Pause for thought...
OK. There appears to have been some shenanigans around this Forex Maestro EA and launch. Maybe. Too soon to tell, and it's certainly too soon to court libel.
For reference, here's some code I found for
Firebird EA v63g. Forex Maestro indicates that it's v63d, so it will probably perform differently.
If TraderSeven owns the copyright (rather than just claims it, which is different), and they are not actually the vendors of the Forex Maestro, then they may have a rightful claim to make - if they can afford the suit. By all accounts several thousand copies of Forex Maestro have been sold, and so there may be a reasonable prize for the contest.
Actually if the code is the same, I like that...I like to be able to look under the hood. It gives me a degree of confidence in the system. AND it can also give something to build on...a framework to modify and enhance. You see, it is perfectly possible to dramatically change the performance of a trading system by tweaking a few bits of logic and a few parameters. Perhaps this is what has happened...although I would imagine that pride of authorship would make me want to change the comment if I had significantly enhanced it...(By the way, the comment on the advanced Forex Maestro says Firebird v2.0).
Certainly, inspecting the code of Firebird v63g I see no signs of an artificial neural network, although it was claimed in the advertising - "
THE AMAZING MONEY MAKING SECRET OF AN ARTIFICAL INTELLIGENCE NEURAL NETWORK FOREX ROBOT" (No affilaite code or hop here.)
The bottom line is that we can't really tell (without decompiling or testing side-by side) whether this is the same system as the comment would indicate. However, it seems likely that it was one of several systems under test that was rushed to market after a winning streak; the advertising really didn't stack up; it claimed 'PROOF' without actually showing real proof - you couldn't tell whether the results were back tests or live trading, or how selective the referrals and examples were.
(To see what could be done, try looking at the FAPTurbo marketing, which has live, regualarly updated trading results. It's still not perfect, but it's a heap better.)
On top of that, I have received no responses from my requests for help or information. So, I have to conceded: It feels like a scam to me.
HOWEVER: Let's not throw the baby away with the bath water!
It's possible that the periodic updates of the system include parameters from such a neural network - many trading systems developers find that they can regularly produce systems that work well in the short term, but which degrade after a while...so a periodic optimisation-based update would make sense...and it would make back-tests pointless.
Let me state, quickly, my objective in purchasing the Forex Maestro system (I know my first post said I didn't know why I bought it - I confess I was being glib): I want a set of a small number of poorly-correlated, positively performing automated trading systems. To do this, you've got to 'kiss a few frogs', and it's quicker to buy them than develop them.
Why poorly correlated? Because all systems draw down, and I want them to draw down at different times and in different timeframes - because trading isn't just about reward maximisation, it's about risk mitigation too. And if your account can peform well from a risk/reward perspective (ie. Sharp or Sortino ratios for instance), with performance beyond your cost of capital, then you can scale up massively.
So I'm evaluating the systems still, as a part of my portfolio - regardless of their heritage! Their performance Thursday - Friday was impressive until the late afternoon. My $5000 demo account rose from $4020 (I lost $ 980 earlier due to wrong settings for my 6 digit forex account) to $ 4450 - 10% up in a day - although it was powering towards $ 6000 at one point!
One of the problems with Metatrader is that it doesn't show the Equity curve in the detailed report, just the Account curve. So you only see the closed trades on the graph...and any system with wide stop losses will look great until they incur losses from undetected long-term trend changes - at which point they typically just blow up! The long term trend will evenutally recover any un-closed position to a profit (until it changes direction). And nobody really has the patience to forward test them for that long!
What about back-testing, I hear you cry? Surely that should save our skins? Well, no, not really...it's useful to a point, but whenever optimisation takes place, it should be viewed with suspicion - it produces systems that are over-fit - they work great on history but fall apart in live user. And you need to accept that the evolution of trading systems that takes place through forums like these and market forces IS A KIND OF OPTIMISATION - we are the infinite number of monkeys!
I know, problems, problems, and no solutions - so let me conclude philosophically: There are no perfect trading systems! There are better ones and worse ones and you can't really tell which is which from hindsight. All systems draw-down; that's OK. It doesn't matter what is in the black box, it's the empirical performance of the black box that counts!
Anyway, it's Sunday, and I need to spend time with the family, so I'm going to stop here.
