Another relatively accurate week.
Isaacson-Tarbell was 12-4. Beatpaths was 12-4. They disagreed on two games, though, and ITB ended up 13-3, missing only on DET>WAS, JAC>HOU, and NE>ATL. I was also 13-3, so I’m still tied.
Also, now that we’re through the first three weeks of the season, I’m going to pick a tiebreaker to use for the rest of the season. After going through the backtesting, I’m picked one – good old uNet. uNet has the advantage of being easy to explain – it’s just the number of teams in one’s beatwin paths, minus the number of teams in one’s beatloss paths. It also performs reasonably well from week to week, and as a fallback, it refers back to the previous week’s rankings, which helps with stability. It’s one of the earliest tiebreakers we’ve used here, and it’s proven to be one of the better ones.
So, without further ado, here is the graph of the season so far, through Week 3. No beatloops so far.
It’s interesting to me that Cleveland has only lost to undefeated teams… they probably are still a very bad team, though.
Iterative and Weighted graphs are up. No loops, and thus not much difference between them. Link on my name or go to beatgraphs.com. It doesn’t look like the graph will get any more interesting next week based on the schedule. Hopefully week 5 will give us some mix up to work with.
That’s pretty crazy that we might get through Week 4 without any beatloops.
The NY Jets, Indianapolis and San Francisco have beatpaths to every other team in their respective divisions.
Best chances for loops forming in week for seem to be in the division games below:
Tennessee at Jacksonville
St. Louis at San Francisco
Of those, a Tennessee win seems more likely than a St. Louis win does. I didn’t look for possible results in week 4 due to multiple game results coming together to form loops, that was going to take more time than I had.
Oh, and to answer something you (rhetorically) asked in the week 2 results summary post, Yes, Washington just is that bad.