Football (American) strategy question

October 12, 2009 10:09 AM

Football (American) strategy question: You're down 15 and score a TD with about enough time left to kick-off, try to force the other team to punt, then get the ball back with a chance to tie. Why do all teams opt for the extra point rather than the 2-point conversion?

I guess the strategy is get within 8, then hopefully the team's pumped up to make a stop on D, then tie the game up at the end of regulation. But wouldn't you want to work with all available info? Shouldn't you go for 2 right away?

If you convert, carry-on with the original strategy, except you don't have to go for 2 on the 2nd TD (in fact, if you're feeling aggressive, you could go for the win with a 2nd 2-pt conversion on the 2nd TD).

If you don't convert, you're now down by 9, and you know you must go into crisis-mode by attempting the onsides kick. You allow yourself the maximum amount of time for strategy adjusting.

My thinking is this: the coaching staff has failed at its job if they score the 1st TD, get the extra point, get the ball back, score another TD with about no time remaining, then fail to convert the 2-pt conversion. You had the opportunity to know whether or not you were going to convert the 2-pt conversion a long time ago with time to adjust your strategy accordingly. But you chose to postpone the inevitable, and leave yourself with no time to adjust.

Where am I wrong here?

posted by glenngulia to sports, hobbies, & recreation (50 comments total)

Sorry, I'm an idiot and accidentally posted everything in question line, without utilizing the more inside feature.
posted by glenngulia at 10:11 AM on October 12

I'm not joking here, I think 90% of NFL coaches make decisions like this because they are afraid of deviating from approved norms. Going for 2 is seen as "not real football", and you only do it when absolutely necessary.

Coaches get vicious criticism when they try something new and it doesn't work. If they try something new and it does work, it's still viewed as suspicious, maybe even gimmicky.
posted by bluejayk at 10:15 AM on October 12

My understanding is that you always go for the sure points. If you dont get the two point conversion the first time, you are basically dead in the water. Onside kicks are rarely ever succeed (less than 10 percent) so rather than lose the shot right away, you take the easy one point, and go for the tougher 2 points when you have the chance to win.
posted by boyinmiami at 10:16 AM on October 12

In my mind, getting a one-point conversion and pulling within one possession is better because what if:

You go for two and miss. You're down 9.
The other team drives for a little while and throws a interception which is returned for a touchdown, or maybe they fumble the kickoff, anything.
You've got a touchdown! But now it's still a two possession game and you don't have enough time to get another score. If you had kicked the extra point, you've got a chance to tie.

I know it doesn't seem intuitive, but NFL coaches have to play PR, and that means playing somewhat conservatively. I know I would be kind of upset, as a fan, if your scenario occurred. My guess is you'd see a lot more of your scenario in lower levels of football, though.
posted by ofthestrait at 10:16 AM on October 12

The general rule of thumb is you take points when you get them. Particularly in the NFL, points aren't always that abundant. You get your points until it is absolutely necessary to go for the 2-pt conversion.
posted by jmd82 at 10:18 AM on October 12

You point it out in your question. If you go for 2 and miss on the 1st TD, you still have to score twice to win. If you go for the extra point, theoretically you only have to score once. Having to score twice with the clock dwindling down is nearly impossible, as converting an onsides kick is even less likely than getting a 2-point conversion.
posted by hwyengr at 10:19 AM on October 12 [1 favorite]

The kick is the safe choice, percentage-wise. Most coaches these days seem rather risk-averse so it gets the nod almost every time. Besides if you miss one or both 2-pt conversions you're back in the same position anyway.
posted by tommasz at 10:22 AM on October 12

1,If you fail on the 2 point conversion, you will be down by 9, this means that you will have to score at least twice more, a touchdown and a fieldgoal, or 2 touchdowns. This means you need to get the ball back twice, and with such little time, game over, pretty well.

2.If you get the 2 point conversion, you will be down by 7, and can either tie with a converted touchdown, or win with a touchdown and the 2 points. Sounds good, but the chance of making a 2 point conversion is alot less than the 1 point.

3.If you go for the extra point, you will be down by 8, and can tie with a touchdown plus 2 point conversion. This means you can tie with just one possession.

