Was BAL's Onside Kick Attempt Smart?

Trailing with 13 minutes left in the 4th quarter vs. PIT, BAL kicked a FG to make the score 13-9. BAL then attempted a surprise onside kick attempt, which was unsuccessful. What do the numbers say?

Surprise onside kick attempts are generally worth the gamble. Based on the Expected Points model, a normal kickoff is typically worth -0.42 EP, as the opponent has a 1st down at about their own 22. A recovered onside kick (1st down at one's own 45) is worth 1.77 EP. A failed onside kick (opponent's 1st down a one's own 45) is worth -2.36 EP. The break-even recovery probability would therefore be:

1.77 * x - 2.36 * (1-x) =  -.42
4.13 * x = 1.94
x = 0.46

From 2000 through 2012, surprise onside kicks have been successful 53% of the time. Since 2009, when rule changes prevent unbalanced kickoff formations, they've been successful only 33% of the time. But we shouldn't expect the rule change to affect surprise onside kicks, because an unbalanced formation would betray the kicking team's intent. So it's likely the drop is due to small sample effects, as there were only 39 surprise attempts since 2009. (It's also possible that the Saints' surprise onside kick to open the second half of the 2009 season Super Bowl may have alerted coaches to the tactic.)

As 53% is greater than 46%, in general, onside kicks are worth the risk when the kicking team has the element of surprise. The receiving team's tendencies are essential to the analysis, and the tactic is best as an option play depending on how the receiving team is aligned and whether they respect the possibility of an onside attempt.

What does Win Probability say? The game was early in the 4th quarter and BAL was trailing by 4 points, so a gamble might be more worthwhile than otherwise. A successful recovery would be worth a 0.38 WP. A failed attempt would be worth 0.22 WP. A conventional kickoff would be worth 0.29 WP. The break-even recovery probability would be:

0.38 * x + 0.22 * (1-x) = 0.29
x = 0.44

It appears the game was still somewhat in "normal football" mode at this point, as the WP and EP models were in relatively close agreement.

Ultimately, coaches would like to see tendencies on the front line of the receiving team before trying a surprise onside kick. Without that kind of analysis, it's impossible to pass a verdict on particular instances. But in general, all things being equal, surprise onside kicks are worth the risk, and this case was no exception.

Of course, all the Monday morning quarterbacking this week in Baltimore is focused on this play and not on BAL's successful 4th down conversion on the previous possession.

10 Responses to “Was BAL's Onside Kick Attempt Smart?”

1. Brian Burke says:

Botched the score on the first edition of this post. Now corrected.

2. NateTG says:

We like surprise onside kicks, and we like aggressive play when behind late. I'm not sure why the call would be controversial.

I wonder if there's a timing and score window where shortening the receiving team's field position will improve the kicking team's chance to hit a game winning drive even if they don't recover the onside kick.

3. bytebodger says:

This would probably never happen, but I'd love to see a setup where a talented kicker is given the freedom to judge, based on his first few steps of the run-up, whether the receiving team is prepared for an onside kick. If the receiving team is not prepared, the kicker purposely tops the ball and the coverage team is trained to always be on the alert for a potential onside recovery. If the receiving team is properly prepared for the onside kick, the kicker boots it deep. Aside from leading to more (successful) onside attempts, this would also improve kick return coverage, because the blockers on the kicking team wouldn't dare sell out so they can drop back into their deep blocking positions.

4. Anonymous says:

What is the best option for kicking off after a 15 yard penalty when you kickoff from the 50? Last night after the Vikings scored their TD, the Giants were flagged for 15 yard unsportmanlike penalty on the kickoff. Should the Vikings have tried an onside kick with worst case scenario Giants get the ball at the 40? It seems most teams just kick it into the endzone for touchback. Last night they purposely kicked short...

5. Jack says:

I agree, that's a great point. I've often said that teams should attempt the onside to take advantage of the not-so-disastrous field position that would result should the kick not be recovered. Otherwise, they're just booting it out the back of the end zone and essentially not taking advantage of 15 free yards. Brian, would you be able to visit that scenario?

6. Brian Burke says:

Yup. Some guys were chatting about that on the WP graph pages during Sunday's game. It would be pretty easy to make the case for it.

7. Anonymous says:

If you didn't catch it because it didn't appear in the stats, the Bears also attempted a surprise onside kick this week, also after kicking a field goal, also after trailing by 4 points in the fourth quarter. The Bears did it with 8:44 left in the game. They succeeded in the recovery, but were flagged for (a very dubious) offsides. They kicked it deep after the penalty moved them back 5 yards, with the element of surprise now lost. Kudos to Trestman and DeCamillis for the gutsy call.

8. J.D. Krull says:

"But in general, all things being equal, surprise onside kicks are worth the risk". If coaches took this admonition to heart, they would onside kick almost all the time. And then it wouldn't be a surprise anymore, eh? I guess to be rational, the frequency of onside kicks would have to reach some equilibrium where it's a little bit of a surprise, but not a total surprise.

9. Anonymous says:

Adding to JD above... I wonder how much Kickoff Return success would change if the receiving teams were more concerned with onside kicks. Meaning, would they be able to protect the returner as well as they do now, if they changed their upfront mentality - by either placing "hands" people up front or by waiting a fraction of a second longer before they commenced their blocking formations for the returner?

10. Jeff Clarke says:

Jack and 2619,