04 - Fast Forex MillionsThảo luận

what risk are you running this on?

I'm using fixed lot 0.01, which is my preference (using fixed lots over risk %).

Btw, for the account balance I have now, these fixed 0.01 lots corresponds to about a risk 1.5% setting.

Well - the back tests looked so good... but reality is so far totally different.
I am also testing this paid EA and honestly speaking so far only my broker makes money on this EA - it is not exactly what I expected to get from a system you pay for....

On the other hand... if it would be a real hit - they would not sell it, or at least not for such price...

let's give it a few more weeks - perhaps there is a sens in this (at first glance) chaotic activity...
well - if not ...

Yes, I hear you, Mark!

Not only that the advertised backtests looked much better than reality, but the backtests I did showed less than half the claimed performance. So here are my opinions so far:

The bad (and I should say very bad):
- vendor backtests not done properly (several of us, independent testers, got the same, less-than-half-performance results)
- the main claim that made me (and many others) buy this EA was that it was broker independent; to some extent that was true, meaning the spread did not affect trade placement. But on the other side, forward testing indicates the trades are very different, even in the same account type (I ran a FXDD real and FXDD demo account, with same spreads, and still the results are different)

The so-and-so aspects:
- market had huge swings in the days following the start of FFM forward testing, and this is not an EA to benefit from that; so the horrible performance I see in my accounts can fairly attributed to that

The good:
- there are so many newbies consumers in this field, I'm actually glad the robot tanked on launch; otherwise we would have heard lots of stories of burned accounts. Sure, that does not prevent other newbies to come next week or month and kill their money, but at least for the people that were witnessing this 'wave' it provided a learning experience (I actually saw people very seriously asking if they can make 1,000,000 starting with 1,000, withing a year - and if not with FFM, what EA would be recommended for that!)

Cheers,

Remus

The best perform BT for an EA is the EA developed data feed. so BT results will be different from the creator results. I think the most important point is as Will noticed that EA uses fractals. Fractals can be different from data feed to data feed. the main reason is the high of high or vice verse can be different. I guess this is the main point of different trades.

Remus,

I have to admit them one thing which is in 100% correct:
the system is broker independent - it loses with all brokers equally 😈

Let's leave it running for a couple more days ... perhaps it is only a bad period...

But it shows again, that if you want to have a good system, you have to design and write it by yourself...

Greetings,

m

I have now stopped the forward tests of FFM in all accounts. There are too many variables in this picture, and sorting them out takes too much time. I would like to gather more statistics and two positive things (the only ones?) about this EA are: it's very fast at backtesting, and version 2.1 generates lots of trades. But I also have more important things to do than test this EA, and also need the VPS resources for other tests.

So let's see what I learned from my forward testing so far... One primary goal of the experiment with FFM trading at different brokers was to see if the EA is taking the same (or very similar) trades at different brokers. I used two brokers that are at opposite ends, as far as spreads go. FXDD was chosen because of its high spreads (4.1 on AUDUSD, 2.0 on EURUSD, 3.x on the rest), and Pepperstone Standard because of its low spreads (I could have gone lower, with Pepperstone Razor, but wanted to keep commission at zero, the same as FXDD). I also ran the EA at FXDD with a demo and a real account, to see if differences occur in that case.

Conclusions:
- FFM takes trades very different trades at FXDD (demo: https://www.myfxbook.com/members/NextLevelForex/fast-forex-millions-fxdd/228516, real: https://www.myfxbook.com/members/NextLevelForex/fast-forex-millions-fxdd-real/106578) and Pepperstone (https://www.myfxbook.com/members/NextLevelForex/fast-forex-millions-peps/228518). The platforms were started at the same time, and used the same version all the time (FFM 2.1 until 1/30, when was switched to 2.5). There are not so many trades taken so far, but it's enough for me to say its very broker dependent (quite the opposite from the sales page claim)
- version 2.5, trading the FXDD demo and real account, took almost the same trades, but version 2.1 took much less identical trades. You can take a look in the attached spreadsheet, where I matched the orders in the demo account (left columns) with the orders in the real accounts (right columns). The middle columns (K and L) indicate the difference between open times and realized pips. I moved trades that were not matching downward in the list, so you can tell by the number of 'holes' in the trade list how close the trading activity was (i.e. few holes mean the EA had very similar activity in the two accounts, while many holes show not so similar activity).
- when taking the same trades (at FXDD demo and real), the outcome was very similar (i.e. the differences in realized pips were pretty minor)

If/when I come back to FFM, I'll probably test version 2.1 again, because I liked the high number of trades it was generating (and look, Birt's account, still running 2.1, seems to be doing fine: https://www.myfxbook.com/members/birt/fast-forex-millions/229091).

Happy trading,

Remus