Originally posted by nd_safety
... when you blow teams out, it hurts their rankings more than if it is a close game, which in turn hurts your SoS.... well, not quite.
The SoS is a rolling average of all of your weekly pregame snapshots, with each snapshot showing how that week's opponent was faring in the season prior to their match-up with your team. I keep it as a pre-game view (as opposed to a post-game view), to avoid the unfortunate consequence of blowing out the other team, because as you point out, that would hurt your own ranking.
40% of each snapshot is based on the other teams' winning percentage prior to the game they play you, and 60% is that team's previous listing in the Power Rankings. To get the Power Ranking (whole number) into a comparable number with the winning percentage (thousandths), I have to use a calculation, which is: 1-(Power Ranking*1.5%).
I don't know that this will help, but I'll provide an example for the Prague Predators' first two games of the season:
Week 1: Danzig Raytheon (0-0 & Ranked
21st in my pre-season starter level comparison)
1-(
21*1.5%) =
.685Since it was the first week, the opponent's .000 winning percentage didn't count, so the SoS was exclusively based on their previous Power Ranking of 21. In other words, the Prague Predators Week 1 SoS (
http://goallineblitz.com/game/forum_thread.pl?thread_id=2182255 ) was a direct reflection of their opponents' pre-season ranking (
http://goallineblitz.com/game/forum_thread.pl?thread_id=2114372 ), since no win/loss records could be factored in.
Week 2: Russian Red Stars (0-1 & Ranked
26th in the Week 1 Power Rankings)
(
.000*40%) + ((1-(
26*1.5%))*60%) =
.366The average of those two weekly SoS calculations (.685 & .366) was .526, which at the time put the Prague Predators as the 30th highest SoS, which was shown on the forum post:
http://goallineblitz.com/game/forum_thread.pl?thread_id=2197828