More Falcons' 4th-Down Decisions

Earlier this year, the Falcons were criticized for going for it on 4th down and failing to convert against the Saints in OT.  Mike Smith was crucified for his "incorrect" decision.  But, as Brian wrote, his decision was correct.  A decision cannot be evaluated based on the outcome, but rather the theoretical expectation of the choice itself.  This week against the Giants, Mike Smith was faced with several 4th-down decision throughout the course of the game.  The first came at the end of a 14-play, 66-yard drive at the start of the 2nd quarter.  Using our Markov model, we can look at how the drive developed:


Once they converted to a 1st-and-10 on the Giants' 33, there was a 68% chance the drive would end in either a TD or field goal.  On 4th-and-1, based on league-average coaching decisions, the drive ends in a field goal attempt 76% of the time, a turnover on downs 10% of the time and a touchdown 9% of the time.  Using Brian's nifty 4th Down Calculator, we see that teams are expected to convert 74% of the time, giving a +1.20 expected point advantage for going for it over the field goal.


On their very next drive, the Falcons were once again faced with a 4th-and-1, this time from the Giants' 42-yardline.  This is even more of a no-brainer; the break even point is a 37% conversion rate versus the league-average 74% conversion, netting +1.56 EP and +0.03 win probability over a punt.  After the failed first attempt, Smith decided to punt the ball (this time, the incorrect decision), netting only 27 yards of field position.

But, that wasn't all.  Midway through the 3rd quarter, after a 14-yard completion to Roddy White, Mike Smith was faced with yet another 4th-and-1 call, this time from the Giants' 21.  At this point, the Falcons were only down 10-2, so there wasn't a huge need for risky play-calling.  Nearly identical to the first 4th down, the break even point was a 47% conversion rate.  Going for it nets +1.19 EP and +0.04 WP over a field goal attempt.  Smith makes the right decision and attempts another QB dive up the middle; once again, Matt Ryan is stuffed.  So, is the decision bad? No.  The outcome is bad, but the decision is good based on expectation.

One thing to note is the way by which the Falcons attempted to convert.  Both times, they lined up in a goal line-type formation and tried to sneak it up the gut with Matt Ryan.  Both times, the Giants overloaded the middle, giving no ground or push to the Falcons' O-line.  The goal line formation allows the Giants to focus on a very small space to cover.  By spreading the field, Ryan would have more room to work with - especially on a potential QB sneak.  The Falcons also could have tried a more risky attempt down the field, because this forces the Giants to cover more space.  Either way, you gain an advantage over the defense by spreading them out.

Unfortunately, all that will be remembered of the 2011 season for the Falcons will be their failed 4th-down conversions, not the predominantly correct analysis by which Mike Smith came to his decisions.

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

20 Responses to “More Falcons' 4th-Down Decisions”

  1. Ian Simcox says:

    I'm always amazed at the apparent disconnect between a coach being bold enough to go for 4th and 1, but then calling a completely predictable QB sneak or a RB dive play. Why call your most predictable play on the most crucial down?

    My suspicion is that teams simply don't prepare for 4th and 1 enough in practice, so they haven't got enough plays to call on when they're needed.

  2. ubrab says:

    I knew that after the 2 failed 4th down conversions, we would hear all about how it was stupid decisions. It started a few minutes after when commenters were talking about "playoffs teams on the road have to take the points", or something like that... Agree with you, good decision, poor play call.

  3. Anonymous says:

    Ian, why would you assume that sneaks and dives are not the best plays to call when you are trying to gain one yard? Everything I have seen suggests teams pass too much on 3rd and 1 and 4th and 1, not too little.

  4. Keith Goldner says:

    A run is the more efficient play call on the short yardage situations, but it makes sense to spread the field in doing so. It makes the defense think and opens up spaces that aren't available in a goal line formation.

  5. Sunrise089 says:

    FWIW Gregg Easterbrook in his TMQ column this year has been advocating going for it (horray) on 4th but incorporating some misdirection. He calls it "do a little dance."

  6. Anonymous says:

    I think the NO coach is given too little credit. He actively won the game vs DET with agressive 4th down play. 17 points scored after not punting in a 17 point win.

  7. Anonymous says:

    Does anyone have Matt Ryan's success rate on QB sneaks on 4th down? I remember last year in the playoffs against Green Bay he blew one too and might have fumbled as well....he was darn near crying after the play. That isn't the guy I want with the ball on a short yardage play, but that's with limited information. Would be interested if he's been successful when I haven't been watching Falcons games.

