- Joined
- Sep 25, 2015
- Messages
- 161
- Reaction Score
- 514
Since you have a ? concerning "who wins", I'll throw my two cents in.
Winner should not be determined by margin of victory and here's why.
Say team A wins a game against team B 80 to 60, a mov of 20.
Poster X's prediction was 81 - 60, mov of 21. Poster Z's prediction was 40-20,
a mov of 20 that is exactly right.
Can you really say the Z prediction was better ?
I propose a comparison of the predicted score for each team against what the actual score is.
I.E. Same 80-60 actual score.
Poster X predicted score 81-60
81 compared to 80 = a difference of 1
60 compared to 60 = a difference of 0
Total difference = 1 + 0 which = 1
Poster Y predicted score 84-64
84 compared to 80 = a difference of 4
64 compared to 60 = a difference of 4
Total difference = 4 + 4 which = 8
Poster X would be the winner between those 2 posters.
I agree with your arguments. Here is a graph of the forecasts for the Nebraska game:
We could calculate who is closest to the actual score (and this is the way DaddyChoc feels it should be done too.) There are two small issues, (1) the person closest to the actual score could have forecast the wrong team to win, and (2) the people who only include a margin of victory would be completely out of the running.
The proper way of calculating the deviation is to use the (x,Y) coordinates of the forecasts and the actual score, and determine the distance between the two for each forecaster. It involves the pythagorean theorem, but I can do that pretty easily.
I re-ran the data for the last two games, and the person DaddyChoc feels should have won the contest would have won the contest, even though neither had the correct margin of victory.