Minnesota vs. Northwestern: Statistics, The Dangers Of Acting Like Tim Beckman And The Falsity Of Broad Claims About Minnesota

UPDATE, 2:05 PM 10/7/14: The post was updated after it's initial posting to better identify the author of each of the articles I linked to. One additional update related to what one of the Inside NU writers meant with one of his quoted statements has also been included.

I must confess I've been left feeling a little confused this week. In a week where the cliche joke is to pretend we're all Ogre and scream NEEEEEEEEERDS really loudly (or perhaps more appropriately, communicate the same sentiment via GIF) I've watched the fans of the more exclusively academic collegiate sporting club go out and use Ogre math to validate their favorite team. What's even more baffling is that I don't think it's needed to make their case the result is that the good arguments get lost behind the bad ones.

At first this seemed to come from a rush of fan euphoria following two unexpected wins, so I wasn't too concerned. After all, we've all been there. Seriously, I'm pretty sure if you looked hard enough you could find examples of almost every fan and most fan bloggers ---myself included--- resorting to the "if only" line of thinking. In fact, I can guarantee I've used it in previous years to explain away problems against Iowa (#FREEFLOYD!). As someone who tries to look at evidence over "gut feelings" I'm not proud of that but I'm not going to pretend it didn't happen.

Nothing however, prepared me for the pairing of these two stories.

I think we all know where the Beckman story is going. I mean, he's Brewster reincarnated as a head coach. So you know the quotes in Roger's Becks piece are going to be good...

#illini Beckman: You want your players to be succesful. We take out 10 plays in that football game, it's a different game.

— Jeremy Werner (@WernerESPNCU) October 6, 2014

Tim Beckman: "You take out 21 points on three plays and we’re in this football game." Yes, he really said that.

— Steve Greenberg (@SLGreenberg) September 14, 2014

What is the real truth? The final score? Or the final score that could have been?

Do not mock Tim Beckman. Beckmanism is not the set of ramblings of a failing football coach. Beckmanism is a philosophy, a world view. Gandhi is often credited as saying that you should "be the change you wish to see in the world." Beckmanism offers a different approach: Don't be the change you wish to see. See the change you wish to see. Look: It's already there, bright and gleaming and beautiful, everything that you had hoped for. Why fantasize about a beautiful future when you can just say it exists in front of you?

(GoAUpher Note: We'll miss you Becks...)

Unsurprisingly the Beckman story was "Becks is a moron" gold. But what about the oddly titled NU rushing defense story? Why was I so surprised to see it next to WIN FIGHT EXCUSE Beckman quotes piece? It's because I knew the context of the headline thanks to Henry Bushnell's post about NU's defensive perfomance vs. Wisconsin:

Plays like that are emblematic of what this defense has become. The box score will tell you that NU was gashed by Melvin Gordon – and at times it was – but as Fitzgerald likes to say, and said again postgame, "stats are for losers." And this time, he’s right.

On the whole this is mostly useless coach speak. It's the same as a fan pointing and yelling SCOREBOARD. And on the most important level (who won) it can be correct. However, the power of the statement disipates quickly when you start trying to talk about "which team is better?" or "what might this mean for the future?" because unless you plan on arguing with only your feelings, statistical backing is going to be important in those types of discussions.

The really troubling thing is the last line in the above quote. The one spoken not by Coach Fitz, but by Inside NU's own writer. "And this time, he's right." What is it about "this time" that makes him right? Well...

The Sweet Embrace Of Beckmanism

Kevin Trahan's Northwestern run defense story from the photo above gets into why "this time" it might be safe to ignore the stats:

Gordon's career high got a lot of the same reactions from Northwestern fans and opposing fans, alike. They were:

1. How do you get 259 yards from Melvin Gordon and not win?
2. Northwestern's run defense must not be very good.

Neither of these points is very legitimate, unless you look solely at the stats. And in order to understand both of these points, you have to look at the circumstances surrounding them. Or in other words, why stats are for losers. Here's a look at why both of them are misleading.

I agree, black and white statements are often silly and context is important. No Beckmanism here. I see some glaring straw man issues creeping in, but let's see where it goes. Continuing on in the post...

How do you get 259 yards from Melvin Gordon and not win?

