Well, after a brief hiatus, The Racing Equation is back in time for New Hampshire! While I haven't been blogging about the predictive model that I use each week to forecast winners, I have been running it and keeping track of its success. To say the least, it has not gone well. In part because of the tracks that were on the recent schedule.
Sonoma posed a problem as a road course. It's not that the predictive model didn't have its favorites (Ambrose, Kurt Busch and Montoya), it's that the data for the model is made up of past performance over 5 years and driver performance on road course has been changing in recent years. In particular, "road ringers" (i.e. Boris Said, Ron Fellows) can no longer be counted on for top finishes, yet they have good historical data which inflates their predicted finishes. Likewise, guys like Joey Logano have no real historical success on road courses but seem to be picking up the tricks of the trade more each year.
Then was Daytona... not a data modelers friend. Nor is any restrictor plate race for that matter. As if the unpredictability of "the big one" isn't enough, now we have to try to guess how well drivers will perform in couples (AND which couples will exist!). No thanks. The model had Jr and JJ #1 and 2, which I thought was a great sign given that they were to partner up. We all know how that worked out. To its credit, the model did have Ragan at #9 and would have definitely been listed as one of my
Opinions expressed in blogs are those of the individual bloggers and do not necessarily represent the views of racing-reference.info.