"Points? 𝗙𝗨𝗖𝗞 points. We have the lead and the ball. We are declaring this game over. Stop us if you want to."
And they didn't! They didn't want any part of it! The softies wanted to go home.
In my foolish undergrad days watching Pac-10 football, I thought the point of an offense was to score, ideally a touchdown. This is *an* important skill for an offense, but it is not the only or the most important skill.
To get a bit nerdy, sports like soccer and hockey tend toward Poisson score distributions, while basketball gives bell-shapes. This is true whether you look at all scores in a season, or simulated scores for a particular game. These are smooth distributions, the chance of a particular score is close to the chance of scoring one more or one less. Baseball is slightly less smooth, because runs come in bunches.
But football is rough all the way. 13, 14 and 17 points are common, 16 is moderately common, 15 is rare. All this is thanks to Walter Camp who invested the system. The key features are different possible scoring amounts and a moderate number of scoring events per game.
If you bet non-football sports, most of your attention is on which team is better and by how much, with factors like weather, team strategies, officiating definitely secondary. But in football those and other factors make a big difference in the likelihood of, say, a three-point or eight-point difference in the final score. If you're trying to choose between a -7 point spread at -110, or a -7.5 spread at -105, you need to do a ton of analysis in football. In other sports you can generally get away with assuming a smooth distribution fit with historical data.
I love the idea of coming up with a term to describe very rare scores that aren't original. I don't think you can just do lower case and upper case scorigami though. Jon Bois has done too much coverage on scorigami as a unique score to start calling rare-but-not-unique scores scorigami.
4th down. I have an idea that the algorithms are still underestimating the value of going for it. The statistics are based on the team structures as they exist, but if they were to take into account the opportunity cost of carrying specialists (both in terms of salary and roster spot), the value of going for it on 4th down would be even higher, at least my intuition says it would. To think of it in practical terms, what if you took the salary and roster spot from your long snapper and used it instead on cornerback or pass rusher? What does that look like in terms of WAR?
Don't think that would be a good idea. A) the amount the long snapper is paid is tiny compared to positions like WR so it will barely increase your budget. B) If you have no long snapper and literally can't punt you will end up going for it in situations where that is suicidal, eg 4th and long near your own goal-line
I'm not assuming that you could swap a long snapper for a #1 WR but rather for a 4th or 5th WR or CB - Long snapper is about $1M-$1.5M against the cap, which would be about the same as the 100th-150th highest paid WR (4th or 5th on the team, probably on a rookie contract). Further, and I have no idea if this is true, but I would assume that your regular center would be a serviceable long snapper in obvious punt situations. But this is just one example. Again, I don't know enough about the nitty gritty of special teams to know this, but does the math change on having a specialist who is a great gunner versus someone who might be a backup running back if you go for it more often on 4th? I'm genuinely asking the question, not assuming my hypothesis is right.
The snap for a punt is about 15 yards, compared to about 6 yards in shotgun which I think is the furthest the center would usually snap it. So it's a substantial difference, but I've no idea how well centers could do it. The fact that teams employ long snappers at all does suggest it's not an easy skill for someone else to do, though.
Nate, this is an exceptionally fun article.
Behold, the sexiest football drive possible: 16 plays, 68 yards, 8:38 time of possession, kneel down, end of game.
https://www.espn.com/college-football/playbyplay/_/gameId/401135287
"Points? 𝗙𝗨𝗖𝗞 points. We have the lead and the ball. We are declaring this game over. Stop us if you want to."
And they didn't! They didn't want any part of it! The softies wanted to go home.
In my foolish undergrad days watching Pac-10 football, I thought the point of an offense was to score, ideally a touchdown. This is *an* important skill for an offense, but it is not the only or the most important skill.
I kind of wish it had been christened the “scoredoku”. That feels more apt.
The scorigami character of American football results makes it much more interesting for bettors (I wrote about this at length here: https://www.bloomberg.com/opinion/articles/2023-02-10/super-bowl-how-do-bettors-love-football-let-us-count-the-ways).
To get a bit nerdy, sports like soccer and hockey tend toward Poisson score distributions, while basketball gives bell-shapes. This is true whether you look at all scores in a season, or simulated scores for a particular game. These are smooth distributions, the chance of a particular score is close to the chance of scoring one more or one less. Baseball is slightly less smooth, because runs come in bunches.
But football is rough all the way. 13, 14 and 17 points are common, 16 is moderately common, 15 is rare. All this is thanks to Walter Camp who invested the system. The key features are different possible scoring amounts and a moderate number of scoring events per game.
If you bet non-football sports, most of your attention is on which team is better and by how much, with factors like weather, team strategies, officiating definitely secondary. But in football those and other factors make a big difference in the likelihood of, say, a three-point or eight-point difference in the final score. If you're trying to choose between a -7 point spread at -110, or a -7.5 spread at -105, you need to do a ton of analysis in football. In other sports you can generally get away with assuming a smooth distribution fit with historical data.
I love the idea of coming up with a term to describe very rare scores that aren't original. I don't think you can just do lower case and upper case scorigami though. Jon Bois has done too much coverage on scorigami as a unique score to start calling rare-but-not-unique scores scorigami.
Nate you can't tell us a delightful story about your childhood gambling ring without telling us how much the house's cut was!
I'm waiting to see a game that ends 6 - 1. It's possible now.
4th down. I have an idea that the algorithms are still underestimating the value of going for it. The statistics are based on the team structures as they exist, but if they were to take into account the opportunity cost of carrying specialists (both in terms of salary and roster spot), the value of going for it on 4th down would be even higher, at least my intuition says it would. To think of it in practical terms, what if you took the salary and roster spot from your long snapper and used it instead on cornerback or pass rusher? What does that look like in terms of WAR?
Don't think that would be a good idea. A) the amount the long snapper is paid is tiny compared to positions like WR so it will barely increase your budget. B) If you have no long snapper and literally can't punt you will end up going for it in situations where that is suicidal, eg 4th and long near your own goal-line
I'm not assuming that you could swap a long snapper for a #1 WR but rather for a 4th or 5th WR or CB - Long snapper is about $1M-$1.5M against the cap, which would be about the same as the 100th-150th highest paid WR (4th or 5th on the team, probably on a rookie contract). Further, and I have no idea if this is true, but I would assume that your regular center would be a serviceable long snapper in obvious punt situations. But this is just one example. Again, I don't know enough about the nitty gritty of special teams to know this, but does the math change on having a specialist who is a great gunner versus someone who might be a backup running back if you go for it more often on 4th? I'm genuinely asking the question, not assuming my hypothesis is right.
The snap for a punt is about 15 yards, compared to about 6 yards in shotgun which I think is the furthest the center would usually snap it. So it's a substantial difference, but I've no idea how well centers could do it. The fact that teams employ long snappers at all does suggest it's not an easy skill for someone else to do, though.
explain this to me like I'm five and never watched this sport
Great read. Thanks