MTPredictor

I don’t know what to say about it, really. Steve Griffiths has done something with the isolation approach to EW that no one has done. The isolation theory works, the software is brilliant when it works. The real deal however is the Position Sizing module. I have never seen a more simple and effective approach to risk control than this (other than the algorithm I am currently running). Having tried over a dozen automated semi black box systems (both predictive and trend following) the concept and ideas that MTP is based on are flawless and works of genius. The ones that have success with it will concur that this works. Steve G, Matt B are excellent, credible people to work with and they are the real deal. Having said that, I give it zero stars. Why, you ask? The basics of any good system are very simple. One of them is being non-optimized and non-curve fitted. MTP fails on both. It will only work on highly liquid markets (or so it says in the brochure) and it will only work 1/3rd of the time. It will not catch trends as it is designed to trap human trading behavior via the EW patterns alone. If that was the only problem I would still be using it. (I am not mentioning the fact ) that you could go days without a single trade in a highly liquid and volatile market like the ES.) All of the above is still not a dealbreaker but any stretch. When the system catches an EW correction, boy is it ever right (and very profitable to boot). The one and only reason that I gave it back wa because I believe (rightly or wrongly) that the algorithm it’s based on is flawed. That is one heck of a statement to make. But I stand by it. The reason is this – it took one signal (just one alone) to cause this conclusion. The system generated a signal and then validated it with a trigger. Then, as the stop was hit the system took the signal away as if it never existed. To me that is absolutely inexcusable. Once a signal is generated AND triggered it must (MUST MUST MUST) stay on the chart for posterity. Whether it is a successful signal or not is immaterial. I was told that it was a rare confluence of events and it may have been that way, I do not know. But once you lose the trust in the signal generation capability of the system, it does not matter. Caveat Emptor.