To any of you football fans out there, I think the outrage over the Seahawks’ decision to throw it on the goal line is a classic example of hindsight bias. Throwing on the goal line is hardly unheard of, and they couldn’t run it 3 times anyway. This FiveThirtyEight article explains why throwing actually was a good decision. Anyway, everyone thinks that the decision to throw it was terrible, and I think that they’re being victims to the hindsight bias.
The 538 article is exactly the kind of context-free argument that justly gives ‘statistics’ and ‘rationalism’ a bad name. Yes, in game-theoretic terms you want to pass a certain amount of the time. Yes, there is a good argument for calling pass on that specific down. But the issue is not whether ‘pass’ in the abstract was a good decision, but whether the specific play-calling actions in their particular context were good. And they were not. They were indefensible.
Firstly, Seattle went out in a 3 WR group, thinking that this would get New England out of goal-line defense (and thus make it easier for the run play). This didn’t work, and they were foolish to think it would. This put Seattle in an awkward position where they would either have to run without enough blockers, or put the ball in the hands of poor players.
Secondly, Seattle ran the play out of the shotgun, making clear their intention to pass. There was no play-action or roll-out. This gives up all the game-theoretic part. If you are going to inform the opponent of your choice, you need to go with your strongest possible choice, whereas...
Thirdly, Seattle ignored their most favourable matchup. Seattle were the second-best offense in power-rushing playing the worst defense in stopping power runs. And instead...
Fourthly, Seattle went with a very unfavourable matchup in very unfavourable circumstances. They asked a receiver known for his downfield speed (and not much else) to fight for the ball on the goal-line. With the centre of the field cluttered, and all the defenders short, they called a quick slant into the centre of the field.
Yes, that play normally doesn’t result in an interception. Yes, there’s an element of bad luck there. But it’s also an example of really poor decision-making that ended up with Seattle essentially running the play New England would have chosen for them. They screwed up, and it’s embarrassing seeing the lengths people go to to defend the indefensible.
I would recommend this article highly for more depth.
The 538 article is exactly the kind of context-free argument that justly gives ‘statistics’ and ‘rationalism’ a bad name.
Are you sure? That sounds exaggerated. It definitely wasn’t context-free. Some examples:
“Let’s spot the Pats some yards, then, and assume the Patriots win1 about as often as a typical team in the AFA model would2 if they started on the 40-yard line. That would give them a 14 percent chance. Maybe that’s generous, but we’re looking for an upper bound.”
“But the Seahawks don’t have an average rusher; they have Beast Mode...”
Firstly, Seattle went out in a 3 WR group, thinking that this would get New England out of goal-line defense (and thus make it easier for the run play). This didn’t work, and they were foolish to think it would. This put Seattle in an awkward position where they would either have to run without enough blockers, or put the ball in the hands of poor players.
An awkward position? They couldn’t have ran on all three plays and if you have to have a pass play, 3 WR against a goal line defense is a good place to use your passing play.
Are you sure? That sounds exaggerated. It definitely wasn’t context-free. Some examples:
None of which consider the specific pass play that they chose to run, nor the specifics of the personnel matchup.
An awkward position? They couldn’t have ran on all three plays and if you have to have a pass play, 3 WR against a goal line defense is a good place to use your passing play.
In a vacuum, it could be a good place! But in the specific context it wasn’t, because New England still easily overmatched the Seattle receivers. Neither Baldwin or Kearse was ever going to get open against that coverage (and they didn’t), which meant that Wilson had one viable target—a downfield specialist not a possession receiver, covered by a specialist corner, running an inside slant (possibly the riskiest possible route in that situation). And all this from the shotgun, meaning there was no worry about a run. Carroll and Bevell knew all this before the snap, but they still chose to run that play. I think they must have known they’d been out-thought, but didn’t want to call a timeout there, and so went ahead anyway.
