We have a tendency to divide teams in to tiers when we look at rankings — whether preseason polls and schedule analysis or when you’re picking games in the middle of the season. For instance, last season it would have been fair to say that Alabama, Clemson, and Ohio State were more or less in a league of their own, with Oklahoma, Ole Miss, Michigan, Florida State, Notre Dame, LSU, and Stanford in the next tier.
The point is, even though we rank teams both in the traditional sense — like the top-25 polls — and using advanced stats like the S&P+, the teams are often grouped in tiers rather than there being an equal distance between each ranked team. The difference between the third- and fourth-ranked teams isn’t necessarily the same as the difference between the tenth- and the eleventh-ranked teams.
David Wunderlich, over at Team Speed Kills, recently looked at peers of SEC teams. For instance, he finds that Alabama’s only SEC peer was LSU, while Ole Miss has seven peers on their schedule for next season. That’s a pretty big difference. Naturally, you’d expect Alabama to have a much easier time with their schedule than Ole Miss, just because they should have more contentious games.
So what happens when you look at Ohio State’s 2015 schedule? Who were the Buckeyes’ peers in 2015 according to the end-of-season S&P+ rankings?
One way the S&P+ is presented is according to an adjusted scoring margin — essentially it converts the S&P+ percentages in to a score against a hypothetical average football team from that season. So Ohio State’s S&P+ margin was 24 last season, meaning that the S&P+ metric would expect an Ohio State victory of 24 points over that hypothetical completely average opponent last year (funny enough, the most statistically-average football team last year was Northern Illinois with a .8 S&P+ margin — and Ohio State only beat the Huskies by a touchdown!).
|Team||S&P+ Rank||S&P+||Proj. Pts Dif||Tier||Actual Pts Dif||Actual Points - Proj.|
A couple of things are clear.
First, Ohio State’s tier includes more or less just Michigan and Notre Dame. Using the tiers, they would have been the only real toss-up games, though the Buckeyes would have been favored in every game of the season (going by the final advanced stats). The projected margins in those games were just 3 and 5 points, respectively. Ohio State played extremely well relative to the advanced stats, however — 26 and 11 points better.
Penn State, another game where Ohio State vastly exceeded expectations, and Michigan State, a game where the Buckeyes had their third-worst underperformance of the year, are the near-peers for the Buckeyes. These were teams that would have been expected to play the Buckeyes closely — within two scores or so. Instead, Ohio State out-performed their expected 14-point victory by two more scores, but didn’t meet expected outputs against Michigan State by 11 points.
But there were plenty of games where the S&P+ margins more or less nailed the actual outcome of the game. Against Virginia Tech, Hawaii, Maryland, Rutgers, Minnesota, and Illinois, the projected margin was within a touchdown of the actual margin in the game. For four of those games, the projected margin was within two points of the actual margin.
So the S&P+ peer system does a good job not only putting teams in to intuitive tiers (thanks, David!), but also projecting the actual outcomes of games (and identifying cases of under/over-performance). Next, we’ll look at Ohio State’s 2016 schedule and set best- and worst-case scenarios for the season.