Ranking Analysis – Week 17 – End of Regular Season
Tom here with the end of the regular season charts. The changes to the rankings in Week 17 were no less dramatic than any other week this season since Week 4. We had two big movers this past week, Houston and Jacksonville. These are both teams that have been difficult to place all year long, especially Houston, which has moved 10+ ranks (either up or down) in six of the last fifteen weeks. Houston returns from #7 back down to the middling #17, whereas Jacksonville has nearly all the support kicked out from underneath it, dramatically revising its rank from #5 to #24—a big reassessment, especially this late in the year.
Large reassessments like Jacksonville’s drop this past week make sure that the variance rises for another week. The way that variance is calculated, two teams moving 10 spots results in less variance than one team moving 20 spots. So even though the overall changes to the ranks are not earth-shaking, that revision of Jacksonville still causes the variance to spike.
The effect of the new tiebreaker method seems to have changed the dynamic of the stability chart. Last year, the rankings became more stable as the year progressed. This year, in an attempt to increase the predictive power of Beatpaths, TT instituted a new ‘EdgePower’ tiebreaker used to calculate the ranks. The new tiebreaker seems to have done better: whereas last year’s record was 150-105-1 (58%), this year’s record is 167-89 (65%). In increasing the predictive power of of Beatpaths, it looks like the new tiebreaker has decreased the ’stability’ of the rankings: teams shift in the rankings with greater fluidity and more responsiveness than before. As you can see on the stability chart, from Week 4 to Week 17 there is no major trend, either increasing or decreasing.
Compare last season’s regular season stability chart alongside this year’s to see the different effect:
| End of 2008–9 | End of 2009–10 |
|---|---|
 |
|
Pick Confidence – Wild Card Round
Last week pick confidence did not do well, in part because New Orleans and Indianapolis sat starters which may have altered the outcomes of those games.
This week, we have only four games, all of which are pretty low confidence games. We ought to expect this, because playoff teams are all generally top tier, and the matchups shouldn’t be as lopsided as most regular season games.
| Matchup (Winner-Loser) |
“Confidence” (out of 100) |
EdgePower comparison (predicted winner – predicted loser) |
Result |
|---|---|---|---|
| New England-Baltimore | 16.67 | 69.17-52.50 | ✗ |
| Green Bay-Arizona | 12.92 | 59.17-46.25 | ✗ |
| Cincinnati-NY Jets | 7.08 | 62.08-55.00 | ✗ |
| Philadelphia-Dallas | 2.09 | 67.92-65.83 | ✗ |
TT gave the official Beatpaths picks for this week. He notes that three of these games are revenge matchups, the same pairings as Week 17. In all three cases, Beatpaths is making the same pick it did last week: Cincinnati over the NY Jets, Philadelphia over Dallas, and Green Bay over Arizona. This, despite the victories of Dallas and the NY Jets last week.
As usual, I’ll return to fill in the results of each game.
