More nuanced approaches in sports that aren't baseball

I just read this on ESPN's NFL blog. Pat Yasinkas explaining why he ranked Andre Johnson sixth in wide receiver power rankings:

Here's the link.

Whether you like Andre Johnson or not shouldn't be the issue here. It really bothers me that a high-level football reporter with years of experience is ranking wide receivers by wins. (Stats should be a non-factor as Johnson was the most productive receiver in the NFL on a per-game basis.)

It's this type of analysis that gets baseball writers ripped as shams on a regular basis, the kind of analysis that suggests Bert Blyleven shouldn't be in the Hall of Fame. But in every other sport, it seems to just pass as normal, or perhaps even a quality argument.

I see it all the time. Does it not say something that every statistically inclined person who writes about the NBA -- from the mainstream or blogosphere -- would have voted for someone other than Derrick Rose for MVP?

This isn't an argument for advanced metrics popping up in stories all the time. This is an argument for making good, sound, logical points. And in sports, statistical backing is a good way to do just that.

Yasinkas could have brought up Johnson's ankle issues last season to defend himself (though they appear minor). But he didn't. He chose to, on the record, say that the Texans' season is a reflection of a wide receiver's play. You don't need hard-core statistical proof to know that's not a legitimate 21st century argument.

He's not alone. I bring this case up because it's only the latest. Rose's sure-to-be MVP is another good one, where every statistical measure is ignored.

Even in baseball, this kind of stupidity still rears its head. The difference is, a large section of the MSM has joined the "sabremetrics are valuable even if I'm not sure how to fully understand them" crowd. The same can't be said for other sports, but it's time we all caught up.

 

A win, an RBI and a save walk into a bar. WAR says "Hey ..."

 

Winners win! Winners don't lose!

 

Johnson plays in a pass-heavy offense and gets a lot of yards when his team is in desperation chuck mode. That devalues his stat line IMO as opposed to someone like White, who plays in a run-first offense for a winning team that game-plans to protect leads. To say Johnson is the best receiver because his stats are best is more misguided than what you are saying Yasinskas is. I agree that you can't pin winning and losing on him, but you can evaluate how he arrives at his gaudy stats. Once upon a time Herman Moore put up some damn fine numbers too, but I don't think anyone was saying he was better than Jerry Rice.

Sabermetrics, advanced metrics, whatever you want to call them, in a game like football they are completely out of place. There are a thousand variables that can affect any play, whereas baseball at heart is a pitcher v. batter confrontation.

 

But Yasinkas did dig deeper into any numbers - he specifically knocked Johnson down a peg because of his team's record. There are plenty of statistically ways to tell a WR's effectiveness apart from his pure counting stats and that might give a clue to his worth after factoring in his team's style of offense and other issues. But Yasinkas didn't do that.

 

There are plenty of statistics, but there aren't statistical ways to tell his effectiveness. That is my point. Advanced metrics in baseball have basically been proven (to what degree is debatable and I debate it myself). Advanced metrics in football are hogwash. You don't know a thing by looking at targets, catch rate or anything else unless you know what defense the team was in, what the score and down-and-distance were, how good or bad the throw really was, and about a thousand other variables.

 

Even if the merits of football stats is questionable -- and the good people at Football Outsiders would have more to say about this than I -- can we agree that using wins as a barometer for a wide receiver is plain stupid?

That is, after all, the metric he cited.

 

It will be a sad day for the NFL if the seamhead-type nonsense takes root there. I can just imagine hearing the philosophy that Matt Ryan should only have thrown if he could get the yardage for the first down.

Actually, ESPN The Magazine had something almost as stupid before the start of last season. IIRC, it was some type of analysis comparing the yardage benefits of rushing vs. passing for each team. Dallas, among others, was branded stupid for not throwing more often. We saw what happened in 2010 when the Cowboys threw more often.

 

In the last two seasons Houston has thrown 1170 passes. Atlanta has thrown 1144. That's a difference of less than one per game. Doesn't seem like it would impact the stat line much. Even more interesting, the straight stat lines are skewed by Johnson's injury and White had better raw numbers two of the last four seasons... though in 2007 White was in an awfully desperation chuck it scheme.

... and it seems odd to invoke Moore and Rice. The comparison doesn't work. Rice had better numbers in all but one season of their careers (1997 when Rice only played two games). Moore only won one receptions title outside of the Rice-less season, finishing one catch ahead, nearly 200 yards and a TD behind Rice. And SF threw 39 more passes that year.

 

There's a lot of good football statistical analysis going. Mostly, they know their limits. Most of it is team-oriented, because it's impossible to pull out individual performance.

 

Some of the individual stuff is good as well. Many positions don't work well individually but some do. The DE's translate pretty well when you go past just sacks. Looking at sacks, QB hurries and QB hits go a long way in showing who is good and who isn't. A DE who has 10 sacks but just a handful of the others was a bit lucky and not better than a guy who gets 3 sacks but is among the leaders in the other two. Football Outsiders does (or did) compile those and put them into a simple rating. It showed how guys like Justin Smith are really good despite not getting a lot of sacks.

 

Right.

Because passer rating is nothing at all like baseball sabermetrics.