To Fake or Not To Fake: Fourth Down Decisions

After we wrote about Mike Tomlin's fake field goal decision in Week 9, it got me thinking about the general success rate of fakes versus typical fourth down attempts. Before moving forward, it should be noted that it is difficult to pull fake field goal and fake punts out of play-by-play data because they are not explicitly specified as such. So, you may be getting an aborted or bad snap or the field goal/punt formation may not have been recorded. We attempted to clean the data up as much as possible, which resulted in 213 fourth down fakes since 2000.

Similarly, we looked at the roughly 5500 fourth down normal go-for-it attempts. Granted, a lot of these will take place at the end of the game when teams are down by significant margins or must go for it in order to stay in the game, but that should not greatly affect their conversion rates. It may give us a sample bias, though, because those teams that are losing are typically worse offensively.

Fake attempts are exceedingly rare and given that there are only 200 or so attempts, we must be cautious of sample size issues. The most frequent fake attempts are on 4th-and-1 or 4th-and-2 which have happened about 30 times a piece. Compare that to over 2000 regular go-for-it attempts on 4th-and-1 and over 600 regular attempts on 4th-and-2.


To assess the value of a fake attempt I looked at success rates based on distance-to-go and then used logistic regression to smooth the data. The results:


The blue represents actual fake field goal and punt success rates while the green represents actual go-for-it success rates. The red smoothed line is an estimate of fake success rates while the purple is an estimate of regular fourth down attempts. It is pretty noticeable that fakes convert at a higher rate, especially with 6 or less yards-to-go. That difference gradually reduces as distance-to-go increases.

The main value of a fake is the surprise nature. Opposing defenses (or special teams) are not expecting a team to go for it and as a result, they are caught off guard without proper defenses. One huge thing to note is that fakes are typically employed when the coaching staff notices a trend for not defending against the fake in their opponent. That is, a coach will notice that there is no safety net on a punt whenever a team defends a punt (similar to when a kick return team does not guard against the onside kick). That means there is a slight sample bias in the data due to pre-existing conditions.

All that said, the choice to fake or not to fake comes down to game theory. There should be an equilibrium when it comes to the success of 4th-down attempts and currently, there is not. As a result, it makes sense for teams to actually fake more often than they currently do until the success rates become equal. Again, keep in mind that the small sample sizes could mean that the difference is due to random variation, but it is definitely a starting point when it comes to strategic decision-making.

Keith Goldner is the creator of Drive-By Football, and Chief Analyst at numberFire.com - The leading fantasy sports analytics platform.  Follow him on twitter @drivebyfootball or check out numberFire on Facebook

  • Spread The Love
  • Digg This Post
  • Tweet This Post
  • Stumble This Post
  • Submit This Post To Delicious
  • Submit This Post To Reddit
  • Submit This Post To Mixx

11 Responses to “To Fake or Not To Fake: Fourth Down Decisions”

  1. Brian says:

    I'm guessing that the logistic regressions you fit are not significantly different in their parameter estimates, are they? Sample size is too small to say that fakes are more successful.

  2. Jeff Clarke says:

    I really like the fake punt with a direct snap to the upback. If you look at a punt alignment, you'll see that the punting team's gunners are both double covered. Those four men along with the punt returner are effectively removed from the play. The punter and the gunners are removed from the play.

    This sets up a 8-6 personnel mismatch for the punting team. If I was a coach, I'd probably run a play from punt alignment nearly every time I wanted to run on 4th and short. The element of surprise would be lessened but I think the numbers benefit would more than make up for it.

  3. Anonymous says:

    > If I was a coach, I'd probably run a play from punt alignment
    > nearly every time I wanted to run on 4th and short. The element of > surprise would be lessened but I think the numbers benefit would
    > more than make up for it.

    That might translate to better net yardage on punts if the other team changes formation to compensate.

  4. bytebodger says:

    >>I'd probably run a play from punt alignment nearly every time I wanted to run on 4th and short. The element of surprise would be lessened but I think the numbers benefit would more than make up for it.

    The problem here comes down to *personnel*. You can line up in any punt-centric formation that you want, but I see your QB and your starting RB in the backfield, I'm going to be screaming to the rest of my defense that you're going for the 1st down.

    Conversely, you can "fix" that problem by only leaving your punt team in for the conversion, but if the guys on your punt team were the best options to be gaining the 1st down, they wouldn't be on the punt team - they'd be in your starting offense.

    If I need/expect to make a first down then I usually want the personnel in the game that give me the best chance of conversion. If I put my punt team out there for ALL 4th-down plays then I seriously damage my chances.

  5. Anonymous says:

    Jeff, samples sizes are 220 and 5500. I'm thinking parameter estimates are going to be quite tight.

  6. Keith Goldner says:

    Jeff,

    There are big enough samples and deviation to consider the fake success probability significant. If we include a dummy variable and run a logistic regression against all fourth down attempts, the p-value for the dummy variable (representing fakes) is 0.013 (significant at the 95% level).

  7. Magnus Alvestad says:

    What about Miami's decision yesterday to kick a field goal on 4th and 5 when they were down 21-0?

  8. Keith Goldner says:

    Magnus -

    They forfeited about 3% win probability with that decision. 53% conversion rate on 4th-and-4, expected win probability going for it is 11%, kicking the field goal is 8%. A made field goal there = 9% win probability, just 2% higher than a failed attempt at going for it.

  9. Anonymous says:

    Actually, the equilibrium has nothing to do with how often a team converts on a regular go for it attempt...The situation should occur, in theory (though since teams punt too often, not in practice), when "EV punt" > "EV go for it" to begin with. Otherwise they wouldn't even be lining up in the punt formation...they'd just go for it. The important equilibrium for a fake punt, then, is where the EV for the fake is equal to the EV of the punt.

    Am I missing something here?

  10. Tim says:

    "The problem here comes down to *personnel*. You can line up in any punt-centric formation that you want, but I see your QB and your starting RB in the backfield, I'm going to be screaming to the rest of my defense that you're going for the 1st down. "

    TEBOOOWWWWW!!!!

  11. Tim says:

    "The problem here comes down to *personnel*. You can line up in any punt-centric formation that you want, but I see your QB and your starting RB in the backfield, I'm going to be screaming to the rest of my defense that you're going for the 1st down. "

    TEBOOOWWWWW!!!!

Leave a Reply

Note: Only a member of this blog may post a comment.