This question is faulty because it assumes that Northwestern did not deserve to win the game. However, one player getting 259 yards does not mean that team deserves to win. Aside from Gordon, who is inarguably one of the three best running backs in the country, Wisconsin was abysmal. The Badgers ended with 422 yards of offense, and while they did out-gain NU (422-385) they didn't do it by much.

In the other areas of the five factors, such as turnover ratio (0-4) and finishing drives in the redzone (2-5), Wisconsin was absolutely terrible. Moreover, aside from Gordon's runs, the Badgers weren't efficient or explosive. Their longest non-Gordon play netted 24 yards and they finished with just 19 first down – only six of which came on passes.

In other words, Gordon was really, really good and his teammates were really bad. If his teammates were better and the Badgers still didn't win, then you can suggest that he gave the game away. But a lot of yards from one player doesn't mean that team should have one. Outside of Gordon, Wisconsin just isn't that good.

Keeping the "deserve to win" part out of it (which I think is important since going down that path involves making a host of assumptions about what they hypothetical question asker's intentions) there is still a big problem. You can't remove Gordon's performance and talk about the 5 Factors honestly. The whole concept of the 5 Factors is to compare both teams against each other. Teams, inclusive of all players and their performances. The quoted text above is an exercise in cherry picking that misunderstands or ignores the whole point of the Five Factors. Gordon being really good and his teammates being really bad only matters if Gordon (A) wasn't actually really good on Saturday or (B) did not play. Neither of those things happened.

This wasn't the only recent NU numbers based story to make the same mistake. This is from Jason Dorow's "By The Numbers" post:

Yet, I didn't find his performance as impressive as the stats suggest. Gordon had two long runs, which were 58 and 61 yards respectively, which bumped his yards per carry to 9.6 for the game. On both of those runs, a Northwestern defensive back had a fairly easy opportunity to take Gordon down, but they took horrible angles to the ball and he absolutely blew by them as only Melvin Gordon can do. Yes, Gordon is an amazing back with a incredible mix of power and speed, who is capable of busting long runs, but his big stats should not have blown up the way they did.

We don't live in the world of "should not have" folks. We live in the world of "did or did not" and in this world Melvin Gordon did in fact have his stats blow up they way they did. Don't embrace Beckmanism NU fans. It doesn't become you. Especially because...

Beckmanism Can Hide Good Arguments

The problem with playing "what if" isn't just that it's Beckman silly, it also gets in the way of valid assertions. Returning to Kevin Trahan's the "Stats Are For Losers" piece, we find this:

Northwestern's run defense must not be very good.

There were three plays that NU's run defense was particularly bad on, when Gordon ripped off runs of 58, 61 and 31. Those three runs accounted for 58 percent of Gordon's runs. While they were major mistakes and shouldn't minimize Gordon's performance, they don't necessarily point out a flaw in NU's run defense, for two reasons:

1. There's reason to believe this was an aberration.
2. The defensive tackles were phenomenal.

I think a more interesting (and reasonable) question of "should Northwestern be concerned about facing teams with strong running backs like Minnesota or Nebraska?" could be substituted without issue. Why the change? Well, I know NU's run defense isn't bad (more on that in a bit). The alternate question is a pretty important one coming into this weekend. Why? Because that question (and the two numerical reasons raised in the piece) directly relate to how we should expect Minnesota to do against Northwestern. The post continues...

Wisconsin (and Minnesota) fans visiting this site will think that calling those plays an aberration is an excuse, but statistically, Northwestern has been very good at defending against big plays.

It depends on the context. If you're using it to downplay Gordon's day or the overall success that Wisconsin had then yes, it's an excuse. If you're singling out the plays not to "exclude" them but to note that this sort of big play has been rare against Northwestern this season then it's actually really important, especially to a Minnesota fan wondering what success David Cobb might have. The post continues...

The Wildcats ran 23rd nationally in that category, according to Football Outsiders, and they've consistently done a good job defending big plays for the better part of a year-and-a-half. One of the best running backs in the country getting to the edge on a few plays does not suggest that there is something to be worried about.

The only thing that matters is the 2014 stats, but even restricted to those numbers you'll see that Northwestern is pretty good at keeping opponents from springing the big play. This is very relevant to the interests of Gopher fans. If Minnesota (or any other team) wants to beat Northwestern the stats say that across a whole game they'll need to grind it out rather than rely on big plays. If you're hoping that Minnesota will have an easy time running against Northwestern, I urge you to read the last phrase again. This stat doesn't mean the Gophers won't consistently grab 5 ypc on their way to a W. That's the job of other stats. That's why despite what Coach Fitz likes to claim...

