Short answer? We can explain about 60% of the variance with our models, but we can't predict baseball politics!
We know from reading and watching Moneyball that baseball loves data analytics. Predictions like this are what made Billy Beane and the Oakland A's so successful, despite having such a lower payroll. As young aspiring data analysts, we thought to do some Moneyball!
We thought this would be a simple model. Plug in some statistics and get a salary. After many models, we discovered that player performance statistics only play a minor role in determining a baseball player's salary. After doing some further research on the topic to figure out why our models were not as easy as we thought we found another side of baseball we were not accounting for: “baseball politics”. This can include the team or league the player is in, how good their agent is at negotiating contracts, the length of the contract, injuries, the player's attitude in the locker room, and how good the player thinks he is. This is only scraping the surface of “baseball politics”.