Limiting possessions key to victory?

Chase, from the comments:

I agree that blitzing is not necessarily a good underdog strategy, because limiting possessions seems to be the biggest underdog key. People talk about controlling the clock, but that doesn’t make any sense in a vacuum. When Miami held the ball for over three times as long as the Colts in that Monday Night game last year, people talked about how brilliant it was to keep Manning off the field for 45 minutes.

But the Colts and Dolphins had the same number of possessions in the game, so who cares? The point isn’t to hold the clock, the point is to minimize variance. That’s the real advantage of controlling TOP, but giving up a ton of big plays on defense and having a methodical offense won’t help you win games no matter how great your TOP is.

So what can an underdog do?

(1) There is one real way to win the all-important possessions battle: control the ball at the end of each half. Combined with other possessions-minimizing techniques, you could end up with 9 possessions to your opponent’s 8 possessions, which is a legitimately valuable edge. If you get the ball with 8 minutes left, it probably makes sense to start thinking about a 2-for-1 with possessions. If you get it with 5 minutes left, figure out if you should go 2-for-1 or if you can drain all 5 minutes. With 3 minutes left, you have to ensure that you have the ball last. Do that in both halves, and you’ve stolen a possession (ideally, scoring a TD with as close to triple zeroes as possible).

(2) Going for it on 4th down is another obvious underdog strategy. In addition to it being a legitimate favorite strategy — going for it on 4th down is the correct play far more often than conventional wisdom dictates, and the correct player is almost always a good favorite strategy — it helps increase variance.

(3) Kicking field goals is almost certainly a loser. Going for it on 4th and G from your opponent’s 10 may not sound like a great idea, but even if you only gain 5 yards, odds are you will prevent the other team from scoring. The more times you can force your opponent to start drives inside their own 10, the better, because research shows that teams are overly conservative in that area. Only in blatantly obvious FG situations should an underdog kick — punting and trying to pin inside the 5 is also a good strategy.

(4) On defense, I think bend but don’t break is the correct strategy. If you can force the opponent to chew up clock and kick a FG, that’s a big win. Chewing up clock conquers all, I think. Once again, the goal should be 9 possessions to 8. Although obviously TOs would be very nice.

(5) On offense, chewing up clock is good but scoring touchdowns is better. I think whatever play is TD-maximizing, whether it’s going for it on 4th down, being run heavy, being pass heavy, being trick-play heavy, whatever, is the goal. A flea-flicker that goes for a 60-yard TD might turn it into a 10-to-9 possessions game, but who cares if you score a TD on that possession?

(6) On offense, I think a modified no-huddle offense following plays where the clock is running is the key. Following a run or completion that lands in bounds (or out of bounds before the clock stops in the final 5 mins), the offense should immediately run up to the LOS to prevent the defense from substituting. Then, they should simply milk the clock for the full 40 seconds (with some variance so defenders can’t time the snap) by doing whatever. Actually calling the play, wasting time, twiddling their thumbs, it doesn’t matter. But preventing defensive substitutions would seem to be a strong underdog strategy.

(7) Special teams would be the overlooked key here. Winning the field position battle, the hidden yardage in football, is an easy way to level the playing field.

My only question: Is it more important for an underdog to limit the overall number of possessions or the relative number of possessions (or both?). This analysis seems to indicate that there are too few onside kicks. For more, see also here and here and here.

But, see this surprising result:

… The team receiving the ball [at the beginning of the first half] consistently lost the half (except in 2008)…. The receiving team will have as many possessions if not one more than the kicking team [during that half]. Yet the data clearly show that the kicking team has won more game halves than the receiving team….

I next ran the data to see how teams receiving the ball at the start of the second half succeeded. The data show the game results of the team receiving the ball in the second half. Again, the team kicking to open the second half won more games than the receiving team.

  • Minimizing overall possessions is still a decent underdog strategy — if you think your opponent is going to average more points per possession than you, then you want to minimize the possessions and keep the game close enough that one big/weird play or two can make a difference. More possessions means more big/weird plays might be needed. If you think you can just straight-up beat your opponent, then this really doesn’t matter. But if your goal is to, for lack of a better term, ‘steal’ a victory, then minimizing possessions is a great idea.

