I'm a new forum member, but I'm an old-time trader. I'm currently researching two neural net programs, NeuroShell Trader and Trading Solutions, to see which better meets my needs. Both have serious deficiencies. NST is a largely undisclosed program which uses nets of unknown topology and which gives very little information to prospective purchasers. TS is far better on net topologies, disclosure and openness, but much weaker on "training/testing" issues like clear separation of in-sample and
out-of-sample data. Both programs use a "programming interface" which pre-dates the modern computer and probably the printing press. The only advantage of these interfaces is that spelling and programming syntax errors are hard to make. They're great for making logic errors. I could elaborate ad nauseum on these weaknesses, but I'll spare the forum.
A buddy of mine, Carl, has been trying to get access to an old Yahoo group that deals with NST, but he can't get admission to the group -- the moderator doesn't respond to e-mails. There may be no useful info in the group, but NST is so-close-mouthed that even a little info would be helpful. Carl and I have pounced on anything published to learn strengths/weaknesses of NST. We now know why they make so much use of DLLs, for example: NST doesn't know how to deal with recursion in function definitions.
If anyone has archived the NST Yahoo group or can otherwise help Carl and me access the Yahoo group, it would be appreciated.
out-of-sample data. Both programs use a "programming interface" which pre-dates the modern computer and probably the printing press. The only advantage of these interfaces is that spelling and programming syntax errors are hard to make. They're great for making logic errors. I could elaborate ad nauseum on these weaknesses, but I'll spare the forum.
A buddy of mine, Carl, has been trying to get access to an old Yahoo group that deals with NST, but he can't get admission to the group -- the moderator doesn't respond to e-mails. There may be no useful info in the group, but NST is so-close-mouthed that even a little info would be helpful. Carl and I have pounced on anything published to learn strengths/weaknesses of NST. We now know why they make so much use of DLLs, for example: NST doesn't know how to deal with recursion in function definitions.
If anyone has archived the NST Yahoo group or can otherwise help Carl and me access the Yahoo group, it would be appreciated.