Suppose Seattle had done the kind of thing teams normally do when they pass from the 1-yard-line—come out showing run, then run a play-action, say with Wilson rolling out, with one tight end and one receiver to look for plus the chance of running it in himself, plus the easy option of throwing the ball away. Then 538′s analysis would make sense. But that’s not at all what happened. 538 doesn’t mention the passing numbers in that situation from shotgun formation. It’s like putting your money in penny stocks, and then defending your decision with the generic claim that equities are a good investment.
I see. I thought that you meant context as in Seattle/NE but it seems that you mean the formation and stuff. I think that what you’re saying makes sense now.
Personally I’d give more weight to:
The threat of running it from the shotgun.
The chances that a SEA receiver gets open vs. that goal line defense.
… and so I still don’t think it’s an awful decision. I think Wilson should have understood the situation and only made a really safe throw, and so the play call wasn’t that risky. But I do agree with you that play action would have been better, especially with a roll out.
I have always thought that any discussion of sports was sort of a playground for human bias and human error. So much passion for no real purpose. Affiliating with a team? The opposite of taking a principled position.
I guess it never occurred to me before that actually making this thought explicit might be valuable. But since discussion of the pass has reached less wrong, here it is.
What does have “real purpose”? Could you elaborate on this?
My thoughts are that most things we do for fun don’t really have “purpose”, that sports are no different, and that they’re an underrated way (amongst this community and most of society) to accomplish the goals of having fun, being in good shape and being happy.
Ah! I may have a meta-contrarian position to contribute:
This is not useful → This is useful for having fun → Fun is a valid goal, but this is a fairly ineffective way to have fun.
In the same way that people are routinely in error about how to improve everything else, they are routinely in error about what things are good at actually providing fun. And there is a familiar resistance to the direct application of thought to the problem, which relies on the normal excuses (“Isn’t it all subjective?”, “But thinking is incompatible with feeling! Haven’t you seen Spock?”).
Playing sports looks really good from an “effective hedonism” standpoint, even up to several hours a week. But for most people, I’m skeptical that regularly watching sports provides a decent long-term return, when done for more than a few hours every month or year.
Tangentially related: My local baseball team is far funner to watch than the top teams, because they make more mistakes, which leads to more unpredictable and exciting plays, but at the same time they’re still athletic enough that you’re not just watching children flounder around. In the same way, I really enjoyed the last superbowl.
To any of you football fans out there, I think the outrage over the Seahawks’ decision to throw it on the goal line is a classic example of hindsight bias. Throwing on the goal line is hardly unheard of, and they couldn’t run it 3 times anyway. This FiveThirtyEight article explains why throwing actually was a good decision. Anyway, everyone thinks that the decision to throw it was terrible, and I think that they’re being victims to the hindsight bias.
I agree. Given that they had one remaining timeout, the sequence of pass, run (timeout), gave them three chances to score instead of 2.
Still, its quite possible that a less risky throw might have been superior, even if it was lower chance of success.
As it was, that throw was inches away from being the game winning touchdown instead of the game losing interception.
The 538 article is exactly the kind of context-free argument that justly gives ‘statistics’ and ‘rationalism’ a bad name. Yes, in game-theoretic terms you want to pass a certain amount of the time. Yes, there is a good argument for calling pass on that specific down. But the issue is not whether ‘pass’ in the abstract was a good decision, but whether the specific play-calling actions in their particular context were good. And they were not. They were indefensible.
Firstly, Seattle went out in a 3 WR group, thinking that this would get New England out of goal-line defense (and thus make it easier for the run play). This didn’t work, and they were foolish to think it would. This put Seattle in an awkward position where they would either have to run without enough blockers, or put the ball in the hands of poor players.
Secondly, Seattle ran the play out of the shotgun, making clear their intention to pass. There was no play-action or roll-out. This gives up all the game-theoretic part. If you are going to inform the opponent of your choice, you need to go with your strongest possible choice, whereas...
Thirdly, Seattle ignored their most favourable matchup. Seattle were the second-best offense in power-rushing playing the worst defense in stopping power runs. And instead...
Fourthly, Seattle went with a very unfavourable matchup in very unfavourable circumstances. They asked a receiver known for his downfield speed (and not much else) to fight for the ball on the goal-line. With the centre of the field cluttered, and all the defenders short, they called a quick slant into the centre of the field.
