We're talking about quarterbacking

Here's the Top 10 passers in the CIS, as listed by passing yards:


  1. Matt Connell, McGill
  2. Josh Sacobie, Ottawa
  3. Michael Faulds, Western
  4. Dan Brannagan, Queen's
  5. Teale Orban, Regina
  6. J.P. Shoiry, Sherbrooke
  7. Adam Archibald, McMaster
  8. Quade Armstrong, Alberta
  9. Kelly Hughes, Mount Allison
  10. Justin Dunk, Guelph

Over at cisfootball.org, there's a post that references a newspaper story from this week that refers to Faulds and Brannagan, whose Western's and Queen's teams meet in the playoffs this weekend, being "ranked No. 3-4 in the nation in passing." It's understandable someone in the newspaper world might say that if they just had time for a quick check of the stats and don't follow CIS football too closely. However, seeing Faulds and Brannagan side-by-side illustrates the limits of that view.


The premier stat, at least in the NFL, is yards per pass; about 80 per cent of the time, whoever comes out ahead in this comes out ahead on the scoreboard. What's really needed is something that:
  • Shows average yards per pass but penalizes for sacks and interceptions (i.e.., "disaster plays);
  • Puts quarterbacks in a good team/bad team context, since play-calling, strategy and risk-taking is always dictated by the score;
  • puts it in a league context.


What this hopefully does is go from passing to quarterbacking... what team manages the ball best, as it applies to their starting quarterbacks?

Step 1 is ranking the 27 quarterbacks who have attempted at least 95 passes this fall by yards per pass, but with a twist: All the sack yardage has been subtracted, along with 50 yards per interception. That seems to be an accurate representation how much a pick costs a team in field position and points scored. Let's call it "adjusted" yards per pass.

The basic equation is:
[Passing yards — sack yardage — 18,468 -(interceptions x 50)] / attempts

All the QBs from good teams (i.e., at least a .500 record) go in one group, while all those from the the not-so-good teams go in a second group. This isolates better on QB performance, and besides, do we have to do math to know the Ottawa Gee-Gees'Josh Sacobie has had a better season than the U of T QBs?

  1. Josh Sacobie, Ottawa: 8.58 (238 attempts)
  2. Adam Archibald, McMaster: 8.44 (196 attempts)
  3. Dan Brannagan, Queen's: 7.84
  4. Erik Glavic, Saint Mary's: 6.89
  5. Bret Thompson, Saskatchewan: 6.84
  6. John Makie, Manitoba: 6.45
  7. Teale Orban, Regina: 6.41
  8. Jesse Andrews, Bishop's: 6.36
  9. Ian Noble, Laurier: 6.31
  10. Michael Faulds, Western: 4.70
  11. Justin Dunk, Guelph: 4.62
  12. Marc-Olivier Brouillette, Montreal: 4.04
  13. Keith Lockwood, Acadia: 3.99

Laval's Benoît Groulx (87 attempts) isn't listed, but for argument's sake, his adjusted yards per pass is 8.86.

Concordia and Calgary, which have had quarterback injuries and have run-oriented offences, don't have anyone listed. The Saskatchewan Huskies' Lawrence Nixon (79 attempts) also doesn't have enough attempts to be considered, but he's averaged 7.67 adjusted yards despite being sacked 11 times in his three starts.

The difference between Brannagan and Faulds, though, in this measure, is, take it away Tim Micallef, maaaaaaassssive.

Queen's doesn't pass as much as the Mustangs do, and while Brannagan's tendency to hang on the ball an extra second does mean they give up sacks (he's been dropped 19 times), he has a great "raw" yards per pass and doesn't throw interceptions (just 6 to Faulds' 15).

The "struggling team" quarterbacks:
  1. J.P. Shoiry, Sherbrooke: 6.96
  2. Matt Connell, McGill: 4.93
  3. Quade Armstrong, Alberta: 4.35
  4. Steve Snyder, St. FX: 4.23
  5. Kelly Hughes, Mount Allison: 3.86
  6. Doug Goldsby, UBC: 3.79
  7. Jason Marshall, Simon Fraser: 3.61
  8. Luke Balch, Waterloo: 3.40
  9. David Hamilton, Toronto: 3.24
  10. Evan Martin, Waterloo: 2.84
  11. Michael Hyatt, York: 2.82
  12. Dan Lumley, Windsor: 2.60
  13. Marc McVeigh, UBC: 1.99
  14. Andrew Gillis, Toronto: 1.80


Shoiry could go in either group since 3-4 Sherbrooke is still alive for a Québec conference playoff berth (they need to win at McGill and have Montréal lose at Bishop's on Saturday). Again, with the exception of Shoiry, the gap between Matt Connell and the rest of the pack really shows how good he has been in a trying situation for McGill this season. His overall rate is reined in by McGill's lack of a running game or a defence.

The next step is converting this to a league context, but it will have to wait till later tonight.
Next PostNewer Post Previous PostOlder Post Home

4 comments:

  1. Thanks . that must have been a lot of work. One problem with fine tuning to such a degree is that the stats you have to work with are not always that accurate. the CIS and particularly the OUA stats have often been criticized for that.
    I remember seeing one game where a QB was sacked five times. Later the official stats said he only lost nine yards or 1.8 yards a sack. Having seen it I knew it was not true as he was losing six or seven yards on every sack.
    I am not sure how you are going to adjust for a league, or account or a starting QB being pulled at the half or end of the third against weaker teams, thus reducing his chances of really racking up the yards per pass attempt. some teams pull their QB in such situations and some do not.
    Anyway, an interesting exercise. From the prelim results , you are making a stronger case as to why Sacobie has to be a candidate candidate for the HEC.

    ReplyDelete
  2. Are your stats updated? I think you have omitted Nixon's stats, last weekend, against the Bears. He has thrown 103 passes (he was 33 of 43 last weekend).

    ReplyDelete
  3. The stats are off the CIS website for everyone... they have Nixon throwing 74 passes.

    The best way to adjust for league is to figure out the league's adjusted yards per pass, then divide a QB's stat by that... alas, I didn't have time to do that today.

    Taking U of T and York out of the equation would be good for the OUA...

    ReplyDelete
  4. Good work, but the CIS website is in error in listing Nixon as only having 73 pass attempts (What he does have is 73 pass completions). He has 76 pass attempts in the last two games alone. I was in the stands watching both games, and know this for a fact. Check it out here: http://www.canadawest.org/football/stats/confldrs.htm#conf.wi2

    ReplyDelete