≡ Menu

Over at Footballguys.com, I identified which quarterback statistics are repeatable and which ones are most likely to regress to the mean. I also ran a regression using touchdown length as my input and future touchdowns as my output.

From 1990 to 2011, 188 different quarterbacks started at least 14 games and thrown 300+ passes in one year, and then attempted at least 300 passes for the same team the next season. After analyzing the lengths of each touchdown pass for those quarterbacks, I discovered the following:

  • For every one-yard touchdown pass in Year N, expect 0.70 touchdowns in Year N+1
  • For every two-to-five-yard touchdown pass in Year N, expect 0.56 touchdowns in Year N+2
  • For every six-to-ten-yard touchdown pass in Year N, expect 0.77 touchdowns in Year N+2
  • For every 11-to-20-yard touchdown pass in Year N, expect 0.70 touchdowns in Year N+2
  • For every 21-to-30-yard touchdown pass in Year N, expect 0.22 touchdowns in Year N+2
  • For every 31-to-50-yard touchdown pass in Year N, expect 0.33 touchdowns in Year N+2
  • For every 50+ yard touchdown pass in Year N, expect 0.33 touchdowns in Year N+2

If a team throws touchdowns from inside the red zone, that reveals an offensive philosophy that is good for your fantasy quarterback. On the other hand, 21+ yard touchdowns might make the highlight feels, but are very unpredictable from year to year. What does that mean for 2013?

You can view the full article here.

  • Shattenjager

    Good work, as always.

    I assume the Year N+2 at the end of each line after the first is supposed to be Year N+1.

    I can’t help but wonder how much of a difference it makes when there is a coaching/system change and whether there is any similar trend if you look at the numbers for a coach or offensive coordinator and whether the ratio of long vs. short TDs is more related to the QB or the OC/HC. While I would think that the correlations would be similar, I could see the possibility that an OC/HC’s system and play calling sets up short or long TDs better or worse and thus the year-to-year correlation is better or worse than it is for a QB. Further, while knowing that longer TDs are less repeatable and knowing who threw the most longer TDs last year gives us a handle on whose totals are more repeatable, it seems that changes in OC/HC could mitigate possible regression there, depending on whether it’s the QB or the play calling that causes the distribution in the first place. (This seems like a very messy paragraph, but I think that I’ve explained what I’m wondering about well enough. Sorry if it’s too muddled.)

    Wondering about this relationship, I looked a bit (albeit rather quickly) and Marc Trestman’s history (I’ve become absolutely fascinated with him!) seems to be rather consistent that his QB gets about half of his TDs in the red zone, whether it’s Steve Young or Jake Plummer (or even Elvis Grbac in a very small sample). Meanwhile (unsurprisingly, given your findings), his teams seem to be very inconsistent in terms of total passing TDs.

    • Chase Stuart

      It’s a good question but I don’t know if we’d have enough data to get a general sense. I think some really pass-happy coaches like Reid and Martz are probably that way, while regardless of the coach, Manning, Brady, and Brees will throw lots of short TDs.

  • Wade Iuele

    For clarification, for each length of touchdown input, you have a future touchdown output…is that future touchdown of the same length as the input? Or is the output a touchdown of any length in year n+1?

    • Chase Stuart

      That’s the output of a touchdown of any length in the next season.

      • Chase Stuart

        I should add that there is also a baseline number of touchdowns to expect from future quarterbacks. I forget, but it was something like 12 + 0.7 TDs from within one yard, 0.56 TDs from 2 to 5 yards, etc.

  • WOW just what I was looking for. Came here by searching for your
    skin is pale white and ice cold quote