So a coach would look at the risk here: If you fail the two points, then you have pretty well zero chance of winning; if you make the one point, you have a slightly better chance of winning. Since a 2 point conversion is not a guarantee, and one point pretty well is, that is what they will do.
posted by BozoBurgerBonanza at 10:24 AM on October 12

Agreeing with above. It's better to score the easy extra point and keep yourself in the game. If you go for the two point conversion and fail, it's game over even though there's time on the scoreboard.
posted by 2oh1 at 10:28 AM on October 12

Don't know anything about football strategy, but that's never stopped anyone on the doctor or lawyer questions, so...

What you say in the OP sounds right if the only uncertainty is whether you make the 2pt conversion or not. But in fact, a lot of crazy things can happen in the closing minutes of a game (turnovers, safeties, etc) and if you're behind by that much, you're probably hoping that one of those crazy things actually does happen. So my take is that they're waiting as long as possible to see if some low-probability event comes along first to save their bacon.

I mean, you still might be right, but you'd have to do the statistics to back it up.
posted by sesquipedalian at 10:46 AM on October 12

I also agree it's better to (virtually) guarantee making it a one-score game than risking still needing two scores, with only the possible benefit of having a slightly easier one-score game. But I'm not following this argument:

But wouldn't you want to work with all available info?

That's a wash, isn't it? Following the first touchdown, depending on your action and its success, you now know you are either 7, 8, or 9 points behind: in each case you have "all available info" and know what you need to do.

You had the opportunity to know whether or not you were going to convert the 2-pt conversion a long time ago with time to adjust your strategy accordingly. But you chose to postpone the inevitable, and leave yourself with no time to adjust.

That only applies if there's enough time on the clock to reasonably expect that you'll get the ball back twice more. Say, with 10 minutes or more left in the game. And I think you'll see coaches do go for the conversion in that situation with that much time left. But if you can't reasonably expect to get the ball back twice more with normal play, then if you end up 9 points down after the failed conversion, you'll need an onside kick, and an expected onside kick has a pretty low probability of success. Doesn't the opposing team also benefit also having "all available info?" Which team benefits more from knowing you still need two scores (including recovering an onside kick)?
posted by DevilsAdvocate at 10:53 AM on October 12

You can treat it as an expected value calculation, but the problem there is that there seem to be so few available statistics on the success rate of the two point conversion. In that article the available estimates vary between about 40 and 51 percent (compared to 96 percent for the extra point) which means it just barely teeters on the knife edge as to whether it's beneficial or not, which pretty much seals its fate given that the extra point is so reliable.
posted by Rhomboid at 10:55 AM on October 12

The odds are better for a tie than a win.
posted by rhizome at 10:59 AM on October 12

I think it's a combination of a few things. The basic idea is that you keep yourself alive by only being down 8. If you go down 9 and the other teams scores another TD and puts you down 16, then you're really in trouble.

But in an end-of-game scenario where if the other team scores again, you've definitely lost anyway, I think there might be some sense to going for the two first.

I think this:

I'm not joking here, I think 90% of NFL coaches make decisions like this because they are afraid of deviating from approved norms. Going for 2 is seen as "not real football", and you only do it when absolutely necessary.

Coaches get vicious criticism when they try something new and it doesn't work.

also has a lot of validity. NFL coaches fear for their jobs. This has led to strategies like the Prevent defense, which everyone knows just doesn't work, but is just now falling out of fashion. The reason it stuck around so long is that losing a game by giving up a long final drive bit by bit doesn't look THAT bad. Blitzing and giving up one big 90 yard play looks very bad. Even though the former happens constantly and the latter is very rare, it's an open secret that most coaches worry about HOW they win or lose, not just if.

So eventually some pioneering coach will go for two up front, it will work, and everyone will copy and praise his genius.
posted by drjimmy11 at 11:04 AM on October 12

Let me try again:

I break it into 2 scenarios:

A- You successfuly convert a 2-pt try in the final few minutes of the game. If that TD + 2-pt conversion is coupled with another TD + extra point, congratulations. You just scored 15 points in the final 4 minutes of regulation and are now headed into overtime. Notice: if the 2-pt try was successful, it doesn't matter whether came on the 1st or 2nd TD.

