Here a copy of my regression spreadsheet as it exists now. I still have a
long way to go to refine the existing factors and add new ones, but it's a pretty good start.
I may not have done it the most efficient way possible, but I'm not an expert at regression analysis. It works.
My spreadsheet contains all dirt Graded Stakes in my database (2015 through current). I couldn't post the entire thing here because it exceeded the upload limit of Paceadvantage, but there's a lot of data here to at least see what I am doing and what weights for each factor it came up with for the entire thing.
I'd be happy to answer any question about the data.
For the record, using just the spreadsheet data alone on Graded Stakes on dirt from 2015 through the tail end of 2022 (need to update it again), it predicted 35.55% winners and had an ROI of .9026.
So with with no handicapping or subjective analysis at all and not even considering some factors the win% was quite good and the ROI far exceeded the track take.
Eventually, I'll break it out by dirt sprint, dirt route, turf sprint, turf route to get more refined weights and post those results also. It's going to take time though because I have to keep up with day to day handicapping.