  8. Keith Goldner says:

    In reference to Matt Ryan's success rate, I can't specifically look at QB sneaks, but I can look at Ryan's rushing attempts with 1-yard to go on each down. I'm excluding playoffs (although I know the one's you mentioned were all in playoffs):

    In his career, by down:

    1st - 100% (2 for 2)
    2nd - 100% (2 for 2)
    3rd - 86.7% (13 for 15)
    4th - 60% (6 for 10)

    So overall, he is at 79.3%. If you just look at 3rd & 4th downs, that falls to 76%. Not a huge sample size though (and again, it's possible some of those were not designed sneak plays, but I would guess the majority of them were).

  9. Anonymous says:

    I don't understand. He had 2 different 1st and 1 situations? We're those 1st and goal?

  10. Joaquim Baeta says:

    Ian,

    That's why the best teams teach situational football.

  11. Anonymous says:

    I really thought Atlanta did a great job on the first sneak...they made it appear that they were just trying to draw the Giants offsides. They let the playclock wind down...had all kinds of motion, then completely shifted the line and had the guard snap the ball!!

    Very clever, but again, Matt Ryan just doesn't seem apt at the sneak. He had no pause after the snap....just blindly stuck his nose in there immediately and had no chance.

    The best QB sneaks are the ones where the QB pauses to allow the line to get a push, find the soft spot to stick their nose into and then have the RB follow closely behind the QB and push the pile.

    The 2nd QB sneak was just plain dumb after failing on the first one. They went empty set giving away the fact that a sneak was coming. Whomever called that one in had a complete brain fart.

  12. chrisb says:

    Couldn't one argue that the possibility of a team mentally checking out should be included in EP fail? I know that this is tought to quantify, but three failed Go4its in a row might be tough to take. Though we are talking about professional athletes.

  13. Whispers says:

    "Using Brian's nifty 4th Down Calculator, we see that teams are expected to convert 74% of the time..."

    But Smith wasn't coaching an average team going for a 4th down play against an average team defending against a 4th down play. He was coaching the Falcons against the Giants. The Giants' D-Line is much, much better than average.

    Context matters. Let's say, for example, I present a statistic that says that 95% of men who propose marriage are answered with "yes." Does that mean that, if I propose marriage to the first woman I see today, there will be a 95% chance of a 'yes' response?

  14. Keith Goldner says:

    Whispers -

    First off, there's a huge selection bias in your example of marriage proposals. But in terms of the probabilities, you are right, conversion occurs on average 74% of the time which translates to average offenses and defenses.

    Atlanta is an above average offense (+0.07 EPA/P) and the Giants are a below average defense (+0.07 EPA/P and +0.02 Run EPA/P). Nevertheless, if you believe that the matchup would diminish the percentage greatly (even though the efficiency numbers may suggest otherwise), the break even point for making the decision is still 46%. That means that the matchup specificities would have to drop the league-average by almost 30% which is highly doubtful. Going for it was the right decision, regardless of matchup.

  15. Keith Goldner says:

    And yes, anonymous, the 1st-and-1's were 1st-and-Goal sneaks.

  16. The Wizard says:

    "On 4th-and-1, based on league-average coaching decisions, the drive ends in a field goal attempt 76% of the time, a turnover on downs 10% of the time and a touchdown 9% of the time" Keith, can you clarify here? It might be obivous to some, but are you talking about 4th and 1 from where they were at (I am not sure where that was). Also, "based on league average coaching decisions," shouldn't you subset by both going for it and kicking a field goal? i.e going for it on 4th and 1 from the 20 gets x points and kicking a field goal gets y points and the difference is x minus y.

  17. Richie says:

    One of the arguments against going for it in this situation are things like "this displays no confidence in defense" and "too much pressure on defense", etc.

    Do you have data to show how defenses do, on average, when their team's offense just failed on a 4th-down conversion in opponent's territory?

  18. Keith Goldner says:

    The Wizard -

    Those probabilities are based on 4th-and-1 from around the Giants 24 (which is where they were on that first drive). Those probabilities come from the Markov model and were not the probabilities used to calculated the expected points for decision-making. Rather, they just show league-wide trends in decision making.

    RIchie -

    I have never looked into that but it would certainly be an interesting study. I would guess that there is no affect on the performance of the defense but would have to do the work first.

  19. ecthompson says:

    The problem with Atlanta wasn't going for the 4th down, it was play selection. Mike Turner on the side lines can't be good. Matt Ryan for a QB sneak? Craziness.

  20. L says:

    @ecthompson: It's funny because after their regular season OT loss to the Saints in which Michael Turner was stuffed at the line, everyone said Mike Smith made the wrong call and should have gone with a QB sneak instead.

    Now I'm thinking maybe it has something to do with the O-line or the formations. With the empty backfield against the Giants, they weren't fooling anyone.

Leave a Reply

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