If I were going to use one of these tools, I would use Trading Solutions not Ward Systems, or maybe both, but never just Ward Systems. nitro
Hi nitro And thanks for the info. The tradingsolutions website looks more "professional" to me. Just a gut feeling that may prove wrong in the end. Have you had any profitable experience with Neural networks ?
Lancelot, Neural Solutions is a more hard core product, and they have networks that are more amenable to the data patterns markets generate. I have never had a single profitable trade with a canned NN. But I think the mathematics of how they work can be adapted to a market trading system that would have predictive use. nitro
Lancealot & Nitro, It's been a couple months since the messages on tradingsolutions. Have you guys been using the product? I started with it about a month ago. It does seem like it could be a great asset to one's trading arsenal IF one can make it through the learning curve, which I find myself a bit frustrated by at the moment. Just curious if you guys have any additional thoughts...
Hi george91197 No, I have not used tradingsolutions yet. I was trading funds at the time I posted (and still do) and needed trading solutions to be able to make a buy decision on D day based on D-n price or a sell decision based on D-m price (n,m being different integers in the 1-5 range). I asked tradingsolutions support if this lag time was supported by their product but got absolutely no answer I seems that some lag factor has been implemented in the trading styles of the new v4.0 and maybe one of these days I will make a second attempt at trying to find out if this product suits my needs. Have a good trading day
Lancelot, I use Tradingsolutions and I know you can lag inputs to the NN by any number of periods, e.g. "-5" (use the input value from 5 periods ago). It's a good piece of software, just don't expect any silver bullet. Once you learn it and build yourself a few dozen good models for some futures, etfs or stocks you can probably make yourself 20-30% returns annually with decent risk/reward ratios (I am not saying more is impossible). Anybody who expects 100% returns every month will be disspointed. I find myself putting a lot of work into this, screening around 200 neural models every day. Easily 4-5 hours work every day plus some homework (=development, testing) on the weekends. But it's worth it ... for me.
Hi makloda I don't know if using lag inputs is the answer with Tradingsolutions. If I apply full optimization to, let's say, lag[f(n1,...,nn),3] wouldn't the "3" will be changed to an optimized value? Thanks
I would be interested to hear about the current position of Neuroshell ? Is it now profitable and do you need an IQ of 150 to understand it ?