Analysing the Lions' Saturday Night Gambles

Contrary to popular belief, Dan Campbell didn't lose the game for his team. I believe he's the reason they nearly won, and I'll tell you why.

I will give fair warning that this article is getting set to be quite stat heavy. If that's not your deal, perhaps you should skip this one, but I promise I'm going to try to be as educational as possible in handling the intricacies of NFL fourth down decision making. To that end, I'm going to make heavy use of Ben Baldwin's fourth down decision calculator throughout, so thanks to Ben for that.

There's been a lot of talk about the Cowboys-Lions Saturday Night game, specifically pertaining to the Lions' persistent gambling throughout the night, which normally works out for them but this time didn't. As with most NFL games, there were many instances where there was a decision to be made. Ignoring the obvious punts on something like fourth and 13, there becomes nine crucial decision points the Lions faced last night, and I'm going to analyse all of them.

Fourth and Two. DET 28. 10:01 of the Second Quarter. 3-7. Detroit Win Probability: 19%

At first glance, fourth and two from your own 28 doesn't seem like a clear go for it position, but when you look at the fourth down decision calculator for this scenario, you see this:

A 2.3 percent expected increase in Win Probability from going for it on this play means this is one of the clearer go for it situations we're going to run into today, but we must ask ourselves why? Surely while watching this game it did not strike any of us (myself excluded. I work with this stuff all the time) as a clear go for it situation, but according to the numbers it's clear as day. Why the distinction?

There are two main reasons: The first is that converting on a fourth and two just isn't that hard. Every offence in the NFL averages more than two yards per play, so just any average play from any average offence will convert this. Many people when doing fourth down decision analysis on the fly are scared by the spectre of what happens in the event of failure, so scared that they ignore how easy success will come.

Please don't take this as me calling any of you dumb or cowards. Even NFL coaches fall victim to it.

The second reason is that Detroit is already in a big hole, a bigger hole than would be denoted by just looking at the four point deficit. Dallas has so far touched the ball four times. One drive was a 92 yard touchdown pass to CeeDee Lamb. Two were ruined by turnovers deep in Detroit territory. One was ruined by an offensive penalty. None of these things are reliable, and if not for some luck breaking the Lions' way, they could easily be in a 21-3 hole right now. The Lion offence has also played fairly well themselves, so it's a good bet converting this fourth down won't just be buying themselves three more plays before the inevitable.

Where most coaches would've caved to their fear and forced their defence to try to get lucky against the Dallas machine again, Dan Campbell doesn't. The Cowboys somehow don't come out ready for one of the more obvious fake punts of the season, and it's first down Detroit, which takes us to our next decision point.

Fourth and Goal. DAL 4. 5:45 of the Second Quarter. 3-7. Detroit Win Probability: 30%

We're later in the same drive as above and interestingly, I have a feeling more people would be sympathetic to going for it here than were a few minutes ago from the Detroit 28, but when you look at the numbers, this situation is actually more marginal:

Generally, any fourth down decision within plus or minus 1.5 percent expected change in Win Probability is quite marginal, which means that it's not as much a mistake as a small missed opportunity if you make the incorrect fourth down decision, so if Dan Campbell had missed the first decision (which recall was 2.3 percent) it would've been a bad miss. I don't feel so strongly about this one.

The reason is obvious. Converting a fourth and four from the goal line is dramatically harder than converting a fourth and two from your own 28, but the reason the model (and the fans) still likes to go for it in this spot is that failing just isn't that bad.

Look at the difference between the estimated WP following a successful field goal and a failed attempt to score the touchdown. The difference between a 22 percent and a 29 percent chance to win just isn't that much, especially so early in the game, because turning the ball over to a Dallas offence backed up on their own four yard line just isn't that dangerous. It's not the same as a turnover anywhere else on the field. Therefore, there's almost no consequence to failure, aside from the assured three points.

If you take the same position that the model (and myself) take, that those three points are not especially important because Detroit has had no success at all stopping the Dallas offence (turnovers are more luck than anything), then you can see quite clearly what Dan Campbell saw, and understand why he went for it here. He failed, but Dallas was so backed up they couldn't score in response, giving the ball back to Detroit to try to score before the half, which takes us to our third decision point.

