Wednesday, June 8, 2016

Naive Baseline Says England to Win Euro 2016

For your (and my) prediction evaluation purposes, this post presents a naive baseline prediction of Euro 2016, based on the transfer market value of each team. This is an exercise I've done frequently over the past years, and history shows that the naive baseline is hard to beat.

The tournament's winner will be England, pictured above. Here is what I have done for this exercise.

1. I take the transfer market valuations from CIES.

2. I use these to rank the teams from 1 to 24, based on transfer market valuation - the higher valued team is ranked higher. (The spreadsheet I am using comes from here.)
3. The higher ranked team wins each match, there are no draws. Group stage matches are shown below, with the projected winner highlighted.

4. This procedure is applied to all 51 games. Here is how the groups are projected to finish.
The evaluation has 2 parts.

1. Total games picked correctly from the start of the tournament, maximum of 51.
2. I'll also run a "reset" when the knock-out rounds start, using the teams that actually make it. That total will be out of 15 games. Here is how the knock-out rounds look from the start.

Skill is calculated simply by the number of predicted victories greater than that anticipated by the naive baseline. I will likely provide a post-Euro evaluation of various predictions. If you would like me to include a prediction in my evaluation, please suggest it in the comments, ideally with a link.

Eventually, all of these fun prediction exercises I've been doing will find their way into an academic paper, once The Edge is out and I have the time. I already have a great dataset. This will add to it.

Enjoy!

No comments:

Post a Comment