Back in 2003, a book came out called "Moneyball". The book is about Billy Beane, the GM of the Oakland A's, and his revolutionary way of using un-traditional baseball statistics (like on-base percentage) to find cheaper, under-the-radar players... who could help him compete with the big-market teams.
"Moneyball" ended up being the battle line that was drawn between old-school baseball types and new-school ones. The old school people couldn't believe that anything was more valuable than batting average or RBIs or ERA -- not to mention things like leadership, grit or scrappiness. The new school people couldn't believe how long people had been slavishly devoted to those seemingly antiquated standards of evaluation.
I fall somewhere in the middle. And I decided to write up my list of the 11 most useless sports statistics because, in every sport, certain stats get WAY more pull than they deserve. And someone's got to talk about that.
"Moneyball". The book that Joe Morgan wants to burn.
Wins. I'm continually shocked that a pitcher's win-loss record is the first statistic people bring up... because it is, without question, the least-relevant major statistical category in any sport. The reason: A pitcher's win-loss record has almost nothing to do with him. It has everything to do with how the offense plays.
Here's the one example necessary to prove this. On April 27th, Indians vs. Yankees, C.C. Sabathia pitches 8 innings, gave up 4 hits, struck out 8 and only gave up 1 run (on a solo HR -- meaning one bad pitch out of his 111 pitches). The Indians offense scored zero runs, so he got the loss.
Meanwhile... on May 21, 2000, Russ Ortiz of the Giants gave up eight hits and 10 earned runs in 6.2 innings against the Brewers. But... the Giants offense scored 16 runs... so he got the win.
QB Rating. I know that quarterbacks do a lot of things, so just showing their yards, or completion percentage, or TD passes paints an incomplete picture. But the QB rating statistic is so convoluted and peculiar that it does an even WORSE job of painting the picture.
You never, ever, ever hear a fan saying, "That was a hell of a game by Tony Romo, his QB rating was 99.3!"
Anything in basketball "per 48 minutes". As Charles Barkley once said, if you have to calculate what a player would do in 48 minutes, it's because he's not good enough to actually play all 48 minutes.
Runs. Looking at how many runs a team scores over a season is a very, very important statistic. Looking at how many runs an individual player scores is basically useless. Yes, a player has some responsibility for it through getting on base. But, outside of his homers, it's entirely up to his teammates whether he crosses the plate or not. This is almost entirely a situational stat, and really has no place as one of the baseball big three.
Pancake blocks. I know the world desperately wanted to come up with a statistic for offensive lineman. But, really, pancake blocks? This was the best they could come up with? At least try to make a convoluted formula based on how many times his guy sacks the quarterback mixed with hurries and running yards and even average time in the pocket or something. (Which would be the most awesome stat ever. But if it existed, I probably would've put it on this list too. The moral of the story: Offensive linemen never catch a break.)
Steals. Steals are the most irrelevant of the major basketball stats -- at least as a measure of defensive prowess. While some steals do show good defensive skill -- swiping one from a guy who's dribbling, sneaking in and grabbing an inbounds pass -- a LOT of steals are generated by playing the passing lanes, gambling and leaving the man you're guarding. In other words: To get a lot of steals, you generally have to have the other four guys on the floor pick up the slack you're leaving by gambling.
Number of Gold Gloves. The Gold Glove awards are generally just a popularity contest. Once you win one, you're basically grandfathered in to win more. And once you're famous, your odds of winning go up exponentially.
How do I know? Derek Jeter and Jim Edmonds win al the time, because they make a lot of spectacular plays. The reason they make spectacular plays? They're both really slow for their positions (thus, technically, making them defensive liabilities), so they have to make crazy diving catches and stops. And that leads to Gold Glove awards.
Looks like an opportunity for a Tommy Point!
Tommy Points. Tommy Heinsohn, who does color commentary on Celtics games, gives out his "Tommy Points" to guys (only on the Celtics) who make a hustle play. That means diving for a loose ball, taking a charge, whatever. There's even a website that tracks them.
Yet I have never heard of Tommy Points being brought up in any contract negotiations.
Saves. Saves are a decent way of quantifying what a closer does. But, just like wins by a pitcher, saves are imperfect.
If a reliever comes in with a lead, pitches the final three innings of a game, and the team wins -- he wins. Even if he comes into a 24-0 game. Even if he comes into a 24-0 game and gives up 23 runs. And... you can get a save for pitching with a three-run lead in the 9th. There are lots and lots of ways to get a save in a game that's all but put away.
PER. John Hollinger on ESPN.com made this stat up. It takes into account everything a player does, both good (points, rebounds, shooting percentages, blocks, et. al.) and bad (turnovers, fouls, et. al.) And that's fine, it's noble to try to make a statistic that puts everything into context.
But in his articles, it's THE only stat he will mention. The only basis for comparison. And while making up your own stat is fine, quoting it like gospel is pretty damn narcissistic.
On a side note, PER does a terrible job quantifying a player's defensive contribution (especially the guys who actually play on their man and don't just try for steals), so while it's a very strong stat, it really isn't perfect. And quoting imperfect, proprietary stats as gospel, again, is just not cool.
Holds. I have no problem with trying to find a stat to quantify what middle relievers do. They're the equivalent of offensive linemen in football or defensive specialists in basketball. But holds are just a joke.
They're so much of a joke that no one can agree on how to calculate them. STATS, Inc., says a reliever has to record at least one out to get a hold... SporsTicker doesn't. In other words, according to the second definition, a guy can come in to a three run game, give up a two run homer, walk the bases loaded... and as long as the guy who relieves him gets out of the jam, the first pitcher would get a hold.
This post was originally published on Monday, June 30, 2008 at 12:35:27 PM under the category Sports.