## Prediction Accuracy Formula

I wanted to answer the question of how well we might be able to predict a teams’ point percentage at the end of the year compared to their current percentage based on how many games are left in the season. It’s easy to say that a teams’ percentage after every passing game reflects their end-of-the-season P% better than it did a few days ago, but we want to know exactly how those margins of probability are changing as the season progresses.

To do this, I got the P% for each team at certain game intervals of last season. I then found the difference of these percentages from the teams’ P% at seasons end. What was found was that the average teams margin of possible P% taken away or added by the end of the season is calculated by y = -.0014x + .1236, in which x is the number of games that have gone by so far (the formula is adjusted to an 82-game season, even though the results were taken from last years 48-game season). This might be a little hard to understand so I’ll explain it in the context of the Jets.

At the time I wrote this article, Winnipeg has played 26 games with a .500 P%. Putting this number into our formula gives us:

y = -.0014(26) + .1236 = .087

This means that if the Jets were an average team, a best-case scenario would leave them at the end of the season with a .587 P%, or about 96 points in the standings, which is comfortable to make the very edge of the playoffs. In a worst-case scenario, they would end the season with a dismal .413 P%, or 68 points (about where Colorado was at the end of 2010-’11).

As you can see, at this point in the season we can’t be absolutely confident about any prediction we make for a team down the road, but what you should know is that the margins become a lot smaller once we get past the 20-game mark of the season. A teams’ P% shouldn’t be used for the first 1 through 19 games of the season because a linear model wasn’t fit for the curve of the 10 game interval compared with the rest of the intervals.

If this FanPost is written by someone other than one of the blog's editors, the opinions expressed in it do not necessarily reflect those of this blog or SB Nation.

## Trending Discussions

forgot?

We'll email you a reset link.

Try another email?

### Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

### Join Arctic Ice Hockey

You must be a member of Arctic Ice Hockey to participate.

We have our own Community Guidelines at Arctic Ice Hockey. You should read them.

### Join Arctic Ice Hockey

You must be a member of Arctic Ice Hockey to participate.

We have our own Community Guidelines at Arctic Ice Hockey. You should read them.