...Stats Are Actually Quite Useful

Kevin's post closes with the following...

Melvin Gordon is a great running back. Northwestern has a good run defense and deserved to beat Wisconsin. These things can all be true, even on a career day for the former.

All true, all verifiable by stats (well, not subjective "deserved" part...that should just be ignored in a stats discussion). Melvin Gordon is an elite running back and despite his career day Northwestern does indeed have a pretty solid rush defense. In fact, it's currently ranked 35th in the country (including stats from 10/4) according to Football Outsiders S&P, which are the stats cited above that said the Wildcats have the 23rd best defense against explosive plays. What does this mean for Minnesota? We shouldn't expect more success than Wisconsin against the NU run defense. That doesn't mean I expect to be shut down, but I'd expect fewer big runs, some TFL, and David Cobb to hover around (or drop below) 5.0 YPC for the day. See? Stats are useful...

...Especially When They Blow Up Fact Free Assertions!

The only problem (well, to me it's a benefit) with using the FO stats to defend Northwestern as both I and Inside NU have is that common refrains uttered about Minnesota this week by NU writers suddenly fail the fact test. What comments am I referring to?

Example 1 - From Kevin Trahan's "Northwestern Is A B1G Title Contender" post:

We've written this twice already, but it's becoming more true each week. Wisconsin was believed to be the best team in the division heading into this game, and how the Badgers have lost to Northwestern. Iowa, like NU, can't move the ball. Nebraska has struggled at times and nearly lost to McNeese State. Minnesota's offense is awful. And Purdue and Illinois are Purdue and Illinois.

On October 4, there is a legitimate argument that Northwestern is the best team in its division. That might be due to the competition, but it doesn't matter. This is who the Wildcats have to beat, and it looks awfully manageable right now.

Northwestern's defense has proven its merit in the past two weeks. There were mistakes on the edge that led to Gordon's big runs, but the front seven is really good, and the secondary has made the plays necessary to win. And the offense looks competent. The offensive line is blocking much better, Justin Jackson is a future star at running back, the wide receivers are finally living up to the hype and Trevor Siemian is improving game-by-game. These wins were not a bad team getting lucky.

The emphasis in both quotes is mine. The takeaway from this piece is clear. Minnesota's offense sucks and Northwestern's is ok. What do the Football Outsider stats have to say about this?

Team	Off. S&P+	Rk	Success Rate	Rk	IsoPPP	Rk	S&P	Rk	Rushing S&P	Rk	Passing S&P	Rk	Std. Downs S&P	Rk	Pass. Downs S&P	Rk
Minnesota	98.2	72	41.80%	81	0.78	94	0.49	86	0.504	57	0.461	107	0.499	101	0.474	68
Northwestern	91.4	99	37.10%	109	0.76	99	0.45	112	0.429	111	0.467	102	0.466	114	0.446	90

Neither offense is great on the whole or even in it's individual components. That includes Minnesota's run game, which the advanced stats find to be only average. That said, by the numbers Minnesota is the better offense across the board with the exception of a single category, Passing S&P (essentially a metric that rates a team's explosiveness and efficiency at passing).

A common retort to this would be that Northwestern has played a more difficult schedule. Unfortunately that wouldn't help disprove what the topline number says. In fact, it actually makes things comparatively worse for Northwestern. Why?

Success Rate and PPP combine to form S&P, an OPS-like measure for football. Then each team's S&P output for a given category (Rushing/Passing on either Standard Downs or Passing Downs) is compared to the expected output based upon their opponents and their opponents' opponents. This is a schedule-based adjustment designed to reward tougher schedules and punish weaker ones.

Emphasis mine. S&P+ is biased IN FAVOR of the team with the tougher schedule. If you grant NU the SOS argument (which I'm willing to do without even checking the numbers) then the claim for Minnesota as the better offense is only strengthened. So which one of the claims is wrong? That Minnesota is awful or that Northwestern is competent? Either way, one of the #NARRATIVES have to go away based on the numbers.