B- You are unable to convert a 2-pt try in the final few minutes. B diverts into 2 different scenarios, depending on whether you went for 2 on the 1st or 2nd TD.

BI- You opted for the easy point on the 1st TD. You kicked-off with 4 minutes left. The other team killed some clock, but you had some timeouts and eventually forced them to punt. You get the ball back with 2 minutes left. You then marched down the field and scored another TD with only seconds remaining. What a comeback! Great, covert the 2, and you're tied! Oh no, you failed to convert the 2. Game over. An onsides kick with only seconds remaining ain't gonna do you any good.

BII- You went for 2 on the 1st TD and fail. OK, game is more than likely over because you're down 9 (2 possessions). But hey, you have info earlier than you had it in BI. Down 2 possessions with 4 minutes left. A win is still unlikely, but it's not over as in BI. I can go on forever about the strategies available (onsides kicks, timeout mgmt on D, more aggressive playcalling on D), but you catch my drift, right?
posted by glenngulia at 11:08 AM on October 12

It's risk-reward and future uncertainty. The two-point is relatively high risk compared to the one point. You don't want to risk the two point until you know you need to, and you don't know if there's going to be a turnover or something in the upcoming minutes that would make the risk not worth it. So you save the risk for when there's no future.
posted by RikiTikiTavi at 11:10 AM on October 12 [1 favorite]

There's surely a psychological side too. You don't really want that downer of scoring then missing the conversion which puts you behind by two scores.
posted by smackfu at 11:33 AM on October 12 [2 favorites]

Yeah, that's a good way of putting it -- you're concentrating on the idea that by knowing that you failed the two point conversion earlier you can scramble accordingly in the remaining time, but you have to balance that against the idea that by delaying the risky move something else might intervene that would make it unnecessary to take that risk at all.
posted by Rhomboid at 11:34 AM on October 12

Let me try again:

You seem to have your mind made up, but I'm nthing ....

Going for the (single) extra point gives you a *really* high probability of making it a one possession game. Going for 2 gives you only a moderate chance of the same. It's a simple as that. Having 'info" earlier really does nothing to change that.
posted by and hosted from Uranus at 11:38 AM on October 12

The goal is to get two touchdowns, one 2-point conversion and one field goal.

You are the coach, and you get to choose which order you want to accomplish these two tasks. 2-point then field goal, or vice-versa?

I think it's clear that you want the 2-point conversion to happen last. If someone tries a 2-point conversion while being 15 points behind, there isn't nearly as much pressure on the opposing team to stop it, and not being under pressure will probably give them a better chance of preventing it.

Whereas if you make the 2-point conversion happen last, there will be intense pressure on the opposing team to stop it, and that is going to hurt their chances. It also makes the game more interesting than the other scenario, and I'm guessing that's one of the jobs of a coach: when faced with two nearly-identical options, pick the most exciting one so that people will have fun and come back.
posted by helios at 11:38 AM on October 12

In BII the other team knows with absolute certainty that they can just run down the clock. They can call zero-risk playing on offense. This means that you are done. In the scenario where you are down 8, that's not the case. The other team knows that you could tie and will try to actually score.

You're conditioning on whether or not the conversion will happen, but that's not how the game is played. The order of events determines the opponents actions just as much as yours. Also, you look worse taking the risk and being anticlimactically defeated early in the game compared to it hinging on the last second.
posted by a robot made out of meat at 11:43 AM on October 12 [1 favorite]

One point conversions aren't guaranteed, either. A block, a miss, and its irrelevant if you have tried for one or two.
posted by wayofthedodo at 11:48 AM on October 12

Thanks, all, for the responses.

I figured a lot of the responses would focus on the “take the easy points that minimize the amount of possessions you trail” method. And judging by how most NFL coaches choose to proceed, you may all be correct.

