TSPCALC Strategies Automation

[ubascuba]

I'm wondering if Matt has thought about setting up a computer to automate creating new strategies? I know there is an almost infinite number of possibilities, but having a computer dedicated to creating new ones would help. I've benefitted greatly from following some of the strategies and would not mind throwing in a few bucks to do that.

[bloobs]

uba, you are stepping into machine learning territory it is gonna cost ya i imagine

but i think the present "crowdsourcing" model in tspcalc works fantastically as is.

[mjedlin66]

I'm wondering if Matt has thought about setting up a computer to automate creating new strategies? I know there is an almost infinite number of possibilities, but having a computer dedicated to creating new ones would help. I've benefitted greatly from following some of the strategies and would not mind throwing in a few bucks to do that.

Solving them all (by this I mean all possible annual strategies) by brute force is actually impossible. There are too many possibilities. Literally the fastest computers in the world can't do it.

Which is why you would have to make assumptions. Start down one path, decide these are junk, discard, and move on. That's where either good models or machine learning come into play.

Instead, I made the seasonal builder. Every possible way to trade a single month has been calculated. Then it's up to you to put them together (although there's a best-fit button if you are using the same criteria on all 12 months).