Sometimes it’s right to let them score…
Here’s the situation. The NY Giants have just tied the game up with the Eagles, and the Eagles get the ball back after a good kickoff return with about 5 minutes left in the game. They succeed in driving down to the Giants’ 32, and have 1st and 10 when the 2 minute warning hits. At this point, if the Eagles get another first down, they will simply run plays until the clock hits 0, and kick a short field goal to win it. If the Giants can hold the Eagles to very little gain on those 3 plays, and use their timeout, they can force them to kick a longer field goal, and still will have 30 seconds left even if the Eagles make it for some kind of a miracle to tie it. But, on the very next play, Brian Westbrook breaks through for a 13-yard gain, and the Giants make the serious mistake of tackling him short of the end zone! Let us see why this is a mistake.
Consider the Giants chances of winning the game at the point where Westbrook is running towards the end zone and has the first down. At this point, even if the Giants use their one timeout, the Eagles can still run 3 plays and kick a ~35-yard field goal to win (assuming they gain almost no more yards). This will only go wrong if they fumble during those 3 plays, or if they miss the field goal. In his career, Brian Westbrook has lost 7 fumbles out of 736 carries, or a little less than 1% of his carries. But, since holding onto the ball, not gaining yards, is the priority here, this is certainly an overestimation of the likelihood of a fumble. I make the true odds somewhere about half that, or 0.5%. The Eagles have to run 3 plays, so the odds of a fumble are 1 – (.995)3 ≈ 1.5% (actually slightly less). If they get through that ok, they have to kick a field goal successfully. In his career, David Akers is 58/66 on field goals of 30-39 yards, or 88%. Maybe this is slightly less, since it was raining/snowing, but Akers is a veteran kicker, and we’re not taking into account the yards they might gain on the three running plays. We’ll call it 82% to be generous. The odds that something bad happens, therefore, is 1 – (.82)(.985) = 19.2%. If they miss, the game will go into overtime, giving them, at least in principle, a 50% chance to win. So their total odds of winning are .5 * 19.2% = 9.6%. Not good, but they can be improved.
Let us now say that the Giants’ defensive coordinator tells his players that if anyone on the Eagles crosses that first down line, they simply miss the tackle or slip and let him through. Sure, this results in a touchdown for the other team, and your team being down 7. But, you now have the ball and 1:55 or so left in the game and a timeout. Overall, the Giants scored a touchdown on 20.9% of their drives this season. Sure, they had more time to do it, but they have a number of factors working in their favor now.
1) They are playing 4-down offense.
2) It’s the end of the game and defenses get tired at the end of the game.
3) They know not to kick a field goal – so many of the drives they kicked field goals on would instead be continued to touchdowns.
4) Teams frequently sit back in a prevent defense in this situation making it easy to get down the field quickly.
The Giants punted on 41.2% of their drives and had a turnover on 13.9% of their drives, which means that 100 - 41.2 – 20.9 – 13.9 = 24% of drives they attempted a field goal. Here, they would go for it on 4th down (with, say, a 35% chance of success, then a 40% chance of scoring after that), and this would have a .35 * .4 = 14% chance of success. This means a successful touchdown 20.9% + 24% * 14% = 24.2% of the time. Actually, I think given the other favorable factors, the real percentage is actually much bigger than this. I haven't even considered the 40% of the time they punted as counting towards their chances to score. All this would just get us into overtime, so again, this percentage would be halved, translating to a 12.1% chance to win. Granted, this is not much better than the other way, but it is better, and that’s all that matters, and remember I was erring on the side of caution.
Teams are constantly throwing away chances like this simply because coaches are unwilling to make a risky-looking, yet percentage play. The media and fans expect their coaches to always make a choice that works out well. Clearly, this is unreasonable – you want your team’s coach to always make the decision that has the highest probability of winning the game you are playing (at least that’s what I want). Now, maybe you’ll argue that I’m overestimating the chances of scoring a touchdown, but clearly the odds could be much better in favor of this move. Replace the Giants’ offense with, say, the Colts’ offense, and now you have arguably the best high-pressure performing quarterback and offense in the league working to score. Now, clearly the odds are highly in favor of voluntarily giving up the touchdown in exchange for the ball back, rather than impotently waiting and hoping that they don’t kick the field goal.
This sort of logic applies also in part to a game where neither defense can stop the other team’s offense. Let’s say that the game is tied and the opponents have the ball with about 4 or 5 minutes left. Doesn’t it make sense to play ridiculously aggressively on defense, blitzing 7 or 8 men each play. This type of defense leads to one of two results. Either you stop them because you get a few sacks and an incompletion, and they have to punt it to you, or they break open a big play and score – then you get the ball back with time to mount your own drive and tie it to go to overtime. Either way, this is better than letting them take the whole rest of the game to drive down the field and score to go ahead with little or no time left.
Now, I’ve neglected the fact that you opponents could (and should) NOT accept the free score. What is good for us cannot be good for them. Even if it’s clear that you are letting them score, most players have no idea about good strategy, and will just run into the end zone out of pure reflex. Because they are never faced with this situation, there is basically no chance that they get it right. Now maybe if you have a track record of doing this, they’ll catch on. But, situations like this are so rare, that if your players act like they’re trying to tackle the guy reasonably well, it’s possible nobody will realize what you’re doing. Even if they do figure it out, you haven’t lost anything, so it’s basically a no-lose play.