But I still feel I need to re-word my stance at least one more time…

If you’re unable to convert a 2-pt try in the final few minutes of the game, then by settling for an 8-pt deficit after the first TD, you were really never within one possession in the first place. I’d rather have the knowledge that I was in a 2 possession game with 4 minutes left than with 10 seconds left.

Those that say that converting a 2-pt try is more likely after the second TD because the opposing D is tenser (or more tired, I guess) make a compelling argument, but I’m not sure I buy it.

Again thanks, all!
posted by glenngulia at 12:12 PM on October 12

I believe in option BII all reactions would be pretty much the same - you manage the clock and the play-calling in such a way that you have your quickest route to the endzone. You could take even wilder, lower-percentage chances of moving down field.

Going for, and missing, a two-point conversion, being down two scores, allows all but the most dedicated fans to change the channel assured that the outcome is known. Delay that risk and you keep your audience. It's a good business AND entertainment option.
posted by yamel at 12:22 PM on October 12

A lot of it is tradition/habit, but it's basically the "extend the game" mantra. By getting to within one score, your "force" your opponent to "keep playing."

Down by 8, you might still onside kick... so do they put the hands team out and risk getting pinned deep in their end by a poor return, or do they put out the normal return team and risk getting onsided? Do they just run every down once they have the ball to take maximum time off the clock no matter what, or risk throwing for the game clinching first down? Might they get the clock down to 10-20 seconds and just run out of the back of their own endzone on 4th down so they can pin YOU further away from the endzone, knowing your only recourse will be a desperation hail mary? They have to keep thinking, and making decisions, and in every decision there's a chance for a mistake that you can take advantage of. Down by 10... they'll just run three safe plays and punt, no risk required to leave you without enough time to score twice.
posted by Pufferish at 12:30 PM on October 12

If you’re unable to convert a 2-pt try in the final few minutes of the game, then by settling for an 8-pt deficit after the first TD, you were really never within one possession in the first place

Okay, so a few problems.

First: admit the fact that the guys who make their livings doing this might have some idea about what they're doing and that you might just be wrong.

Second: from a momentum standpoint, it's going to be easier to get that 2-point conversion after scoring, stuffing the other team and scoring again -- you have all sorts of adrenaline and momentum working for you. Also, if you miss that initial 2-point conversion (which you have a very high likelihood of) you have just taken all of the wind out of your comeback, the other team is hyped up, and the odds of them scoring on you (something that would probably effectively end the game) has just sky-rocketed.

Third: You can just forget about going for the win. Nobody's going to do that. That's a false "benefit" because no coach in his right mind is actually going to risk losing the game on a 50/50 chance rather than having a nearly 100% chance of going to overtime and getting a chance to win the game there.

Fourth: You're assuming that if you fail to get the 2-point conversion after the second touchdown that you'd have failed to get it after the first. This isn't true. Even if it was, the same can be said of the PAT . . . the kick is not guaranteed. In either case, you're dealing with a late-in-the-game attempt to tie -- an attempt you do not know the outcome of.

Finally, as pufferfish points out, you've completely neglected the "what is the other team going to do" side of the strategy. You've gotta force them to make tough decisions. If you go for two and miss, you've just given them a really, really easy shot to end the game without doing much of anything.
posted by toomuchpete at 12:39 PM on October 12

glenngulia, that's fine if you're flipping coins, but you are playing against a strategic team trying to run down the clock. The opposing team could theoretically kneel down 6 times with two possessions ahead and take 0% risk over a fumble, interception, or -even better- a safety. With only 8 points down, you can limit them to 3 running plays (which you can minimize the impact with time-outs), and you have control over the field AND more importantly the clock.
posted by yeti at 12:39 PM on October 12

I think the answer to this depends alot on exactly how much time is left and how high scoring the game has been (which affects how likely you think it will be that you will score twice more if needed).

...you were really never within one possession in the first place. I’d rather have the knowledge...

If you were the only role agent in this game, your position would be completely correct. But you are playing against another person, and he wants to have that knowledge too. By going for 1 point, you deny him the knowledge of whether it's really a one possession game or not. Because he's on offense next, that lack of knowledge probably affects him more than it does you because he is forced to assume it is a one possession game. How valuable that knowledge is overall to both sides would be really difficult to quantify, but it's not as clear cut as saying "the more knowledge I have right now, the better".

