EA performance real vs backtest vs demo

Has anybody had much success with EAs in real trading? I’ve been working on one and backtest results are pretty good, but I’m not sure that this performance will reflect in real trading. There’s still a few things to sort out, but progress is looking good.

I will be running it on a demo account, but it needs to be running 24/7 to function properly and won’t be able to do this for a couple of months. The results are shown in the image below. Does anybody think it will actually hit this kind of performance in reality?

1 Like

Personally I have not yet found an EA that makes money long term. They all seem to rely on martingale systems.

That’s probably because nobody that has a working EA is going to give it away. Even if you sold it for $20k, that’s small fry compared to what you could get from an actual working EA.

Mine isn’t martingale, but it does have a R/R that is very unfavourable hence the big drop downs on the chart. I would be manually stopping these trades as I do with my strategy that this is based on, but until I start forward testing and then live, I don’t know if I believe the results that I’m seeing.

1 Like

That is very true! If people really had a working EA there would be no need to sell it, especially if it was achieving the results these marketers band about.

1 Like

There are EA that twork long term but they stack profit in a very unpleasant way.
Maybe they profit 2 months and stay in DD 10 month every year even if the final result is positive.
Robust EA and strategies are difficult to endure, while martingales grow 95% of time until they blow.
Many EAs are designed to be sold to rookie traders that think it is possible to profit every month.

it’s possible, but there’s no way of knowing

nobody in their right mind with a consistently working EA is ever going to share, sell, advertise or promote it, that’s for sure

and that’s the reason it can’t ever be possible to give the answer “no” to your question (however much one thinks that’s probably the answer)

lots of people (including many financial institutions) have had long-term success with automated trading software that enters and exits trades on their behalf, on the basis of indicator-patterns, price-action-patterns, or combinations of the two, and books describing and discussing them have been published

but those aren’t “EAs”, because those aren’t people using MT4/5, and they’re not people betting against counterparties rather than trading in a real, transparent market, either - and there’s something to be learned from that

still, it’s a similar kind of thing, based on the same principles, and it at least suggests that it might be possible to do it with an EA (and that fact is what keeps the EA-selling industry going, of course)

if anyone does, you should stop listening to them - that’s for sure

that really isn’t possible, for many reasons

that’s a very big red flag, and an additional reason to be very skeptical about the backtesting outcomes


yes all is possible. here is test from 2015 till 2020.
it contains 3 safety levels: 1) stop loss in pips 2) stop out level in money on each pair (size 2-4%) 3) equity stop 30% size on account fully. it trades on 6 pairs same time and all of them in sum never reached 30% account decrase from 2008 till today :slight_smile: as I cant post the monitoring link you can search it for on myfxbook as David’s bot