Gamblor Raids Again: Week 1 NFL Picks: Welcome back to another season of fun gambling after predictions that inherited a neon light, nightmare ill-tempered Tempest, one of the few, GAMBLOR! Last year, introduced a new model, which I called the Son of Gambler, and The Unholy Alliance between the two abominations made a small profit of $ 222. As usual, I tinkered a bit 'of the program during the offseason. I was hoping to develop a model for predicting winners that could be useful in betting on the money line “i.e. predicting the winner or loser of the actual game without the spread factored in” My efforts delivered a model that could accurately predict winners 59.9% of the time, which sounds pretty good. But someone else has already developed an even better model, which picks winners at a rate of about 64%. That’s even more impressive, right? Not so much. The best predictor of winning real football, the spread damn, yes, if you just pick the favorite to win every time, it is only 67.4% of the time. If you are building a betting strategy based on this? Of course not, most probably on the money line that will prevent you to see any benefit. But it just goes to show you that Vegas is pretty smart when it comes to predicting who has the edge in the NFL. Even though this exercise didn’t provide me with any new gambling tools, it did provide me with a new way to show you what Gamblor is thinking in each game. Gamblor simulates each game close to 150,000 times, and comes up with a distribution of scores for each team. I use the expected score is simply the average of all these simulations; a value that represents the exact average score curves. For example, the first game of the NFL season in 2011, the computer provides for the Green Bay Packers beat the New Orleans Saints with a score of 27-20. Tracking the spread of the results of the vision for each team:
0 comments:
Post a Comment