Calculated Reactions: The 2011 quarterback rankings

For several years we have tried to quantify the best quarterbacking performances in CIS football. We started in 2005, where the top QB was Laval's BenoƮt Groulx, and continued through to 2010 (Kyle Quinlan) and now 2011 (uh...Kyle Quinlan again). Below are the rankings for the 2011 season.


First, in a vain attempt to provide some year-to-year continuity, here are the top 10 quarterbacks in adjusted net yards per attempt, relative to their conference average and adjusted for the level of competition each player faced. (These are not the official 2011 rankings.) All playoff games are included.

Leaders in adjusted net yards per attempt relative to conference average (100 = average), 2011 (all games)
1K QuinlanMcMaster290225
2A KennedyWindsor293182
3E DzwilewskiCalgary269166
4B GreeneUBC359161
5K GravesAcadia253160
6B Prud'HommeLaval333148
7J CreightonSaint Mary's157143
8A ColbonOttawa289139
9R QuestConcordia222127
10J Doyon-RochSherbrooke343124

Note that those are indexed to the conference average, and not per-attempt numbers (it would be very difficult for anyone to gain 225 yards per pass attempt, unless they were playing Waterloo). Kyle Quinlan's 225 means that for every 100 yards gained through the air by an average OUA quarterback, Quinlan and his receivers gained 225. No, that is not a misprint.

These rankings are based on slightly different methodology than those in previous years, so we cannot compare Quinlan's 225 directly to the previous leader (Josh Sacobie, 2007, 176), but it's safe to say it is probably the best passing performance by anyone for as long as we've been tracking it here.


Now, because I like to make things difficult and change the methodology each fall, here is a new explanation of how the 2011 rankings were developed.

We're trying to estimate each quarterback's win value for 2011, which would give us a more accurate list of who the most valuable quarterbacks really were. It's kind of silly to have Jack Creighton ranked above Aaron Colbon when the latter was involved in many more plays; there's value in playing more often, provided you aren't terrible.

We start with each quarterback's adjusted net yards (ANY) — and that's all QBs, not just the ten above — and as noted above, we are giving more credit to those who played tougher competition. Specifically, we take the strength-of-schedule component from last year's RPI, and convert it into a multiplier where a value above 1 means this player faced better opponents. We then compare his total to what a replacement-level quarterback would achieve in the same number of opportunities.

"Replacement level" refers to the level of production you would get out of a player acquired for minimal cost. In CIS football terms, this would mean someone not highly-recruited and maybe not even on a roster in the first week — in other words, whatever yahoo happens to be walking by the stadium, so long as that yahoo has some eligibility left (note: such requirements not necessary at Laurier). A rough estimate of replacement level, and the one we'll use, is based on the results for every player who had no more than 50 pass attempts in the last few years. "QB2" is not a terribly high bar to meet in CIS, after all.

(If you wanted to be cruel, you could just define replacement level to be the entire Waterloo team. But I'm not sure they were even that good.)

Once we have the difference between a QB's actual ANY and the ANY his replacement would put up in the same amount of playing time, that gives us that quarterback's "yards above replacement." Which we can then turn into points above replacement (15 yards per point), and wins above replacement (32 points per win), based on known conversions for CIS football.

So, again, we see that Quinlan had an awesome season, providing Mac with nearly five wins above a replacement-level quarterback. (My favourite stat: he passed for 1,341 yards in four playoff games, more than the entire York team had in eight regular-season games.) Quinlan's figure of 4.8 WAR doesn't necessarily mean that without him, the Marauders would have lost five more games; his backup was better than replacement level, after all. Marshall Ferguson was actually 10th in QB WAR, and could have started for several CIS teams. These numbers suggest that if you gave Ferguson all of Quinlan's playing time from last year, including the playoffs, Mac would only lose about half a win — which says quite a lot about Marshall Ferguson. But having a better backup doesn't make Quinlan's season any less valuable ... though, yes, he did put up a lot of these numbers against not-very-good OUA teams.

Austin Kennedy and Hec Crighton winner Billy Greene are second and third, but each is a win or more behind Quinlan. They would probably be first and second in any other year.
Next PostNewer Post Previous PostOlder Post Home

1 comment:

  1. It's a little early Rob but it will be interesting where Myles Gibbon's ranks next season. His "WAR" index will be at least a 2 or 3 I'm guessing.

    York currently sit 6th in the nation in yds/game after playing Queen's and Ottawa generating over 300 yards of offense and over 70 offensive plays against the highly touted Queen's defense.