Guess What? Points Matter

facebooktwitterreddit

If may be hard to believe, but scoring points matters in terms of wins and losses. Score more points than your opponent and you usual win the game. I know, right? Who would have guess it would be so? OK, now before you all roll your eyes and chalk this up to being a John Madden or Chris Collinsworth worthy expression of the obvious, here me out for a second. The purpose for looking at this topic was to answer this question:

How many points does a team need to improve by to expect 1 additional win?

The answer: 36.

That is, if a team scores 36 more points over the course of a season, or their defense holds opponents to 36 less points, or some combination of the two, then that team should expect to win 1 more game than they would have otherwise.

How did I come up with that number? That’s a good question. It comes from 10 years with of data from every NFL team. Below If a graph of all the data.

The black line is a linear regression that was performed on the data, and as you can see it fits the data extremely well. The correlational coefficient of .837 (which varies between 0 for no correlation, to 1 for a perfect correlation) indicates that the correlation is very good. The equation for the line is also given on the chart. The y intercept of 8 is what you’d expect for any correlation to win total, since that’s the average number of wins that teams in the NFL have given the 16 game schedule.

Its the slope of the line that gives us the useful piece of information here. It tells us that each point is worth .0278 wins. Or in other words, it’s 36 points for each 1 win.

Down the road, when we start to figure out how certain players or certain plays (like turnovers) directly lead to points, we’ll have an understanding of how those plays contribute to win and losses.