|
As a far as the WC scenarios, the game space is huge 2^32. My software was predicting that the Steelers had clinched a playoff birth last week, but none of the reporting services were showing that the Steelers had clinched. 5000 simulations were not enough to find it, even when I set the Steelers to lose the rest of their games. I eventually found the scenario by which the Steelers would miss the playoffs and when I moved my sliders accordingly, my software made the correct prediction.
So I'm not saying that there is no way we can make the playoffs by a WC, but the chances must be impossibly small. We can increase the power of the simulations forcing the Chiefs to split their last two games. So every simulation is one in which the Chiefs finish 10-6.
Let's also move the Chiefs SOV and SOS up by having Arizona beat Dallas. That is the only game I can easily identify that would help the Chiefs while hurting other teams.
I'm also going to set all sliders of remaining games to 0.5 to evenly sample all of the remaining possibilities.
Any WC scenario must also involve some combinations of wins and losses between the Pats, Jets, Pittsburgh, Ravens, Jax, and Colts. Each of these teams plays 2 independent games, so that is a total of 12 games.
2^12 = 4096 different outcomes. That is a reasonable number of simulations.
Now if we need common games, SOV or SOS to go our way, that would need more simulations than 4096. Maybe a lot more. But I ran 30,000 simulations for each way in which the Chiefs could go 10-6 (beating Tenn or beating Oak) and did not find a single instance in which we made the WC. Still not proof that ESPN was wrong, but I would like to hear the scenario by which we got a WC birth before I believe it.
|