Well, I can tell you that the algorithm is not that smart. Predicting an F1 race is actually very hard.
So, no matter what you get "told", it would be a bad strategy to listen.
I my case, this year started a bit bad, as I wanted to keep away from all the rookies, so I went for Alonso and Sainz...
You can also think of it like this: Someone (maybe me) takes the time to listen to the FP3 training and wait for radio messages talking about how comfortable the driver feels in his car. This is information that for sure gives an advantage. Is this now also unfair?
Also, I am 100% sure that all good players at the minimum have some kind of spreadsheet where they calculate something like expected points divided by costs.
Fantasy leagues without at least some minimal statistics would be a bit strange.
With my system, I am sure though that the budget optimization part is more helpful than the machine learning prediction part, because it is simply pretty much impossible to predict the outcome / positions of the race. When I figured this out, I was a bit happy, because it kind of proves that F1 (currently) is not boring at all and qualifying does not determine everything.