Fourth and Five. DET 45. 0:44 of the Second Quarter. 3-7. Detroit Win Probability: 21%

Dan Campbell has been perfect on his fourth down decisions so far, but that streak is about to break as Detroit's aggressiveness suddenly falters right before the end of the first half. They dropped back and punted without much fanfare, but the model says maybe they should have looked at this one a bit harder:

This is again a marginal spot, because the punishment for failure is again just not that great. The difference (in Win Probability terms) between being down four and down seven is a lot smaller than you'd think, and even in the event of failure, Dallas is at the 45 with less than 40 seconds left and one timeout. Their chances of scoring a touchdown from here is exceedingly small, but Detroit's chances of scoring given a conversion are significantly higher because they're left with two timeouts.

I won't begrudge Dan Campbell for missing this possibility, as it isn't big enough to swing the game, but it felt watching the game as if it wasn't thought through hard enough.

Fourth and Two. DAL 41. 9:01 of the Third Quarter. 3-7. Detroit Win Probability: 26%

I almost didn't include this one, because of how obvious a spot this was to go for it, but with how often I analyse games from the 2000s and use the phrase "now-extinct punt from the 40," I figured it'd be disingenuous not to include it. Here are the numbers that explain why it's so obvious to go for it on the other side of the 50:

The rationale is very simple. Look at the WP after failing at an attempt to convert, and compare that to the WP following a punt. They're almost the same number, and the punt eliminates any chance of a reward, meaning that unless you intend to try a 60 yard field goal there's really only one option if you find yourself in fourth down from the 40. You must go for it. There just isn't enough of a difference in the opposition's chance of scoring whether they're starting from the ten or the 40 to justify a punt.

This is all compounded by the fact that this is a fourth and two, meaning the Lions have a well over 50 percent chance to convert it. It all adds up to a very obvious fourth down try out of the Lions, which they do manage to convert, and go on to score a touchdown to take a lead, helped quite a bit by their fourth down skills.

Fourth and Three. DAL 7. 12:21 of the Fourth Quarter. 10-10. Detroit Win Probability: 52%

Would you look at that? All of a sudden we've gone from trying to make shrewd decisions just to hold ourselves in the game to having a higher than 50 percent chance to win. Now tied, we find ourselves in a weird spot now, again down by the Dallas goal line, but this time not needing a touchdown to avoid a turnover, which is a big difference. The model again sees this as a marginal spot:

A field goal means a lead, but it also means forfeiting a wonderful chance at a touchdown lead, and because there's no need to score a touchdown on this play, it's actually quite a bit easier to convert than would be a fourth and goal of the same distance. You can see the chance of conversion is 51 percent, meaning success is more likely than failure.

Your opinion on this call comes down to how much you value a touchdown lead vs a field goal lead. If you're like me and think having a touchdown lead is immensely valuable, especially against a team like Dallas that doesn't tend to have trouble scoring, then you can easily see the rationale for going for this. For a team with more faith in their defence, I think a field goal attempt could've been justified right here, but for a team like Detroit I think they should've gone for this, especially because even in the event of failure, the Cowboys are starting inside their own ten again.

In a moment that's a bit out of character for them, perhaps indicating they have no play call for this situation, the Detroit Lions decide to err on the side of safety and accept a field goal and a 13-10 lead from this spot. The Cowboys immediately prove this decision incorrect by easily and immediately scoring a touchdown to go up 17-13, and Detroit in their response touch is quickly faced with another tough choice.

Fourth and Five. DET 30. 5:59 of the Fourth Quarter. 13-17. Detroit Win Probability: 19%

The fairly good spot from the beginning of the fourth is gone, and we are now faced with a potential three and out on our response touch. It's fourth and five from the 30. This went down as a relatively uncontroversial punt, but it shouldn't have, and here's why:

This was a clear go for it spot that was just missed by a team that does not miss many. The reason it's so clear is because the Lions are staring down a serious risk of never getting the ball back. With six minutes left on the clock, facing a four point deficit, looking across the field at a Cowboy offence that just got done hanging an easy touchdown on you, this is an obvious spot to go.