Harold Sackrowitz is a statistician who has consulted for various NFL teams. An article and his chart on conversions can be found here (pdf). Unfortunately, the entry for this situation (down 9 after scoring a TD, < 4 possessions left) is blank, meaning it doesn't matter.
posted by ghostmanonsecond at 12:40 PM on October 12

Hmm, maybe third time's a charm because I do catch your drift better now than before. I still say the "simple" (aka one-possession) answer is the real reason, but momentum and other psychological factors probably do play into it. If you fail the 2 -point on the first TD, yes you now know for sure it's a two-possession game and that knowledge is a good thing, but you've killed all the momentum you just got by scoring a TD. Take the point and your team is still saying "get the ball back and we can tie this thing". Fail the two and the team will know the game is pretty much over. If the team is down 8 with 4 minutes left, nobody's mind is over-thinking about still needing to get a 2-point conversion...they're playing to get the ball back and to score. If a team is down 9 with 4 minutes left, they know there is little to play for except for hoping for something flukey.

The only situation where, if I were a coach, I might think along the line you've set out is if I had all my timeouts left and we had just run the ball down their throats to the end zone. If they can't stop my rush, I might be tempted to run at them again and count on my defense getting the ball back twice if we fail. But, if we had relied on passing to score (which is more probable in a late game scenario), the risk of failing to get 2 would just be too much.
posted by and hosted from Uranus at 12:41 PM on October 12

Put another way: the game hinges on a 2-point conversion.

Do you want to try this 2 point conversion on a defense that has played 3.5 quarters of football and probably still feels like they're in control of the game, or do want to try it against a defense that has played four quarters of football and feels demoralized (and very likely extra worn down) after giving up 2 touchdowns in the closing minutes?
posted by toomuchpete at 12:42 PM on October 12 [2 favorites]

Third: You can just forget about going for the win. Nobody's going to do that.

Not nobody. It happens. Rarely, but it happens.
posted by yeti at 12:43 PM on October 12

yeti: Not nobody. It happens. Rarely, but it happens.

See also, and for (arguably) higher stakes (national college championship vs. regular-season NFL win).
posted by hangashore at 1:16 PM on October 12

I agree with most responses here-- you don't get the 2-point conversion, game is essentially over. You get the PAT, the game continues-- and, a small side benefit, you force the other team to try to move the ball, and maybe enhance the chance for a turnover. Nobody cares if they punt up 9 with 30 seconds to go.

But people are right that coaches have irrational biases-- they're more likely to go for the FG from the 32 when it's 4th and short than if it's 4th and long.
posted by ibmcginty at 1:29 PM on October 12

Okay, sure, so in some very bizarre situations a coaching staff is going to see something bizarre in the defense and think "we can get the win RIGHT HERE by running this magical play" or they're going to hate the idea of playing sudden death (in the NFL) with the current team playing the way they are.

But the Denver example isn't very analogous. They hadn't scored since the first half, whereas San Diego had scored in every quarter. Usually twice. Overtime didn't favor Denver in the slightest . . . it's not as though Shanahan got two two-point conversions in the final quarter as some sort of strategic move. If you're down 15 in the fourth quarter, nobody is going to make the win-or-lose decision on the first touchdown when you can put it off. When your team scores two unanswered touchdowns in the fourth quarter, you're going to feel pretty good about your chances in OT.

And hangashore's example is from (essentially) a completely different sport: ties still existed in 1984 and in that particular situation, a tie could easily have killed his chance for a national title . . . essentially meaning that a tie is a loss.
posted by toomuchpete at 1:43 PM on October 12

Those in favor of going for two after the first TD assume that (1) the losing team will act more urgently if trailing by 9 than if trailing by 8; and (2) the success of the two-point conversion attempt is independent of when the two-point conversion is attempted.

No one would dispute that a team trailing by 9 would act more urgently than a team trailing by 8. A valid counter-argument would have to demonstrate that a two-point conversion attempt is more likely to be successful the second time. I don't think there's enough data to evaluate this. Where this original poster's argument usually breaks down -- as with my beer-drinking buddies and their failure to come around -- is a lack of desire by the listeners to walk through the entire argument.

