-
Thank you for the answer, Petko!
Cheers!
-
Hey guys,
A quite interesting topic that should not fall down in the Forum.
I see that using the right Acceptance criteria is what gives us the start.
Recently I started to use the Min Count of Trades + R-squared. I think robustness and a nice balance line works great because it just makes sense…
-
Hi Guys,
I have used a very lenient acceptance criteria such as 90% backtest quality, minimum trades 100 and profit factor 1.1 (also for in-sample and OOS).
I realised once the strategy count in the collection area reached 100, the rest of the strategies are getting moved to “Pruned Excessive Strategies”. It looks like there is a limit of a maximum of 100 strategies per portfolio. Not sure though.
Is there anyway I can get these Pruned Strategies back to my collection so I can filter them manually?
Please advise. Thank you.
-
Hello Faraz,
yes, in the collection you will always the top 100 strategies. You can not see the rest. Anyway, you can set on the side additional acceptance criteria, and while the Reactor is working you can filter the EAs so it won’t reach 100 by using your extra criteria.
If you want to have just more strategies than 100, you will need to download the collection when it reaches 100, and delete the strategies so you will collect the new ones.
By the way, you use a very small value for count of trades. Go for 300 as a minimum.
Cheers,
-
Hi Petko,
Thank you for the advice. I will follow the same accordingly.
Have a nice weekend.
-
Hi Petko,
As shown in your courses, I tried to open 2-4 windows of EA Studio to calculate many strategies at once. However, the speed significantly reduces, which you have told as well.
Is this speed something relating to my computer, internet connection speed or is it related to the server of EA Studio?
I have a VPS connection with Contabo with 60GB Ram, 1Tb SSD and 1000mb/s Bandwith. Yesterday I ran just one strategy window for 600 mins and it was able to calculate more than 5 Million strategies within this time period compared to 4 windows which was able to calculate only around 250,000 strategies within a 600 min timeframe.
However, I did another one strategy window today and it was only able to calculate around 350,000 strategies within the 600 min timeframe.
I am starting to think that the speed is not related to my Computer or Internet Connection, it maybe the EA Studio server.
Please correct me if I am wrong and do you have any other suggestion to increase the speed of the calculations, as I feel the specifications of my VPS mentioned above is pretty good.
Any of the group members who can contribute to this? Please do.
Thank you very much.
-
Hey Faraz,
It looks like you have a very powerful VPS!!! Are you using it just for trading?
If you have something else in there which uses the RAM(even you have a lot) this might slow down the process.
EA Studio runs on a server and on the browser at the same time. Make sure you use Chrome.
-
Hi Mery,
I have a separate VPS setup for trading only. This one I use it solely for EA Studio only. So yes as you mentioned, it is a powerful VPS and I’m only running 01 EA Studio session at the moment and nothing else. The speed is still very low. For a 10 hour session, its generating only around 250k-300k strategies and yes I’m using Chrome as well.
Initially it was generating around 5 million strategies every 10 hours, thats why I’m a bit confused on what happened now.
I’ve checked the RAM as well, this is the only program thats running. Feeling that its the EA Studio server that is slow.
Generally how many strategies does your EA Studio generate in a 10 hours timeframe?
-
Its strange, EA Studio is more slower than before and it use many CPU like FSB .
Im sure its wasn’t like that few month ago .
-
It work good now, was a problem of my PC I think .
When do you consider you’re acceptance criteria as too tight ?
More than 10000 strategy calculated without any of them passed validation ?
-
Hey guys,
I have been using EA Studio every single day, and I don’t see any issues with the speed.
Faraz, keep in mind that when EA Studio is loaded, it does not have anything to do with the server. It works just on the browser.
It works as it is on your computer.
It might calculate fewer strategies on the lower time frames because you have more bars. On the higher time frames, it will calculate millions. Of course, it depends on the acceptance criteria which will pass.
