Tuesday, October 16, 2012

Rant of the Week: Onside Kicking From Opponent's Side of the Field

Because there are small things in just about every game that make me angry (and a lot of big things, too), I've decided to start a weekly column where I complain about something. Sometimes, like today, it will be a coaching decision. Other times it will be on refereeing, or clock management or whatever is bugging me that week. This week, it was Jim Schwartz's decision to not onside kick the ball from the Eagles' 45-yard line (or even consider it).

Let me set up the situation in case you forgot or didn't see. The Lions had just scored a touchdown to get within three of the Eagles, 16-13. On the ensuing extra point, the Eagles encroached, costing five yards on the kickoff. On the ensuing, ensuing extra point, Fletcher Cox punched a guy, costing the Eagles another 15 yards on the kickoff.

So the Lions, down three with 10:30 to go, were set to kickoff from the Eagles' 45-yard line. Jason Hanson decided to take the opportunity to blast the ball through the endzone, resulting in a touchback (or a net kickoff of 25 yards).

This is an all-too-common play. In fact, I can't remember a single time a team has elected to onside kick in this situation. But the onside kick was the right play, and a kick through the endzone was a wasted opportunity.

But I'm an understanding person. Please, convince me, Schwartz. Why not onside kick the ball? (quotes via Pride Of Detroit)
"We gave up four special teams scores in two weeks. We weren't going to get cute right there."
Hmm...okay. You don't want to onside kick because its cute? So you don't want a chance to capitalize on a huge opportunity because someone watching may look and say, "Oh, Jim Schwartz, you cutie!" There must be a real reason...
"They have one of the best returners, a pro bowl returner in WR DeSean Jackson." 
Uhh...what? We're talking about an onside kick. You do know what an onside kick is, right? The deep kick returner doesn't usually get his hands on an onside kick. And if the receiving team does recover an onside kick, you know how long they return it 99% of the time? ZERO YARDS. And if you're that worried that Jackson will return an onside kick for a touchdown, KICK TO SIDE OF THE FIELD WHERE HE IS NOT STANDING. I would say Jackson's chances of returning a onside kick for a touchdown go from about .01% to .00000000000000000000000000001% if you KICK IT TO SOMEONE ELSE. I cannot believe the media let him get away with this terrible explanation.

What makes everything so much worse was his answer to an earlier question about challenging a play that clearly was not going to get overturned:
"And I said ‘If it's close I'm going to throw it,' because [...] the reward is so great, I mean, if we're able to get a turnover. Imagine this, imagine if it had been the other way and it had been slightly backwards and we didn't get a replay and we didn't challenge it and we were sick to our stomachs after the game saying ‘Jeez, we could have gotten a turnover in the red zone, taken points off the board and everything else.'"
Now, this mistake isn't all on Schwartz. He challenged the play approximately 30 seconds after the play was done, and if the guys upstairs couldn't find a conclusive replay (which everyone at home had already seen) by then, then there is seriously something wrong with whoever is watching tape up there.

So Schwartz's rationale is that despite the low chance of an overturned review, the potential huge reward (opponent turnover in the red zone) is worth the minor risk (lost timeout, challenge). It's a fair assessment. A successful review takes at least three points off the board, and the Lions still had plenty of time to come back, so losing a timeout isn't particularly damning. Obviously it looked like a bad decision when we all knew the play wasn't going to be overturned. But if Schwartz really thought there was a chance the play would be called his way, he was correct in his actions.

But you could use this EXACT LINE OF LOGIC to kick the onside kick. There is a small chance that the Lions recover an onside kick (especially considering the Eagles were expecting one). Football Outsiders puts that chance at around 18%. But lets look at risk versus rewards, since that seems to dictate Schwartz's decision.

The reward of a successful onside kick (which, may I remind you, the Lions successfully did in Tennessee, when the Titans were expecting one), is absolutely huge. The Lions would recover somewhere around the Eagles' 35-yard line, basically already in position to tie the game. It's a move that's almost certainly worth at least three points (just like the overturned replay).

The risk is 15 yards. Since the Lions had a guaranteed touchback if they kicked the ball off, a regular kickoff would result in the Eagles starting at their own 20. A failed onside kick almost assuredly results in a recover somewhere around the 35 yard line (the required ten yards from where the kickoff started). Sure there's a very small risk of a return, but standard protocol for the receiving team is to grab the ball and get down so you don't fumble.