If the assumptions are sound, the original poster's conclusion is valid. But this is a complicated scenario, and it is not obvious that the assumptions lead infallibly to the conclusion. An additional question: What is the best way to introduce this argument so that others will allow you to convince them, or at least disagree with your assumptions rather than your logic?

By the way, if you trail by 22, there is a much clearer probability-based reason to go for two the first time. Trail by 8, the choice is trivial. Only with 15 comes the confusion...
posted by RobinFiveWords at 1:46 PM on October 12

I would say there's a psychological advantage by delaying the 2 point conversion. You have to realize your chances are relatively slim to begin with. If you don't convert the 2 point conv. on an early try you still have a shot, BUT... you're team is going to be deflated by the miss AND their remaining task is now at least twice as hard (2 possessions needed rather than 1). Also, you lose any sense of surprise if you go for an onside kick, since it's now going to be required (or at least advisable).

On the other hand, if you get the extra point, you have the other team reeling a bit, you can kick deep or go for the onsides kick. And if you march down the field and get another score, the defense for the 2 point conversion this time will REALLY be up against the wall, knowing they've given up two quick touchdowns and the game is on the line - all the pressure will be on them.
posted by lubujackson at 2:01 PM on October 12

Being within 1 score radically affects the way the other team will play its next posession. They will take more risks and aim to obtain a first down. This will create chances for the defense. The key variable here is not the score, it is time.

Also there is no such thing as "it was never a two-score game anyway." You are applying hindsight bias.
posted by Ironmouth at 2:12 PM on October 12 [1 favorite]

The urgency issue is a red herring at best and a counter-argument at worst. It can be assumed that a team will play as urgently as it needs to, so if you are proposing a strategy because it will make the players play more urgently, the reason is likely because your strategy put them in a position where they had to.

I'm not seeing where the difficulty is.

You reach a point in the game where you are down by 9 points and have to choose between kicking or going for two. In both cases, if you succeed, you'll be within one score of tying the game and will need another touchdown. In both cases, failure essentially ends the game.

As the coach, do you choose the ~95% chance to pull to within one score, or do you choose the ~60% chance to pull to within one score?
posted by toomuchpete at 2:16 PM on October 12

You had the opportunity to know whether or not you were going to convert the 2-pt conversion a long time ago with time to adjust your strategy accordingly.

Put another way, by kicking the xp, the coach has the opportunity to deny the other team the chance to know if the two-pointer and adjust their strategy accordingly.
posted by Ironmouth at 2:18 PM on October 12

Don't forget, the other coach has all the info you do.

Let's say you take one. It's a one-possession game. With 4 min. left, the opposition doesn't know whether you might kick off deep or go for the onside, and might put the wrong receiving unit on the field...or you might change your decision whether to kick deep or onside based on who he puts on the field.

Further, say you do kick deep. You have timeouts and can get the ball back with a lot of time if you make a quick stop. The opposition now knows he needs a first down or two--and will be more tempted to throw, potentially stopping the clock for you on an incompletion, and opening himself up to a strip sack or pick.

Say you go for two and get it. It's still a one-score game and none of the above changes.

Now say you go for two and blow it. You need two scores...you need to onside kick and the opposition knows it. Full on hands team. If he gets it, it doesn't matter how many time outs you have--he's going to run into the middle three times, make you burn them, and punt if he has to--since you have to come down the field twice. He has more information--he knows you're in a bigger hole, just as you do, but he has more effective responses to that information.
posted by stevis23 at 2:32 PM on October 12

I'm in the "go for two early" crowd, and I'm starting to understand why I prefer the Early arguments to the Late arguments. The Early arguments are based on assumptions that contextual differences are negligible, and then pure deduction. Early is a mathematical proof. In contrast, the Late arguments are based on assumptions that contextual differences are significant. As a math nerd, I don't like the Late arguments because there is very little data to quantify the context of the situation.

