All,
TL;DR: I think P123 has quite a bit more potential than is being realized. Which in a normal world would be embraced by P123. Seen as a good thing with an effort to realize that potential.
I have been requesting an array for 10 years. I found a way to get around needing that array but my method is limited. These are my only ports, with real trades. No survivorship bias as these have been my only ports since their inception. I would be the first to say it is not statistically significant.
But it suggests machine learning might work at P123 EVEN IF MY METHOD IS LIMITED:
So here is my question: “I wonder what I or anyone at P123 could do with serIous support for that matrix I have been requesting for the last 10 years at P123.” What I could do with a random forest or what Jonpaul could do with XGBoost or someone else with PCA regression etc.
Maybe with a healthy diet and exercise I will live long enough to find out.
I think that might happen but here is the thing: When I finally get that array at some time in the future I will ask myself: ’ I wonder if P123 could have been a machine learning legend, written about in books that are also about Ed Thorp and James Simons." That possibility gets a little more remote each day we wonder if we should put any effort into that matrix, IMHO.
Best,
Jim