So the risk is 15 yards, but how much is 15 yards of field position worth? Well, according to Advanced NFL Stats, less than one point. Teams, on average, score around 0.25 points starting from their own 20, and score around 1.1 points starting from their own 35. So let's break down the math:

The Lions recover an onside kick 18% of the time, and if recovered around the Eagles' 35 yard line they are expected to score around 3.1 points. So the Lions, on average, score .558 points if they elect to onside kick (3.1 x .18).

The Eagles recover 82% of the time and score an average of 1.1 points from the 35. So the Eagles have an expected point value of .902 (1.1 x .082).

Together (.902 - .558), an onside kick results in .344 expected points for the Eagles.

A regular kickoff, lets assume, goes for a touchback 100% of the time. So a normal kickoff results in .25 expected points for the Eagles.

So the breakdown is as follows: an onside kick attempt results in .344 points for the Eagles, while a regular kickoff results in .25 points for the Eagles. I guess I'm wrong, huh?

Well, those averages don't consider any context. The Lions defense had been stout all day, and the offense needed all the help it could get. I think that the Lions would've held the Eagles to little or no points from their own 35 (they ended up intercepting Vick on the very first offensive play from the 20). And given how the Lions offense had been playing, the benefit of a successful onside kick was even larger than normal.

But even considering that the statistical breakdown favors a normal kickoff, doesn't the onside kick at least deserve some consideration? The difference in choices is .09 points, and given how the defense had played, I believe the onside kick choice was preferable. But according to Schwartz, no thought went into kicking an onside kick. And, to me, that is unacceptable.

If Schwartz would've come out and said, "we did the statistical analysis, and it shows that a normal kickoff was preferable," I would have no problem with his choice. But hearing that he didn't even consider it, then backed up his decision with irrelevant, irrational excuses, is terribly disappointing of a coach I respect so much.


  1. As a baseball and football fan, I'm personally bugged when stats are thrown out there carelessly to justify decisions.

    Biggest pet peeve: citing some unconditional (or partially conditional) probability.

    Meaning: Your probability of success is not some value X between 0 and 1, nor is it some value X given A, especially if you have not taken B or C or D... into account. It is X given A,B,C,D... It is in reality, impossible to measure exactly.

    Some number off a website is NOT going to give you that true conditional probability. Over a longer term, a lot of conditions average out to noise (meaning zero effect), but in that one single instant, you really don't know your true conditional probability.

    For example, a batter's probability of success is not his batting average. You could take lefty/righty matchups, park, etc. into account for one thing. What if the batter had a hip issue that moment? What if he was distracted by some text message from his girlfriend? Of course, that affects it too, and you'd probably never see that reflected in the probability that some sabermetrics website spits out.

    Anyhow, in the case of the Lions, yes, given that the special teams blew basic punt and kickoff return coverages, and perhaps the special teams spent the week drilling it more than other things, maybe you want to adjust that probability upwards, and the probability of onside kick coverage success downwards, because that may not have been the focus of practice the prior week.

    So I have no problem with what Schwartz said or did, given the special teams issues over the prior two weeks.

    Schwartz has always seemed to emphasize with his teams to do the things they can do well, rather than take chances on things they might not do so well.

    1. Appreciate the well-thought-out response. In some degree, I agree. In fact, assuming you read the entire article, you saw that I mention that statistics don't cover the context of the specific situation both teams were in.

      However, your complete dismissal of statistics, calling them careless, is a bit unwarranted. As you said there are nearly in infinite amount of variables affecting a single decision. That's why it's not fair to say that the Lions would recover an onside kick at a rate above or below the average, and why using the league average (not a random number from "some website") is, at the very least, a good starting point.

      As for the notion that maybe the Lions had drilled on kick coverage more lately and therefore weren't properly prepared for an onside kick, I hope that theory is as ludicrous as it sounds. Given the competitive nature of the NFL, you have to be prepared to kick an onside kick every week. If Schwartz didn't have the team prepped for the possibility of an onside kick, that's an entirely different failure on his and Crossman's part.

      In the end, the purpose of the stats was to show that, in general, this decision was a toss-up. And Schwartz's absolute refusal to acknowledge other options was disappointing. It shows that we can place him in the large majority of head coaches who are too afraid to consider unconventional methods.

  2. I am with Anonymous here. While I would have liked an onside kick there I have not problem with the choice or the explanation. When your special teams have been historically bad the past few week I see no reason to risk it. On the other hand I still think we should have gone for it because "Fortune Favors the Bold!"