How This WorksJust as the Predictalator plays any individual game 50,000 times before it's actually played, it can also play entire seasons 50,000 times before they are actually played. That is what we have done for the 2014 college football (FBS) season. The actual regular season and bowl schedules have been played 50,000 times, allowing us to compute likely records, expected bowl matchups and power rankings.Using the Predictalator is not as much of a concern to us as making sure that we have the best possible inputs for players' statistics, progression over time and age, roles, health and playing time as well as teams' coaching styles and weather.In general, we apply strength-of-schedule-adjusted, relevant statistics from the player's collegiate games to a fairly traditional player development curve that is driven by class and previous playing time. Not only does this curve help to set average player inputs, it combines player health history to dictate the variance ("boom or bust" potential) of a player's inputs.Playing time is more art than a science, especially in college football. We have done our best to estimate playing time for all players who are expected to see significant starting time in 2014. That being said, some playing time gaps have been filled in by our best approximations of average bench and role players on teams.After all of the math and analysis is conducted, we import each team's schedule and play EVERY game 50,000 times.
Reply With Quote