Yes, that play normally doesn’t result in an interception. Yes, there’s an element of bad luck there. But it’s also an example of really poor decision-making that ended up with Seattle essentially running the play New England would have chosen for them. They screwed up, and it’s embarrassing seeing the lengths people go to to defend the indefensible.
I would recommend this article highly for more depth.
Are you sure? That sounds exaggerated. It definitely wasn’t context-free. Some examples:
“Let’s spot the Pats some yards, then, and assume the Patriots win1 about as often as a typical team in the AFA model would2 if they started on the 40-yard line. That would give them a 14 percent chance. Maybe that’s generous, but we’re looking for an upper bound.”
“But the Seahawks don’t have an average rusher; they have Beast Mode...”
An awkward position? They couldn’t have ran on all three plays and if you have to have a pass play, 3 WR against a goal line defense is a good place to use your passing play.
None of which consider the specific pass play that they chose to run, nor the specifics of the personnel matchup.
In a vacuum, it could be a good place! But in the specific context it wasn’t, because New England still easily overmatched the Seattle receivers. Neither Baldwin or Kearse was ever going to get open against that coverage (and they didn’t), which meant that Wilson had one viable target—a downfield specialist not a possession receiver, covered by a specialist corner, running an inside slant (possibly the riskiest possible route in that situation). And all this from the shotgun, meaning there was no worry about a run. Carroll and Bevell knew all this before the snap, but they still chose to run that play. I think they must have known they’d been out-thought, but didn’t want to call a timeout there, and so went ahead anyway.
Suppose Seattle had done the kind of thing teams normally do when they pass from the 1-yard-line—come out showing run, then run a play-action, say with Wilson rolling out, with one tight end and one receiver to look for plus the chance of running it in himself, plus the easy option of throwing the ball away. Then 538′s analysis would make sense. But that’s not at all what happened. 538 doesn’t mention the passing numbers in that situation from shotgun formation. It’s like putting your money in penny stocks, and then defending your decision with the generic claim that equities are a good investment.
I see. I thought that you meant context as in Seattle/NE but it seems that you mean the formation and stuff. I think that what you’re saying makes sense now.
Personally I’d give more weight to:
The threat of running it from the shotgun.
The chances that a SEA receiver gets open vs. that goal line defense.
… and so I still don’t think it’s an awful decision. I think Wilson should have understood the situation and only made a really safe throw, and so the play call wasn’t that risky. But I do agree with you that play action would have been better, especially with a roll out.
I have always thought that any discussion of sports was sort of a playground for human bias and human error. So much passion for no real purpose. Affiliating with a team? The opposite of taking a principled position.
I guess it never occurred to me before that actually making this thought explicit might be valuable. But since discussion of the pass has reached less wrong, here it is.
What does have “real purpose”? Could you elaborate on this?
My thoughts are that most things we do for fun don’t really have “purpose”, that sports are no different, and that they’re an underrated way (amongst this community and most of society) to accomplish the goals of having fun, being in good shape and being happy.
Ah! I may have a meta-contrarian position to contribute:
This is not useful → This is useful for having fun → Fun is a valid goal, but this is a fairly ineffective way to have fun.
In the same way that people are routinely in error about how to improve everything else, they are routinely in error about what things are good at actually providing fun. And there is a familiar resistance to the direct application of thought to the problem, which relies on the normal excuses (“Isn’t it all subjective?”, “But thinking is incompatible with feeling! Haven’t you seen Spock?”).
Playing sports looks really good from an “effective hedonism” standpoint, even up to several hours a week. But for most people, I’m skeptical that regularly watching sports provides a decent long-term return, when done for more than a few hours every month or year.
Tangentially related: My local baseball team is far funner to watch than the top teams, because they make more mistakes, which leads to more unpredictable and exciting plays, but at the same time they’re still athletic enough that you’re not just watching children flounder around. In the same way, I really enjoyed the last superbowl.
Good points. Particularly about watching vs. playing. I’m a lot more skeptical about the value of watching.