That's assuming the math authority actual knows something about the game of football and are using relevant data and situations or generalized assumptions. I currently have a sig bet with someone about 4th down %. He's using research from a couple different "math authorities" showing that because the probabilities they calculated that teams should go for it almost all the time on 4th regardless of distance needed or field position. Looking deeper at their data, it was inherently flawed. The "math authorities" cited, didn't use 4th down stats because there wasn't enough to attempts to draw a reasonable conclusion, so they used 3rd down stats, one study only used 3rd down stats from 1st and 3rd qtrs to eliminate the issue with time running out, iirc the other study only used 3rd down stats from the 1st qtr of games. Each study also left out other 3rd down attempts on different criteria. These "math experts" are starting with flawed data to start their calculations, the strategy a coach would use to call a 3rd vs 4th down play are different, and the consequences are different for failure. The data they used was over a 12-15 yr span in each case. They concluded that coaches should always go for it or at least go for it in almost every case, regardless of yards needed or field position. Our bet was that he thinks there will be a 20% increase in 4th down attempts, based on these studies and that the Eagles were effective with them. He doesn't get that different teams are better or worse than others and that those stats are for generic team A or B, it doesn't work for each real life team. The usual difference on 4th down attempts each season over the last decade have only varied 1-2% up or down.
In your attempt of a basketball analogy, If Curry and Shaq each made 9/10 FTs, the chance of each of them then making 900/1000 is much different. The teams or players don't rise or regress to the league average, they rise or regress to their own ability. In such a team dependent sport as football, there also are numerous other variables to evaluate.