This is a very similar spot to the one in the second quarter, a spot in which Detroit went for it without hesitation and converted. I'm not sure what they're so scared of here. A 48 percent chance to convert is pretty good, still less than inspiring, but you only punt this ball if you have serious faith in your defence to get you the ball back, and get it back within a touchdown. If you have that much faith in your defence, why not go for it and rely on them to hold up on the back end, if it even gets that far?

Going for it and succeeding doesn't give Detroit a great chance to score. It only gets them to the 35, but it keeps the ball out of the Cowboys' hands. Going for it and failing gives Dallas the ball at the 30, already in field goal range, but it at least guarantees the ball back for Detroit. Punting doesn't guarantee anything. If not for a clutch Aidan Hutchinson sack on a third down, the Lions wouldn't have gotten this ball back at all.

I normally love Dan Campbell, and I think he's a great coach, but this was a clear and obvious decision that he missed, and it came within one Cowboy third down conversion of costing his team the game. It's the type of mistake that he doesn't normally make, but he made it here. I'm not sure if the pressure got to him, or if it's even my right to say such a thing, but this is one of the bigger games Detroit has played all year, and in the biggest spot of the game Dan Campbell made an out of character mistake.

Just a thought.

Two-Point Try. DAL 2. 0:23 of the Fourth Quarter. 19-20. Detroit Win Probability: 31%

Thanks to that key sack, Detroit did get a chance to get the ball back and here we are. They've just scored a touchdown, and now they must decide whether to try for one or try for two. In fact, it's not much of a decision, and I can prove it:

The numbers reveal an obvious spot to go for two, and it's based on a simple assertion that evidently Dan Campbell agrees with: The Lions are not going to win a straight fight in overtime with the Cowboys.

Going for two in this spot raises the Lions' chances of winning. It also increases their chances of losing. It manages to do both by eradicating the chance of a tie and an overtime. The rule of thumb is that if you're the better team, it's okay to be content with overtime. If you're not the better team, you better go for the win right now, because you're likely not going to win an overtime.

It's not a controversial statement to say the Cowboys are better than the Lions, especially at home, and so in this rule of thumb, the Cowboys want overtime. The Lions do not. Both sides prove my assertion correct because the Lions line up to go for two. The same philosophy holds for all three two-point tries. The Cowboys wanted overtime, and the Lions didn't.

For a man who nearly cost his team the game not being aggressive enough just six (in-game) minutes before, I have immense respect for Dan Campbell for doing what's right to try to get his team the win. It would have been really easy to have faith in his players and just choose to go to overtime. It's much tougher to be realistic and come to the necessary conclusion that your players are just not good enough to win when it comes down to an OT. NFL coaches are not known for making the tough choices, but Dan did, and he deserves big credit for it.

In Summary

The Detroit Lions faced nine decision points in this game. Six of them were quite obvious (but only five of them were chosen correctly). Three were marginal and easy to miss, or which two were missed. In all, the Lions got six out of nine fourth down decisions correct, which is about in line with their season average, denoted in the following graph:

The Lions' fourth down daring held them in a game they could easily have fallen out of in the first half, and their sudden conservatism nearly lost them a game they could have won in the fourth quarter. In all, I think they did a good job, and their fourth down superiority has done nothing to hurt their standing, either throughout the season or in this game.

I've heard many talking about Detroit's aggressiveness, and how it potentially lost them this game. I disagree, I believe their daring is what held them in a game against a superior opponent that they easily could've lost, and I think the players respect a coach that has the balls to make the hard choices to try to help them.

Many people ignore the successes of Detroit's many gambles in an effort to focus on the failures. You can continue to do that if you'd like, but now that we're all informed on the positives, perhaps you won't want to anymore. That's my wish anyway.

Enjoy the gamble. It's fun to gamble, and it's fun to watch your guys do it in your stead. In fact, it can actually benefit you quite a bit, and it's important we all remember that. Sure sometimes you'll shoot yourself in the foot, but more often than that you'll win. Especially when your coach is Dan Campbell.

Great job, Dan.