4th down
Ball on opponents 2 yard line
Down by 10
Under 3 minutes left in the game
Also, assume you'll probably need to get the onside kick because your defense hasn't forced many three and outs
The question is, do you go for a touchdown, knowing that you then only have to get a field goal? Or, do you kick the field goal, knowing that you need points on 2 possesions and that you are much more likely to get the field goal than the touchdown? There are some strong opinions around here about this. What do you think?
10 comments:
Kick it! You need a FG anyway, so take the sure points while you can. USC made the wrong choice.
I'm sure someone on the internet has already posted all the necessary calculations to give the "right" answer, but here's my opinion (after a few caveats, of course).
First, I'll say that I didn't watch any football games this weekend, so I missed any game this might hypothetically be referring to. Second, I don't think you've provided enough information. 2:59 vs 2:03 left is a big difference, as the former leaves you with at least some sort of chance should you force a 3 and out. Also, what about timeouts? If you've got 3, you give yourself all sorts of opportunities, as they're worth about 2 more minutes of time and making moving the ball much easier when running or passing up the middle remains more viable.
So, my answer: I'd kick the FG. (Provided you have a better kicker than BYU.) If you go for it on 4th and miss, the game is effectively over. Your objective should be to force the other team to keep playing for as long as possible, giving them extra chances to mess up. Going for it on 4th down turns the game into a coinflip, after which you still need some crazy luck (recover onside kick) in order to win. (Expect onside kicks only have about a 10% recovery rate.) I'd like to hear your thoughts on it, and perhaps I'll expound more sometime when I don't have kids to look after.
4096 characters per post? Seriously, Google?
This will be long since there's still a lot of football left to be played at this point and you left out a lot of details. You do realize that where the ball is on the field is going to affect the odds, as will the ambient air temperature and any winds.
Anyway, let's cover the scenarios where you win down ten first. For simplification we'll call overtime as good as a win and neglect the discussion of going for two. I will also neglect the possibility of a safety and giving up 1 or 2 points on the PAT attempt. (These would turn out not to matter much anyway)
W1) Score a touchdown, score a subsequent touchdown or field goal, successfully prevent an opponent score. [same fuzziness that plus only requires you to prevent a touchdown]
W2) Score a field goal, score a subsequent touchdown, prevent an opponent score.
W3) Fail to score a touchdown, get the ball back, score a touchdown, and a field goal in either order (or more likely two touchdowns)
W4) Fail to score a field goal, get the ball back, score a touchdown and a field goal in either order. (or 2 TDs)
In W3 and W4, you'll not have enough time left to worry about the opponent scoring again anyway.
Note that in every scenario a failed onside kick is basically a loss (and extremely dominant as a term) so we'll assume the odds of onside recovery are 100%. (Not to mention it's a constant and whatever value I pick it will cancel out).
So the question is which is more likely W1/W3 or W2/W4. From the 2, we're basically looking at an extra point. You don't mention if the ball is on the hash or not, so let's take a minor penalty over the usual odds. For convenience we'll say 47% chance of a touchdown and 94% chance of a field goal.
So attempting a touchdown:
47% TD Ball at the 50
53% Opponent ball at the 2.
Attempting a field goal:
94% FG Ball at the 50
6% Opponent Ball at the 20.
Giving the opponent the ball at the 2 with 3 minutes remaining and a safety making it a one score game, you're likely to see three runs somewhere between the inside of the tackle and the outside of the center. So depending on the number of timeouts left, you're going to see between 2:30 and 1:00 minute left on the clock when you get the ball at about the 45 with what I will say is a 50% probability. (I think it's [much] higher than this but we'll yield to your claim that the defense isn't playing well.) To simplify the math, I'm going to move the ball back to the 50 here.
Giving the opponent the ball at the 20 with 3 minutes remaining you're much more likely to give up a first down with more of the playbook open. Using the same deference to the defense not playing well I'm going to give it a 75% chance of losing the game. That's 25% chance of getting the ball back at the 25. But you're down two scores with the full field. At 2:30 it's not over but at 1 Minute it is.
So now we're at:
TD play
A) 47% TD
B) 26.5% No Points (1min to 2:30 on the clock depending on timeouts)
26.5% Lose
FG Attempt
C) 94% FG
D) 1.5% No Points / 1.5% Lose (if you have no timeouts)
4.5% Lose
Let's cover D first though it won't make a lot of difference. If you're having a monster day on offense perhaps you have a 33% chance of scoring twice from the twenty five in 150 seconds with no timeouts. (But then if you're having a monster day on offense, how did you end up down so far so late?) So let's say that there's a 1 in 200 chance you can win the game by missing the field goal and playing great defense. It's actually pretty generous.
Now in C we're looking at the odds of getting a touchdown from the 50 using all 4 downs. Let's call those X for a moment.
.94X + .005 = Field Goal Winning percentage
In A we're simply needing a field goal attempt to win but we'd be happy to win with a touchdown. Let's call the odds of MAKING a field goal Y and the odds of MAKING a touchdown Z. (Note Z != X because we only use three downs for part of the field).
