Has there been any progress fixing the Lot/% bug.
I started week with account of 8600, opted for 5% but EA used 0.01 Lot.
I ended week £0.27p up.
Here is MT5 EA version attached. Please tell me if the algorithm is ok according to this thread?
You can use reverse logic as well - by parameter is_retrace_mode setting to true.
Please share the best backtesting results.
Made some random backtesting and one completed optimization for DJI30 - see optimization results attached (3 years, real ticks).
Seems there can be found good opportunities. Believe there can be combinations of multiple symbols as well - will continue optimizations. Saw promising results with DAX as well (see attached Excel) - will run an optimization on DAX too.
Please post here other optimizations as well - have to spend several hours on 1 optimization so if we will do it together then we will get wider opportunities to split investments for several baskets (EA setups).
A lot of work must be done here - please join backtesting and publish your results. Together we can do it!
tried to do backtesting and got unsupported filling mode. I am using ICMarkets MT5. any idea?
By mistake found better DJI setups... I set in tester DAX but in EA settings forgot to change symbol and so left with DJI.. as a result I have got different optimization results for DJI (see attachment).
Compared the best result in 2nd optimization Excel and see that it is missing in 1st one. MT5 strategy tester chooses testing scope according to some rules. Ok, I am using the option "Fast genetic based algorithm" and not "Slow complete algorithm".. that's the reason?
wait for a bit.. hmm.. idea: strategy tester was executed on DAX ticks (EA uses function onTick) but with DJI data.. maybe this gives some impact.. actually.. so you set up EA on DAX chart but set in parameters to track DJI prices. onTick will execute with DAX ticks.. interesting...
you can see in Excels attached differences when the same EA parameters for DJI executed on DAX ticks or on DJI ticks.. quite a big difference when SL2 and TP2 case.
Finally, have got DAX optimization results (8h running) - see attached.
And here is set up with combinations of DAX and DJI (5 EAs on 1 account) - different days and/or different SL/TP to diversify risks:
1) DAX - day 1, hour 15, SL 42, TP 22, lots 0.1
2) DAX - day 1, hour 15, SL 3, TP 9, lots 0.3
3) DJI - day 5, hour 6, SL 20, TP 20, lots 0.1
4) DJI - day 5, hour 11, SL 38, TP 26, lots 0.1
5) DJI - day 2, hour 14, SL 2, TP 2, lots 1 (on DAX chart must be executed)
In ~3 years profit ~3000 USD. Starting balance 1000 USD. Profit ~100%/yearly. But here is no swaps/dividends included in backtesting. It should be less profit, therefore.
Here calculated balance (not equity!) drawdown of all 5 combinations:
2016-08-09 18:04 - 2016-09-15 18:28 down 108 from 1062 was up 62
2017-06-20 09:03 - 2017-06-29 19:00 down 108 from 1655 was up 700
2018-07-24 17:21 - 2018-08-10 10:00 down 139 from 2900 was up 1353
2019-09-12 02:24 - 2019-10-22 21:50 down 103 from 4067 was up 1306
According to courses from my broker which I attended - backtesting drawdown in real life could be 2x more and profits 2x less. So that's what we can expect here: ~50% yearly profits with ~30% drawdown. Fully automated. Think that's not bad only for 2 symbols combination but here can be found much more - need more time.
But yes - here is 1 dealbraker.. according to courses must be at least 700 deals to consider statistics. What do you think? Will you give a try to this setup with real money?
Hello guys, ok we have fixed up some minor bugs with the EA.
As it turns out it is impossible to calculate lot size using risk % without having a set SL, and as this is a basket with no set SL it can't be done (I thought it could) so it has been replaced by a lots per balance instead. Now you can choose the lot size based on the account balance which I think will work well in place of risk %.
We also fixed some things with the trailing stop as it was not working correctly and small bug in how the current profit/loss was displaying on the dashboard. hopefully now all is good and we can have another profitable week!
First post also updated!
Follow The Candles EA.ex4
The question is who is going to invest in those live tests.
Eas are good ideas but People must understand that the Brokers and market makers simply wont let them work against them.
Your best bet is to design something that is semi-automatic that you can control after clearly spotting the trend.
Even then you must tighten control over the bot to ensure that the brokers and market makers dont send out opposite signals before pushing live trade in other direction. Or flood the bot with high volume of transactions to get you into over trading zone!
anyway seems I missed out on the topic of this thread and my posts here are misleading from author purpose. sorry about that. please delete my posts from here if possible.
Let me know, I'll live test this for the team.
© Forex Factory