Forum Replies Created
Perfect thanks guys much appreciated
Thanks again Samuel
I have now watched those lectures on Petko’s professional course re identifying the 12 balanced robots and swapping them monthly. This is very different to the approach I have been instinctively pursuing, but it is one I am going to try straight away. I have seen many an EA do brilliantly on demo, and then totally flop when I switch it to live, which is why i thought i just need to worry about the 2-3 year picture and not worry about the 3-4 month stagnation periods. My thinking has been – Maybe i have just been unlucky in switching performing EAs to live accounts after good periods and at the start of stagnation periods.
Re Petko’s approach – I’m intrigued as to why last 6 month performance, and then last 1 month performance is a strong indicator of good performance in month 7, but not beyond. Any views / insight on that?
Also I note in the professional course that this method – ie swapping out 12 balanced EAs each month was being demonstrated in November 2020.
Petko (or Samuel if you follow this approach) can you confirm this approach is still working in August 2022, and also provide some performance information for this approach over the last 21 months since Nov 2020, ie how many months of profit / how many months were not profitable, what has been the average profit per month etc etc.
Thanks Samuel, you are being super helpful.
All of that makes sense, but this comment caught my eye:
“2-3 months would be a better wf amount in my opinion.
Also remember you are not managing robots in anyway with that system but simply putting them on and leaving them for 7 months (which is ages).”
Apologies what did you mean by “wf”? Again my mindset here has been – what robots can I identify and then just leave on charts indefinitely, knowing that they always trade on profitable win rate over the long term. But it sounds like you manage your robots and swapping them in and out on a more frequent basis, is that right? How then do you determine which ones to swap in and which to swap out?
Do let me know, I should add I am going through two of Petko’s courses:
Forex Strategy Course + 12 EAs I Trade Live
EA Studio Professional Course
…so if the answer is in there, feel free to tell me to review those courses first!
Re your second point Samuel, I think you’re right. I have become fairly obsessed with the principle of the 1:1 Risk: Reward strategies. If you set a strategy that has an equal pip value of stop loss to take profit, the win rate would average out to 50% in totally random market, exactly like a coin toss. But we don’t believe the market is random, and we believe that indicators DO have some predictive qualities (if they didn’t, again we would expect to see a 50% win rate over the long term). So I have left compelled to find those indicators that can tip the win rate to 52/53/54/55%. But as you say it is a very restrictive approach and i think i need to focus on other methods. In my 1 2 1 Petko suggested EAs with large stop losses and smaller TPs have performed better in his tests, have you found the same? Is there any particular approach which has generated strong results for you in the past?
“It sounds like you are generating a collection using the reactor and then adding them all to the portfolio and then checking OOS and not getting a profitable single asset portfolio? This isn’t too surprising really”.
That’s basically right, for example I create a collection of my best looking EURUSD 5M strategies, generated using the Reactor with above settings, and over data period 1.1.20 to 31.12.21. I load them into portfolio and calculate and check their collective profit figure at 31.12.21, then change data horizon to 1.2.20 to 31.7.22 ie 7 months extra data, but invariably the collection has lost money over the extra 7 month period.
I take your point re a multi asset portfolio. I am running my tests across balanced pairs. Then i load collections of EAs and check them in the Portfolio, asset class by asset class, but it is not as if I am seeing great results for say, EUR JPY, which i know will compensate for mediocre results for USDCHF. None of the collections seem to be showing continued profitability for the newly added 7 month data period.
Unless I am missing something the Portfolio feature can only be used for one asset at a time? You can’t create a portfolio showing EAs from USDJPY, EURUSD, AUDNZD etc at the same time?
Ok thanks Samuel and Petko, I really appreciate the advice. I am now working on a new 1:1 risk / reward project but the EAs will be generated using adjusted settings as per your recommendations.
The last couple of days i have been playing with the 100K / 200K bars method, or actually i have been testing using different date periods, for example generating EA using data horizon 1.1.20 to 31.12.21. Then downloading collection and re calculating based on data 1.1.20 to 31.7.22, so each EA shows performance in green on the chart for the “new” 7 month period ie 1.1.22 to 31.7.22. So far i have had mixed results with this, I would have expected that for EAs generated using robustness tools that performed well in 1.1.20 to 31.12.21 MOST (but not all) would continue to perform well in the 1.1.22 to 31.7.22 period. But that has not been happening, in fact i have yet to generate a collection which has a higher profit figure in Portolio as at 31.7.22 compared to 31.12.21. Is there any reason why that might be? I am sure they have not been over optimised as i have used the settings as per above.
As ever any advice always appreciated.
Also Petko, you mention running the Reactor using 100K bars and then re calculating with 200K bars. Is that not the same as running the reactor and in Generator Settings, setting 50% out of sample? That way your data history could be 200K bars, but the strategy would only be formulated using 100K (50%) bars and then then green zone on the chart would show you how the EA performs in the other 50% (100K bars) out of sample data?
Thank you for responding so quickly Petko, I really appreciate it.
I am happy to leave out the full data optimisation. but is there a way to incorporate “Entry Time” into my strategies? I am conscious that almost all EAs perform better at certain times of the day and I’d like that setting to be in my EAs to ensure they only open trades at the most profitable times of the day.
The only way i could see to incorporate time of day into the EA was to have it as a preset criteria and then ask the Reactor to optimise it (ie select the most profitable hours of the day) in Full Data Optimisation, but if there is another way, please let me know? Thank you
ok great thank you! I will try this out.
HI Petko, I have checked out FSB Pro and I see the feature where the EA can add lots to open trades. I am going to trial a few of these EAs but that was not quite what i had in mind.
I have a couple of promising EAs that have low TP values, high SL values and high win rates, making them profitable. They trade frequently, sometimes 10-15 trades a week, and when i have long winning streaks the pips cumulate nicely. But I have had a couple of occasions recently when the EA gets stuck on an open trade and price is fluctuating between open price and SL. The longest open trade I had was 8 days. Even though that one turned out to be a win, it was frustrating that I could the EA was not taking other trade set ups because it was stuck in the open trade and is obviously programmed to only trade one at a time. Over that 8 days it probably missed another 20 trades, and when i am reasonably confident this EA is profitable over the long term those missed trades obviously mean missed pips and profit. Is there anyway the EA can be changed to trade every set up, and not just one at a time? This feels like it would be an easy tweak in meta editor, but i am not a programmer, so i don’t want to start playing around with the code and disable my EA in some way. Please do let me know your thoughts, I would also be interested to understand the rationale in having EA studio programme EAs to only trade one trade at a time. Could that not be a user choice option in the Generator, ie: trade one at a time / trade all set ups?
Great thanks Petko, I will check out FSB Pro.
Can you upload EAs generated from EA studio into FBS Pro, and then add the trade all set ups feature?May 5, 2021 at 17:29 in reply to: How satisfied are you with the automated trading software EA Studio? #85449
Hi, I am new to EA Studio, and am absolutely blown away by this tool. I used to pay a developer 2-300 Euros per EA, wait a week for it to be written and then hope for the best that the back testing might match up to some minimal manual testing. Now I can generate hundreds of EAs in a matter of minutes. I have no idea how long it took to code and build, but I am so glad I found it. Amazing work, I haven’t had any serious results yet, but some early test accounts are showing encouraging gains. Thanks guys
Fantastic, thanks Petko
Hi Petko, thank you for responding. Having spent more time on the site I realise i need to upload historical data from my broker, instead of using the meta trader demo data. That way I will be comparing the same data source. If i am still having problems after that I will post again. Thanks!