.47(Y+Z) - .005 (which will be the other teams odds of scoring based on what we have above and assuming great clock management)
+
.265(X)*(Y+Z)'
== Touchdown winning percentage
Where (Y+Z)' is smaller than Y + Z because time actually matters here. [150 seconds is plenty of time to go half the field, not so much twice]
Now, you can get any answer you want by plugging in different values of X, Y, Z, Y' and Z'. For example, if you let X = 40%, Y = 35%, Z = 33%, Y' = 33% and Z' = 20.5% then you can do whatever because the odds of winning are 37.1% either way. But perhaps you like some other values. For example, perhaps you consider the touchdown from the 50 unlikely [.25, .33, .15, .1, .1] yields a .00385 advantage to going for the touchdown. Another interesting vector is [.5, .65, .35, 0, 0] which yields a tie at 46.5%. That's interesting because it suggests a sure thing points on the second drive and a 50% shot of scoring the touchdown from the 50. But allows no opportunity of getting the ball back and scoring. But the zeros are certainly two low if we believe that we can score from the 50 in 4 downs in 150 than surely that's not zero.
You can of course supply your own values to try and make the TD wrong, but basically for most reasonable values it's either quite close or in the TD's favor. The reason being the same reason you should NEVER kick this field goal if game situation isn't playing into the discussion. A team who turns the ball over inside their five is more likely to score next then their opponent.
Now the more interesting discussion is who this makes right. Initially, my take was that Sabrina probabloy opted for the field goal as a throwback to her days of loyalty to Utah (5-7). They have a history of taking points off the board by kicking which afflicts many schools with good kickers like Louie Sakoda.
Alas, I've decided to place my money instead on Brett being the one who is wrong here out of an abundance of preferring to get the quick points.
Inquiring minds want to know.
[By the way, if Alisquire is right and you did mean the ND USC game, that was a nine point win which adds another scenario to win with (3 FGs) I doubt that makes a difference since it means I can longer neglect the onside kick recovery term which is probably a pretty small probability especially since the two recoveries are not independent probabilities.]
Either way, while the discussion is interesting, you're still losing the game. Ben puts the odds of winning in the neighborhood of 40%, including the onside kick recovery as a free bonus. Expected onside kicks have a recovery rate of about 18% according to the internet (about 60% for "surprise" ones), which moves your odds of winning the game firmly into the single digits. If the FG-then-TD plan has a 35% chance of success and the TD-then-FG plan has a 45% chance of sucess (seems like a big difference, right?) then the required onside kick drags those down to 6.3% and 8.1% respectively. My point is that while, yes, there is a correct answer to the question that was posed, the first 57 minutes have all but sealed your fate - you should have not sucked earlier in the game, because now even the right choice only raises your odds of winning by 1 or 2 percentage points.
advancednflstats.com puts the conversion rate for 4th and 2 from inside the 10 at about 55%. 0.55*7pts=3.85pts, which means you should take the points now, as you've already "earned" more than 3.
outsidethehashes.com uses college data and shows essentially the same thing. They don't seem to have any 4th down data, but value 3rd and 2 from the 2 at about 4.5 points. 4th down would certainly be worth less, but I'd expect it to still stay over 3 points.
The ability of your kicker is another important factor here. From the 35 yard line, a hail mary might be 5% successful. Anyone want to suggest odds on BYU hitting a 52 yard FG? At 40 yards, the hail mary won't have changed much, but the FG odds will be fading fast, and the overall game calculus will need to be adjusted for this. If you've got a great kicker, go for the TD, if you recover the kick, you only need 10 or 15 yards and you can attempt a FG. If your kicker stinks, you might need 30 or 40 (heck, I know a team that has trouble on PATs).
http://www.advancednflstats.com/2009/09/4th-down-study-part-3.html
http://outsidethehashes.com/?p=199
http://www.advancednflstats.com/2012/09/hail-mary-probabilities.html
I certainly didn't run any numbers and am not really into stats. I agreed with the decision to go for the touchdown, but a lot of that had to with how close they were to the end zone, as Ben points out is an important piece of info to consider. They were only a couple of yards from goal line, if I remember correctly. It's hard to ever even get that close and it's much easier to get a touchdown at the close range. My reasoning was they might not get the chance to get that close again and they need a touchdown to win no matter what. They might as well take their best shot at a touchdown while they have it and then get as far down the field when they have possession again for a field goal. If they end up not getting as close, they have a better shot at getting a field goal from further back. So basically, I feel like you might as well take your best shot at the touchdown while you have it. Sure, they didn't get the points and the game was effectively over so I see why people are saying it was a bad call, but I think I get why they called it the way they did and I would have done it the same way.
In reviewing the post and looking for a response from Brett (which is apparently something I'll have to learn to live without). I not that he didn't even specify that it was 4th and goal. So, I'm going to assume that it's 4th and inches and then going for it is a no-brainer.
Post a Comment