For example, stevis23 raises a valid point of the decision facing the receiving team on the ensuing kickoff, but how do we quantify this? How much more likely is the kicking team to recover an onside kick when trailing by 8 than when trailing by 9? How many yards of field position will be gained through the uncertainty when kicking deep? Will any of this be enough to outweigh the potential benefit of knowing you need two additional scores? We're talking about a very small potential benefit, because the two additional scores are so unlikely, but we're also talking about very small potential benefits from kickoff tactics, or player motivation, or any of the other contextual differences.

I'd love to see some data that would quantify all the factors involved. Until then, I'll just be happy if people avoid the fallacy of cutting the lead to one score at all costs. This is like running the ball every time on 2nd and 10 just to avoid 3rd and 10.
posted by RobinFiveWords at 3:14 PM on October 12 [1 favorite]

As Napoleon said, "the reason I beat the Austrians is, they did not known the value of five minutes."

The Early arguments are based on assumptions that contextual differences are negligible, and then pure deduction.

Context matters in everything but pure math. In the NFL, the offense for the other team will play vastly differently if it knows you need two scores to catch up with only a little time left. They will run conservative plays, making a turnover far less likely, becuase they don't need a first down. If they are only one possession behind, they will play for a first down, because the four extra downs will likely eat up the remaining clock (they've got a 30 second clock on each play and at best the other team can only stop the clock 3 times). If they need a first down, they will likely get in a position where they have to throw, vastly increasing the chance for a game-changing turnover.

On defense the situation is the same. If the defending team knows the other team needs two possessions to tie, they are going to run a nickel or dime package every time--putting in extra defensive backs to stop the pass, as any run will keep the clock going, unless it (unlikely) goes out of bounds. Pass plays are easier to run with the object of getting out of bounds and are more likely to go for more yards.

Put another way, the team that is trailing wants to have the maximum number of options because then the other team must defend against all of those options, reducing its overall ability to face any one option. A team that trails by one score has far more options than a team that trails by two scores. Reducing the number of threats you must defend against makes it much easier to defend.

Seriously, for real football people, this is a no-brainer.
posted by Ironmouth at 5:24 PM on October 12 [1 favorite]

For me, it comes down to the terrible success rate of the onside kick (someone said 10%).

In addition, I bet the success rate for a two point conversion is pretty low. It has to be less than 50% per play, otherwise everbody would go for a 2pt conversion every time. For instance, if you had 4 downs to score a touchdown with a 25% chance of scoring per play, that would end up being an overall 69% chance to score a touchdown, which seems about right for the NFL average.

If you go for one, you do still have a chance for your defense to stop them, get the ball back and then score another touchdown. You go for two because you have to, and if you fail, then you fall back to the onside because you have no other option. Even if you get it with minimal time left, you still have a chance for a long field goal or a hail mary.

If you go for two and fail it, you've made it a two possession game and you are essentially forcing yourself into a position where you're going to need to kick at least one, if not two, onside kicks. (I know you could kick away, make a stop, score a touchdown, and then go for the onside as in the previous scenario, but I think you're more likely to go for an onside immediately knowing you need two scores). Also remember, if you go for an onside kick and fail it, you give your opponent much better field position so that even if they end up punting it, you have to go that much farther to score a touchdown.

TL:DR Going for one gives you the option of avoiding an onside kick unless it's absolutely necessary. Going for two and succeeding doesn't buy you much (unless you're planning on going for the win, which you're probably not), and going for it and failing makes ithe onside kick necessary.
posted by cali59 at 6:05 PM on October 12

Would anyone feel differently if the team that just scored to cut the lead to 9 is out of timeouts, and there is 1:30 left on the clock? That's the more extreme scenario in which the leading team going on offense can be ignored, because the trailing team must kick onside and will lose either way if it doesn't recover the onside kick.

@cali59, I think the two-point conversion success rate in the NFL is around 40%.

@Ironmouth, earlier today I thought that the context was negligible. You folks are convincing me otherwise. I just don't see any statistical evidence either way. The Late arguments are based on gut feelings and experiences, but not hard numbers.