The logic is simple – the more bars you have the more calculations it does. It is a lot of mathematical formulas behind the generator… The more values there are (bars) the fewer strategies will be calculated.
-
Hi Petko,
Yes, you’re absolutely right. I just opened 2 browsers and ran an M1 and a D1 reactor for 24 hours. At the end of the day, the M1 reactor had calculated around 500,000 strategies and the D1 had calculated more than 11 million.
So yes, now it makes sense. Thanks for clearing it up Petko.
-
Sure, let me know if there is anything else.
Cheers,
-
Hello,
This is my acceptance criteria. But I am looking forward for some feedback. I’m pretty new to EA studio (4months).
COMPLETE BACKTEST:
min back test quality: 98
min r-squared: 60
min count of trades: 200
min stag %: 30
min equity dd %: 20
min prof factor: 1.2
IN-SAMPLE:
min prof fact: 1.2
min r-squared: 60
min backtest qual: 98
OOS
min prof fact: 1.2
min r-squared: 60
min backtest qual: 98
I been focusing on GBPUSD: 15M AND H1.
This are my settings on my acceptance criteria and validation and looking for some feed backs:
Running on Reactor:
Data Horizon – 02/01/2018 to 06/11/2020 (14,706 bars) data source from FS DUKAS
S/L: 50-100(ALWAYS USE/FIXED)
TP: 10-50(ALWAYS USE)
GENERATOR SETTINGS:
OOS: 30% ( GEN STRATS W/ PRESETS and USE COMMON ACCPT CRIT)
WFO SETTINGS:
SEGMENTS: 5
OOS 30%
+- 10 STEPS
OPT PRESET IND
VALID SEG 5
MONTE-CARLO SETTING:
20 @80%
default settings for validation (im looking to change this, but i just cant think of a good setting for validation)
I hope this info help some people build a basic..
lately I have been messing around min count trades on acceptance criteria and sometimes I use it to filter and or sometime use a lower min count trade.
Im really looking forward for some feedback to improve my validations/acceptance criteria for reactors/wfo
-
Hello, and welcome to the Forum!
I think you are doing a pretty good job and you realize what you are doing.
However, the min count of trades are small, you need to keep at least 300 as a min.
If you tried and you fail to get any strategies it is because you have a small number of bars?
If you use the Free Forex Historical Data app, there are more bars, why would you use less than 15k?
Also, you may want to try generating EAs not just for the GBPUSD because this will bring you more diversification.
Cheers,
-
What I recently use for my acceptance criteria is:
PF > 1.4
A number of trades 700.
I noticed that if I use Walk Forward Optimization in the Reactor it finds more strategies with bigger Count of Trades.
Did anyone notice that? It works great by the way!
-
I have noticed the same. I guess because it divides the backtest period into zones, and in each one, it requires min count of trades.
However, what I wanted to share recently is that I use more and more the R-squared and the tests show that if I generate EAs with above 80 as a value, the strategies perform pretty well after that.
-
Hi johnbrown7,
When you use number of trades 700, how many bars do you then have in your data set?
-
Recently I noticed that we don’t need to have too many bars to get more count of bars. With the default 50 000, you can still get 300 to 500 count of trades.
And this way we get more active strategies. That is what I Explain in the upcoming trading course.
-
Hello Petko!
I realized what you said in this last post after watching the course :) Amazing!
I did many tests and right. The generator still produces strategies within my acceptance criteria of 500 trades and 1.3 PF.
You are always right! It is so evident that if we have fewer bars and we still want 300 to 500 trades as a minimum, we would have more active strategies. I didn’t think it is possible.
Now comes the question: How much should we decrease the number of bars? :)
-
Hey Andi,
It seems I missed your kind words, mate! :)
Well, I do not go below 100000 bars on the lower time frames. It depends on the symbol but even we have 20 000 bars if the strategy shows 500 count of trades and a good Profit Factor, passes Monte Carlo, and is not over-optimized, then all is good.
