Back in November, Cian Fahey tweeted me a simple question: “What is Alex Smith’s record in games where his D gives up 21 or more points?”
I made a note to run the numbers in the off-season, and guess what? It’s the off-season. Smith now holds a career record of 7-38-1 (including a 1-1 mark in the postseason) when his team [1]Yes, that is not the same thing as his defense. allows 21 or more points. That’s really bad, as it turns out. In fact, among quarterbacks who started such a game last year, only Ryan Fitzpatrick (5-43-0) has a worse career record.
How am I defining a “good” or “bad” record? Well, that’s somewhat subjective. But you might be surprised to learn that over history, teams have won about 25% of their games when allowing 21+ points, and there hasn’t been quite as much variation as you would expect. [2]Well, there’s some variation, but that’s to be expected. But if the rate was dramatically difference in ’92 compared to ’89, that’s probably a small sample size issue … Continue reading So while it would probably help a little to adjust for era, it’s not as necessary as I would have originally thought. [3]And since I’m lazy, I’m not going to adjust for era.
Here, take a look: [4]I will pose a question to the crowd. I know there is a very loud crowd that hates graphs where the Y-axis does not go to zero. Here, I could do that, which would show that the variation is not that … Continue reading
The best quarterback by this metric? Tom Brady. Brady has played in 106 games (including playoffs) where the Patriots allowed at least 21 points. In those games, Brady is 60-46, for a 0.566 winning percentage. Given that we would “expect” Brady to win 25% of those 106 games — that is, 26.5 games — Brady has exceeded expectations by 33.5 wins. All the usual caveats apply: wins should not be applied to the quarterback, there are lots of reasons not to read too much into the data, etc. But sometimes, it’s just good to do a data dump, even if you only find the information useful for trivia purposes.
As for Smith? There are 121 quarterbacks who started at least 40 games where their teams allowed 21+ points. Smith has the 8th worst winning percentage among that group, behind David Carr (0.065), Archie Manning (0.068), Fitzpatrick (0.104), Lamar McHan (0.116), Eddie LeBaron (0.120), Ron Jaworski (0.130), and Jeff George (0.156).
References
| ↑1 | Yes, that is not the same thing as his defense. |
|---|---|
| ↑2 | Well, there’s some variation, but that’s to be expected. But if the rate was dramatically difference in ’92 compared to ’89, that’s probably a small sample size issue and not an era issue. Rates are around 25% now, which they were in the ’80s and ’60s and even in ’74. |
| ↑3 | And since I’m lazy, I’m not going to adjust for era. |
| ↑4 | I will pose a question to the crowd. I know there is a very loud crowd that hates graphs where the Y-axis does not go to zero. Here, I could do that, which would show that the variation is not that great. But I think people know how to read axes, so they’d benefit from something like this over a lot of blank white space. What do you think? |
