Change to PI?

[w_sohl]

I'm a little cornfused. Can someone with a statistics background better explain this chart?

Advanced NFL Stats: First Down Probability

I would think that break even is at 50%. Also, wish the graph had all the lines. I'm also assuming that .6 on the chart is 60%

[pob14]

Originally Posted by w_sohl

I'm a little cornfused. Can someone with a statistics background better explain this chart?

Advanced NFL Stats: First Down Probability

I would think that break even is at 50%. Also, wish the graph had all the lines. I'm also assuming that .6 on the chart is 60%

The text explains; break-even is 66% (.66 on the chart), because the probability of a new series, given that you start with first-and-ten, is .66 - that is, in the NFL, a team gets a first down (or a TD) two thirds of the time when they start first and ten.

So if they gain 5 yards on first down, you look at second and five on the chart, which (if I'm reading it right) is about .7; we've gone up a little from .66, so that's a success; we're now more likely to get a first down than we were when we started.

However, if they throw an incomplete pass, it's now second and ten; that's something like .52 on the chart. They still are more likely than not to get a first down, but now it's less likely than it was on first down, so that play leaves them worse off (obviously). So that's below the break-even point.

What he means by "break-even" is: is this situation better or worse than first and ten?

[Rich]

We had DPI today on 3rd and 27. First time with the no rule. Nobody complained.