-
Hi,
I am Very new to EA Studio, I am unable to generate even one strategy! please find the screenshot below and anyone please correct me what is wrong?
-
Just my opinion….. You are using to much filtering. I would drop any optimisation and normalisation. Also Just stick with the monte Carlo for the robustness testing. I would also always use a SL and TP. After 2 years of using EA studios and probably a 1000 + hours invested in this program and with the help of Petko I have realised it is best to set the parameters lower, get lots of strats and filter them out on demo before putting them on live. Currently I have taken this to the extreme by creating about 50000 strats a week on the M1 and then pre-filtering them down to 100 / 200 with the prior weeks data and then running them on demo and doing a final filter down to 30/40 strats to run live. Again this is just what is working for me. Watch some of Petkos videos on the filtering of starts, I believe this is key…… Good luck.
-
Hi richard, what default Acceptance Criteria’s do you use to get that huge amount of strats per week? How many bars do you use? Do you use the Generator or the Reactor? Many thanks for sharing your knowledge.
-
-
Also…. Once you have got the reactor sorted. Take time to learn the validator. This is a very powerful part of the program that will allow you to input 1000’s of strats and “Run them forward” on say the last weeks / months data. I use this as a pre-filter before I take them to demo for a final filter (For example wait 3 trades and then any strats above PF of 2 transfer to live)…..
-
I only go back 8 weeks on the M1. I use the reactor. I leave out the previous week as I use this as a pre-filter. So actually start back 9 weeks. I use COT = 100, PF = 1.2, Max losses = 3. I don’t use any robustness or optimisation. I just want raw strats as I run my own form of robustness tests later by comparing many slightly different versions of my core strategy. I run 12 instances of EA studios at the same time. 6 on each PC. It normally takes 6 hours to get 300 strats on each EAS. So about 4 or 5 days to get 60000 strats. Seems like an overkill I know……. I have done many tests with different amounts of strats up to 60000 and have found that the more I have the lower the DD and the higher the PF becomes. I am aware that I may be curve fitting the strats by having that many so am watching for that. After I have the 60000 I run them forward on the previous weeks data in the validator. With much stricter criteria: COT = 10, PF = 2, Max losses = 1. This shows me how those 60000 strats would have run “Live” for that week. There is normally 300 strats that pass this criteria. I then select the top 100 / 200 and create a portfolio. I run them on demo. I use a custom made filtering app that does the final filtering and sends the ones that pass through to a live account. Normal settings are something like wait 3 trades, if 2/3 trades win with a PF of 2 and R2 above 50% they go to live. I monitor them in the live account and if they start to drop below my criteria they are removed.
I am always continuously testing. So if I have an idea that I want to test I usually run about 10 slight variations of it side by side for a month or so and then evaluate and move forward with the one that is working the best. This way I keep narrowing down what works for me.
After lots of testing with the M15 and the H1 I found I just could not get consistent results. Not to say it can’t be done, just doesn’t work for me. The markets have also been pretty crazy for the last year and there is potentially massive changes ahead to the fiat currency system as we know it….. So i want to trade a TF that can adapt quickly to changes and not look back to far in the past. Again this is just what I am comfortable with…..
I have also narrowed it down to one particular pair that performs well and I also only run my EAs on a 4 hour window each day when the markets are most active and the markets are not crossing over from one another…..
I still have a way to go but the results are improving and getting more consistent as I go…..
I am working towards opening an account with FTMO and some other prop firms in the next month or so. So I have been trying to create a system that stays with the tight requirements of the prop firms….
Good luck and just keep testing and practicing……
-
Hello,
Can I ask you why you don’t use OOS of EA Studio and do it manually ?
> I have also narrowed it down to one particular pair that performs well and I also only run my EAs on a 4 hour window each day when the markets are most active and the markets are not crossing over from one another…..”
So you are generating your EAs also on a 4 hour window each day by using the “Trading Session” ?