  • It seems to me that increasing variance is the best way for a big underdog to steal a win. Run a lot of man, blitz a lot, take a lot of shots, along with your suggestions. (TD v FG, etc.) If you accept that losing 49-3 is no different from losing 27-21, jacking up the variance is the best way to steal a couple of scores through turnovers and other big splash plays.

  • Chris: Just a thought — I understand the frequency and depth with which the concept has been covered here, BUT — you might want to slap a hyperlink on the first occurence of “variance” in order to help out newbies…

  • Bah. I should have read the original article before commenting.

    Never mind. Old ground.

  • Paul

    One thing about the comment regarding how scoring a TD beats chewing clock, I believe this ignores the effect that resting time can have on defensive players. If you score a TD after a long drive, you may be more likely to prevent the opposition from scoring on their next possession.

  • Will

    Paul: true, but that matters only in real time, not game time. The two may or may not be comparable, depending on whether or not you huddle on offense, how much you throw incomplete, etc.

  • Charles

    Surely there’s an alternative here, that you force your opponent to increase their variance with your defence, and you play a more conservative game on offense? It seems like it should be obvious that the expected outcome for each team is related, and I would predict that the reason most upsets don’t end up as close as one would expect from more stay in the game strategies is that it changes the standard deviation of the ‘goliaths’ score as they increase their risk through panic.

  • mitch v.


    not sure i’m following everything you’re saying…

    “But the Colts and Dolphins had the same number of possessions in the game, so who cares?” I think an underdog minimizing the combined possession of both teams is certainly a legitimate strategy. Lofty TOP can certainly be indicative of an offense trying to minimize total possession – by making their possessions take longer, the game is in effect, shorter. Which, as has been said, increases variance in that a single flukey/trick play has a greater weight in determining the final outcome. More can be said about defensive strategies to force opponents into lengthy drives when the score is close (obviously not preferable to 3&outs or turnovers, but hey we’re underdogs here), as others have mentioned “bend don’t break”. If an underdog we’re doing both, TOP may not be an important stat. But, in the Miami Indy example, TOP meant Miami was limiting total possessions of both teams, increasing variance and their chance to win, not just by preventing defense subs, opponent’s ‘rhythm’, etc.
    Interesting here is that most of the talk about variance leads to a ‘more of a chance to win’ vs. ‘more of a chance to get blown out’, which doesn’t seem to be the case with limiting possession. If an underdog can force a game where each team gets two possessions (which no one can), they would be increasing their chance to win and could only ever lose by 14. What coach is going to get fired for that?

    “Is it more important for an underdog to limit the overall number of possessions or the relative number of possessions (or both?).”

    Seems that both is the answer. But, to be pedantic, limiting relative possessions a la Chase’s (1) shouldn’t be thought of as an underdog strategy – as it doesn’t increase variance so to say, and is an equally legitimate strategy for the favorite.

    Just my quick thoughts

  • Chase

    Is the only difference between overall number of possessions and relative possessions the idea of onside kicks? If so, I certainly agree with Burke’s findings that onside kicks are an effective underdog strategy. That said, I don’t consider it much of a long-term strategy for underdogs. Surprise onside kicks can only occur once or twice before they cease being surprises. So in one game, yes, underdogs can do this, but at some point soon that strategy will melt away (just like you can’t run flea flickers every quarter).

    As for your surprising result, well, yes that is very surprising. But absent a legitimate explanation, I’m not inclined to believe it. It goes against just about every bit of football research ever done to suggest that having the ball at your 30 is worse than your opponent having the ball at his 30. I also think we’d want to see more than just a record, i.e., a points differential. Are teams that win the coin toss outscoring their opponents just 45% of the time in the first half but on average outscoring them by 2.2 points? That would be relevant information.


    When people speak about controlling TOP, they don’t speak of variance. I certainly think underdogs should try to chew the clock on offense, but they should also try to chew the clock on defense. If your opponent has a 10-minute, touchdown drive, I’m not sure that’s all that bad for an underdog. One lucky play and you’ve turned this into a 48-minute game instead of a 60-minute game. Generally, I think underdogs should hope for long touchdown drives for themselves and, at a minimum, no quick TD drives for the opponent.

    The flip side to this is that dominant teams should *not* be focused on winning the TOP battle. But commentators never mention that (of course, winning the TOP battle because you force a ton of 3-and-outs is still advised ;)).

