How is espn poff calculated

The key holding call. For a lot of people, this was a traumatic event, but from a Win Probability perspective, what the Patriots did was so unlikely that the graph actually told the story of the game.

BA: We start with what the pre-game odds. That 25 percent is harder to put in context. Sports are relatively controlled in what we can measure. In politics there are so many intangible factors. Nate Silver and FiveThirtyEight does the best political modeling around, and his projections were less certain of a Clinton victory than many others. You can build really intricate models but there are so many factors to capture. BA: Our models do a really good job, but as we saw in Super Bowl 51, there will always be rare events.

The Atlanta Falcons suffered a historical collapse, and it led to an epic turnaround. Those are really, really rare. This is going to be a rare historical event. As I said, this really did tell the story of game. Washington State is coming …. Now, by checking the ESPN baseball standings, you can learn that -- as of Saturday -- the previously hapless Rays have a These percentages are calculated by coolstandings.

These playoff chances are shown in the last three columns of the ESPN. As of Saturday, here are the teams with the best odds of making the playoffs POFF , according to coolstandings. See standings page for more:. How are the percentages determined?

The key is estimating how many runs each team will score, on average, against every opponent on its schedule. Coolstandings takes into account factors such as home vs. Previous season performance is also used, especially near the beginning of a new season. Coolstandings even takes into account all of the MLB tiebreaker scenarios when determining divisional and wild-card chances. Some factors aren't considered, including starting pitchers, trades, free-agent signings and injuries.

Roster changes are indirectly taken into account because the model emphasizes recent team performance over older data. The estimated runs for each team are plugged into a modified version of Bill James' Pythagorean Theorem to determine the odds of each team beating the other teams on its schedule. The data is regressed to the mean, so hot or cold starts don't weigh too heavily on the simulation.

Random error is also added, since a. Luckily, baseball has been around for quite a while, so there's a lot of data for math geeks to drool over.

The geeks at coolstandings have optimized their model against historical data going back to , the first year the World Series was played. Since then, there have been playoff teams, 1, non-playoff teams, and more than , games played. By simulating every season since and comparing the projected results against the actual historical results, coolstandings has been able to evaluate and improve the accuracy of its model.

For each range of playoff chances -- say between 40 and The average absolute error across all teams and dates is less than 2 percent, and is continually being improved. So will the Rays bust through and clinch their first playoff spot? We then look at how many "seasons" a team won its division or won the wildcard, and voila - we have our numbers.

The trick, of course, is to determine what chance each team has of beating every other team. Our method is to use simple team statistics e. For those of you familiar with baseball prediction, we use a variation of the Bill James "Pythagorean Theorem" to predict results. Pretty smart, huh?