Thank you
-
I hold back the previous week of data and then run the strats forward on that week using the validator. This is basically doing a one week OOS. I prefer to use the validator to do this as it gives me more filtering options and allows me to load large amounts of strats. It allows me to set slightly different run times or whatever other filters I choose for testing purposes…
Yes After much testing I have decided to focus on the 4 hour London / US overlap. This is the most active session of the day. Also I have noticed that when a new market session opens it can sometimes reverse the current market, so I try to line up as many things as I can in my favour….
I also only want to trade a pair that has both of their respective banks open during that session which is why I am focussing on the GBPUSD for now….
I create my strats on the 4 hour overlap only. I have done comparisons of creating them over a 24 hour period and the overlap and the overlap shows more promising results….. So for me I set the sessions to 1515 – 1845.
-
This is my summary of algo trading and what has brought me to this point that I feel works for me. (I could be wrong???)
Put a large amount of time / filters / robustness tests into creating a smaller number of EAs that are tuned to run for a longer period of time and ride out the drawdowns with confidence that your EA will work in the longterm.
Create a larger number of EAs with less robustness filters etc… Allow them to run forward and filter them on the fly….. Replacing them on a more frequent basis to keep them aligned with the market.
I am just more comfortable with the second.
If you watch enough of Petko’s videos ( And I recommend that you do) you will see there are 2 main filtering systems that he utilises.
1.) Create your strats, run them forward on demo, filter the top performers ( with PF, WL etc) and transfer them to live account.
2.) Create your strats. Leave them for a week / month. Load them into validator and pick the top performers and transfer them straight to live. Or hold back a week / month run them through the validator and transfer to live.
So for me I chose to use both filtering systems as this gave me better results….. I create my M1 strats over 8 weeks of data, not using the previous week. Then run them forward on the validator for one week, giving me the top 100 / 200 or whatever. (Pre-filter). Then to demo. After they meet my criteria on demo they go to live for the rest of that 4 hour session.
I hope that makes sense….. I have probably gone on too much…….
-
Thank you very much for your precise answer.
More I am using EA Studio and more I am thinking of putting less effort into Filtering ea with robustness tests.
Because we can never be sure the EA’s will perform in the future no matter what robustness tests we are doing.
I may just want to generate a bunch of EAs (3k+) then put them on Demo and transfer the top performer to live account.
I’ve do that in the past but with only 100 EA’s so with a much larger portfolio I think I have much larger chance to have profitable strategies.
-
Hey Richard,
I have been following your work in the Forum, but you haven’t posted anything recently? How is it going with the automated software that you developed to track the EAs?
-
Hi, I’m trading with .o1 lot; I tried setting up the strategy like Petko’s beginner video. I have taken a screenshot of my results. Can you please take a look and let me know what I’m missing? Or a way I can do improve the strategy, so it’s profitable.
Thank you,
Paul
-
Hey mate,
It looks like you had only 2 trades…are you sure you implemented the strategy as it is shown in the video?
Which video are you looking at?
Did you try to see the results with the Premium Data?
-
Plus, you can a rule Long Only which I doubt I showed in the video. Not for Forex. :)
-
Hey Richard,
Been reading through this thread and really resonated with your posts. Are you still on here? Sounds like you were honing in on success for sure to me so expect you are too busy enjoying your millions.
If not then it would be great to hear how you are getting on, especially with regards to getting stable results that stay within the strict guidelines of the prop challenges?
From what I have read my approach and use of EA Studio is very much aligned with yours so I really do hope we haven’t heard from you because you are too busy enjoying your profits!!
-
Also lots of other strong voices in here that haven’t been heard from in while (Andi, Haliffa, Roman, Jaquin2002).
Anyone about, how have your results been going?
-
Hey Samuel,
Thanks for keeping up the spirit in the Forum! I appreciate it! :) I am sure the others will join again.
-
- You must be logged in to reply to this topic.