  • Charles

    I don’t quite see how limiting possessions is hugely advantageous. If we’re talking about increasing opportunity to win. If there are 20 possessions split evenly the stronger team has a higher probability of scoring on each of it’s ten drives then the weaker team. To increase the probability of winning the weaker team needs to increase it’s chance to score, decreasing the number of possessions only increases the chance of winning when a lead is achieved, and that assumes that the weaker team can hold onto that lead.

  • Brian

    One thing overlooked in the Colts-Dolphins example is how many drives Indianapolis had in the game and how many they AVERAGED for a normal game. If both teams have equal possessions but only have, say 10 each, and the Colts normally have 13, then they gave the Colts three fewer possessions, and the strategy would have worked. Chewing the clock but giving them the same amount of possessions they expect to have in a normal game doesn’t really provide a benefit.

  • mitch v.

    ok Charles, hang with me for a second

    pretend a Team A (underdog) is playing Team B (favorite)

    Team A scores a touchdown on 25% of their possessions and
    Team B scores a touchdown on 50% of their possessions

    think flipping coins or the like, the fewer trials there are the greater possibility Team A scores more than Team B. you could certainly see Team A winning or tied after each team has had one possession, and not because they did anything to increase their 25%. but, it becomes increasingly improbable as the number of possessions increases.

    Team A, whether winning 7-0 or losing 7-0, has a better chance of outscoring, and ultimately not losing to, Team B in 2 possessions than in 20.

    Still, though, I’m not disagreeing when you say it’s not “hugely advantageous”. They numbers I used are probably a wider spread than say the best NFl team to the worst, but something worth considering still.

  • Old South


    It’s because of variance. The effects of variance are most profound over small samples.

    Think of it this way: you’re in a 3 point shooting contest with Michael Jordan. Do you have a better chance at beating him if it’s best of 3, or best of 300? The obvious answer is best of 3; you might just get lucky and hit all 3 when he misses one. Over the long run, though, his superior shooting will win out.

    Same concept applies with possessions in football.

  • Charles

    If I’m trying to maximize my possibility of winning, then I want greater ability to control the outcome. If each team has one possession, once the other team scores I cannot win, with two possessions even if they score on their first possession I still have the opportunity to win. Also smaller number of possessions lowers the amount of risk that I can take because I won’t have as many opportunities to recover, and that again lowers my ability to win. Not only that but of course teams with a lead will choose to play differently from teams who are behind. Underdogs want to increase the variance of the favoured teams expected points, so that the probability that they score less increases, the best way of doing that is to make them chase a lead. The less possessions the more a ‘goliath’ can play the brand of football they want to play because it’s unlikely they’ll feel as much pressure.

  • Dubber

    I’m most troubled by the kickoff thing…….

    I have to know WHY?

  • Dennis

    I’ve long believed that TOP is an overrated stat, largely because of the aforementioned issue that each team will still get the same number of possessions, +/-1, with the exceptions of scores and/or turnovers on special teams. (And onside kicks, which was why I so approved of the Saints doing so in the Super Bowl.) What matters is what a given team does on each of those possessions. I recognize that giving the defense a rest is a legit issue, but they’d also get some rest if they made a 3rd down stop.
    I haven’t had the time yet to do the research, but I suspect we (fans, media, pundits, and coaches) mis-identify what causes the TOP edge. Is it really how long it takes for a given play to run – using every second on the playclock? I believe, though without evidence yet, that converting 1st downs (and preventing them) has a much larger effect on TOP than how long the average play takes to get started. I thought of this back during the Bills 4-yr Super Bowl run; after they lost to the Giants there was such a big deal made about TOP, and the Bills responded by trying to milk the clock on each play in the subsequent seasons. And their offense regressed from the previous heights. There was more than just their per-play clock average, but I believed then it was a factor and still do. The team that does a better job converting 1st downs and preventing them is likely going to win the game, and that success leads to an edge in TOP. Then we give credit to winning the TOP battle as one of the key factors. It seems to me that this is like the record when RB-x runs the ball 30 times stat.

  • RP

    this is an intresting post that fits into “limiting possessions”. Do you know of the rule change that Ryan talks about, and in your opinion would this be a legitimate tactic for a scenario such as the Jackson return in the giants/eagles game.

  • RP
  • 4.0 Point Stance

    Chris, I think the “kick off to start both halves” strategy deserves a post in itself. What could realistically be an explanation for this? Noise (seems like there’s too many data points for that)? Field position? The fact that an offense needs a drive or two to get the kinks out?