In post 873, AGar wrote:In post 869, mith wrote:AGar and Yiley are the luckiest compared to their expected records, by almost 2 games each, while I have won 1 more than expected.
By my math, Yiley is actually super lucky by 8 games?
The 8 game difference is between two different power ranking stats; EPE is everyone plays everyone, while WDST is a more complicated stat attempting to account for not just the luck of who you face each week, but also the luck inherent in how the scores are distributed. For example, based on your score vs. the mean and sd of the league's scores, you might expect to win 75% of the time, but it happened that six teams beat you that week by a few points, killing your EPE; or the converse of that, which is what's happening with Yiley. As a more concrete example, this is what the WDST value is for the third best score each week (EPE of 9, or 9/11 = .81 repeating for comparison sake):
.77, .79, .77, .88, .79, .87, .62, .91, .82, .69
Yiley's was the week 7 one, much lower than average for a EPE of 9 (.62*11 teams = 6.82). In that week, we had two scores in the 100s, and then a bunch of scores in the 80s, with Yiley the highest of that pack at 87.68. It was only slightly above average for the week, but because of the way the scores were distributed it was good enough for third.
The 2 games comes from comparing actual records with EPE (divided by 11). So your expected record is slightly better 6-4, and Yiley's is about 5-5, whereas your actual records are 8-2 and 7-3 respectively. (Using WDST, Yiley has a 2.5 difference, so definitely luckier by that measure.)