A better scoring system for match predictions

The outcome of a match can either be 2-0, 1-1, or 0-2. Let’s assume each possibility is equally likely.

Suppose someone predicts 1-1 for a match. If you score 1 point for each correct game, this predictor will get either 1 or 2 points. But on average, he will get (1 + 2 + 1) / 3 = 1.33 points.

Suppose someone predicts 2-0 for a match. He will get 2, 1, or 0 points. But on average, he will get 1 point.

So a 1-1 prediction always gets at least a point, and scores better on average. In order to account for this imbalance, I propose a correct 2-0 prediction scores 3 points instead of 2.

**

Addendum:

Assuming each outcome is equally likely may not be a fair assumption, though. If teams are equally likely to win a game, it should be 1-1 50% of the time.

This means a 1-1 prediction will score on average (.25 * 1 + .5 * 2 + .25 * 1) = 1.5

A 2-0 prediction will score on average (.25 * 2 + .5 * 1 + .25 * 0 ) = 1

In order to compensate, a correct 2-0 would have to score 4 points!

Leave a comment

Design a site like this with WordPress.com
Get started