I could say that the defense is likely to try much harder to prevent a last-second tying conversion than to prevent a conversion that would cut the lead to 7 points. Would this improve the offense's chance of converting early to 45%? Or that the offense won't risk its best two-point play unless it would be the tying conversion (maybe they'd use it later in the season). Would this drop it back to 40%?

All of your points are valid, but without being able to measure their impact, we don't know whether they make a 5% difference or a 0.05% difference or a 0.0005% difference in the trailing team's likelihood of tying the game. The Early argument relies on a single point. The Late argument relies on a large number of points, and most people listening to these arguments side with the Late argument, but that could be psychology. Then again, everyone knows that on fourth down, unless there are extenuating circumstances, the wise team punts. And there's a high school coach in Arkansas who never punts, ever.
posted by RobinFiveWords at 8:57 PM on October 12

The late argument as you call it does rely on logic, not gut feelings. It's logical that the receiving team will expect the onside kick and therefore field their hands team if it's a two possession game due to missing the two point conversion early. According to wikipedia the onside kick is much less effective when it's anticipated than when it's a surprise, to the tune of about 12 percent to 20 percent, though in actuality the gap between anticipated-to-surprise will be even wider since the 20 percent figure is the overall success rate of all onside kick attempts, both anticipated and surprise. And so on for all the other things already mentioned (offense will be running conservative plays.)

Having that information early benefits the defense more than it benefits you.
posted by Rhomboid at 12:27 AM on October 13

If you kick the extra point to pull to within 8, now you have to get the ball back and go for another TD, and then go for 2. Your defense has to stop their offense preferably in one series, but maybe you need a little longer, they get a first down, you can still stop them. They can't take as many chances, maybe they don't pass the ball at all (depends on whether they trust their QB). Essentially you're playing to get the ball back & win a one-score game.

If you try for 2 after that first TD and miss, it's now a two-score game no matter what. This will make it a lot easier for the other team & is going to affect how they make decisions in many ways, maybe now they can afford to take a risk and run some pass plays on 3rd down for instance, because even if something goes wrong and they give up the ball and a touchdown, they'll still get the ball back and the lead.

So choice #1 is giving your team the best chance to win.
posted by citron at 12:29 AM on October 13

Er, and if I'd actually read the wikipedia article rather than skimming it's more like 12% vs 50% for anticipated vs surprise.
posted by Rhomboid at 12:29 AM on October 13

OK, after reading, Ironmouth is right. Kind of dumb advice on my part that in the two-score situation, a team who's got the lead should take a risk to get a first down. I am actually Vinny Cerrato.
posted by citron at 12:38 AM on October 13 [1 favorite]

Would anyone feel differently if the team that just scored to cut the lead to 9 is out of timeouts, and there is 1:30 left on the clock? That's the more extreme scenario in which the leading team going on offense can be ignored, because the trailing team must kick onside and will lose either way if it doesn't recover the onside kick.

Huh? If they attempt the two-point conversion and fail, leaving the opposing team up by 9, they must then recover two onside kicks to win in the "more extreme" scenario.
posted by DevilsAdvocate at 5:31 AM on October 13

This is a deceptively tricky problem. I keep switching back between the early and late scenarios every few comments. But here's an strictly quantitative argument for Late:

1. Assume the odds of getting another possession and scoring on it, and the odds of success in a two point conversion under any circumstances, are all fixed and independent of each other and your current decision. I understand that the football strategy people would object to this, but it seems to be the premise of the question.

2. There is still the other team's intervening possession. Assume the distribution of possible outcomes is also completely independent of all the decisions/outcomes/odds above. Still, if you assign any positive probability to the most positive outcomes, i.e. a safety or a quick interception --> td that gives you a shot at a third possession, then there's at least a chance that you *won't* need to get 15 points on just two scores, so the initial premise changes. In other words, right now the odds are 99% you will need to make at least one two-point conversion not to lose, which figures in a 1% chance of one of the above outcomes. If it doesn't happen and you just get one more possession, the odds are now pretty much 100%. So if you want to equate those two situations, doesn't it make more sense to go for two when you're slightly more certain that you'll need to?

What am I missing here?
posted by pete_22 at 2:27 PM on October 13