Example 2 - From MountainTiger's "Film Breakdown" post:

Minnesota is another team that relies on running and defense; fortunately, they're worse at both of those than Wisconsin. Just like against Wisconsin, taking an early lead will be important; without Melvin Gordon and the Wisconsin line to lean on, Minnesota will be even more reliant on their limited passing game to fuel a comeback.

Emphasis mine. Without a doubt this is true when it comes to rushing offense. Wisconsin's Rushing S&P rank is 16th and Minnesota's is 52nd. Defensively though, this simply isn't factual.

Team	Def. S&P+	Rk	Success Rate	Rk	IsoPPP	Rk	S&P	Rk	Rushing S&P	Rk	Passing S&P	Rk	Std. Downs S&P	Rk	Pass. Downs S&P	Rk
Minnesota	117.3	11	35.00%	18	0.62	4	0.41	6	0.385	11	0.423	12	0.422	4	0.4	17
Wisconsin	106.2	35	35.20%	20	0.79	46	0.44	18	0.413	21	0.472	35	0.473	29	0.402	19

Minnesota is better across the board. Remember again that S&P+ is biased to Wisconsin here, given the fact that they've played the more challenging schedule.

UPDATE, 2:05 PM 10/7/14: InsideNU wanted to note that they were referring to rushing defense not total defense in their UW/MN comparison. As of now the text is unchanged in their original article but I felt it was worth noting.

Northwestern Fans Reading, Don't Misunderstand My Intentions Or Arguments

- I'm not claiming Northwestern has a bad defense. In fact, the same stats I just used to fisk the dismissive comments about Minnesota say the Wildcats have a pretty good defense!

Team	Def. S&P+	Rk	Success Rate	Rk	IsoPPP	Rk	S&P	Rk	Rushing S&P	Rk	Passing S&P	Rk	Std. Downs S&P	Rk	Pass. Downs S&P	Rk
Northwestern	109.9	22	38.30%	41	0.73	23	0.45	26	0.439	35	0.467	32	0.464	21	0.447	52

- I'm not trying to say the Gophers will blow Northwestern out. My goal was not to suggest this would be an easy win for Minnesota. Instead I was trying to:

1) Call out some pretty egregious cherry picking (and yes, troll a touch by pointing out that it's a milder version of what Beckman engages in).

2) Show that the very stats used to defend Northwestern's defense also refute the broad statements used to downplay Minnesota's units and the associated over-stating of Northwestern's supposed advantages.

3) Try to steer the conversation back to fact based territory.

- The stats don't guarantee a Minnesota victory. They mean Minnesota comes in with a better chance of winning (I'm still expecting to see the advanced stats models say 60/40 or so), but I can see a multiple paths to a NU victory and none of them are farfetched.

- I'm not saying Northwestern fans should be glum, pessimistic, or that they shouldn't be happy following two big unexpected wins. Go nuts! It's what TDG folks have been doing following the Jug win. Unexpected wins should be fun. I'm just saying that saying things like "Minnesota's offense is awful" while claiming "Northwestern's offense is competent" need support and that's why they're getting challenged.

Gopher Fans Who Haven't Jumped To The Comments Yet

- It probably won't be an easy win. There isn't anything statistical you can point to that would suggest an easy win. I know Gopher fans are annoyed with some of the broad comments NU fans/writers have been making. The reverse is also true. If you want your argument to be taken seriously make sure you can back it up.

- I'm still confident in a win. This is a combination of their stats edge and and some gut feeling. On the statsy side, I think the Gophers will need to turn the ball over, play much worse defensively/have NU play much better than usual offensively, or some combo of the two in order to lose this game. That's because I think Minnesota's better defense will give NU more fits than Wisconsin did. On the gut feelings side, I think playing at home and coming off a bye week will allow Minnesota to avoid falling into the early hole that Northwestern would prefer to create in order to improve their chances at winning. I'll probably end up predicting a 3-7 point win, but if it were more than that I'd be shocked.

- In the end, these two teams are pretty darn similar. Both are stronger on defense than offense. Both have had recent performances that create excitement but that haven't done enough to drastically improve either teams statistical rankings (at least offensively). This also means both have more to prove until we can say small sample sizes aren't an issue. Both have losses to a Power 5 school that look better today than they did at the time.

I think the Gophers are objectively ahead in the stats and I'd argue the TCU loss looks better than the Cal loss (no stats there, just an opinion), but on the whole these are teams that are pretty well matched. If this game were on the road I'd be calling this a straight toss up.