Very frigid. Some questions:
– Are the fashions expert on files most moving from this year or historically? How attain you story for adjustments in rules/teams and so on. if historical
– It seems like your fashions very highly ticket teams (that can additionally embody engines) which on the total makes sense when teams delight in a clear revenue or procedure back (Merc, Williams). Nevertheless the set the racing is end (midfield) or a disparity in driver capacity (Pink Bull, Renault) this is snide. Is it seemingly that the relationship is non-linear. A easy neural community could match better (as an belief) fairly than the lin regression.
– Accept as true with you thought to be a likelihood model for speed wins (bayes?) In the intervening time your monte carlo sim accounts for quail variance however there isn’t any accounting for variance in the speed model unless I misunderstand.
– Basically could you save the elo model in rather extra depth, you yell a speed is a series of head to heads, however what number of attain you attain and so on?
– Enact you need to also delight in ancient performance metrics for your fashions, I could perchance be astronomical drawn to seeing those.