I was asked the other day if Francisco Liriano should be considered for closing duties. And then I was lectured on how he is just too valuable as a starter. Before we tackle that question, let's talk a little about statistics, and where they fail.

Photo courtesy of aturkus

There are some who say that statistics have ruined baseball. That goes too far, but I'll heartily agree that they distort our view in subtle ways. And I'm not just talking about new-fangled ones like WARP (Wins Over Replacement Player), UZR(Ultimate Zone Rating), FIP(Fielding Independent Pitching) or even PGKMNBISTAA (Please God, Kill Me Now Before I'm Subjected To Another Acronym).

No, I'm talking about more basic stats. Like home runs, hits or innings pitched.

We've counted these for so long that we have forgotten that they aren't distributed neatly. Some at-bats are essentially useless, while some are the difference between a win and a loss. That goes for innings pitched, too. There is a reason that Bobby Keppel pitched some innings and Joe Nathan was saved for others. Nathan's innings were worth more.

We inherently understand this when we're watching an important game. We'll praise the single to lead off the eighth in a one-run game. We'll shrug off the two-run dinger in a blowout. The latter looks better on the back of a baseball card, but the former is the one we value. If only there was a statistic that did the same.

Well, there is. And not too surprisingly, it is often disregarded by the very people that usually love statistics, because it attacks the neat little counting they like to trumpet. It's called Win Probability Added (WPA) and it tabulates how much a player has added to his team's probability to win.

Now this is the part of the entry where I usually detail how WPA is calculated. And it's also the part where your eyes glaze over. But I still do it for one very good reason: I just can't help myself. The details of this little stat and how it works are just so very elegant that a math major is helpless to resist. Unless of course, he's in a hurry to fill out his NCAA bracket. Which I am. So consider yourself spared. If you don't want to be, you can dig into the pretty details here.

Yes, I said pretty.

Here's the Cliff Notes version:

  • Each player gets points based on how much they increase their team's chances to win a game. So, since each team starts with about a 50% chance to win, if a player led off a game with a home run and then pitched a perfect game, he could win 50 percentile points.
  • A player's score isn't dependent on how anyone else on his team does. A leadoff single doesn't count for less just because the batter ends up stranded. A pitcher who throws a scoreless eighth isn't hurt because the closer gives up the save in the ninth.
  • A player can either have a negative or positive score. If they do something good, it's positive; bad, it's negative. The points are dependent on how other players did in the same situations throughout baseball history.

Now, back to Liriano starting or closing.

When you study WPA for pitchers, you find some interesting truths. One that is very tough for stat boys to accept is that innings in games can get MUCH more valuable as the game goes on. There is a simple reason for this: if you blow a lead later in the game, your team has less opportunity to win it back for you. It's bad to give up a lead in the third inning. It's brutal to give it up in the ninth. Alternately, it's worth a lot more to hold the opposing team scoreless in the ninth than it is in the third inning.

You can see this clearly when you compare the WPA scores for relievers and starters. Even though a reliever might pitch one-third fewer innings, his WPA scores that can exceed those of starters. In other words, a reliever who consistently shuts down the opposing team in late innings can be more valuable than a starter who pitches well for most of the season.

That happened last year for the Twins. Not only did Nathan lead the pitching staff in WPA (3.89 which is 7.5 games to the good), he contributed three times more than the Twins best starting pitcher. And the Twins weren't alone. Thirteen other teams had a reliever play the role of the most valuable arm. Another eight had a reliever as their second most valuable arm. And only one team, Atlanta, didn't have a reliever in the top three.

That's the power of late innings, and it's a door that swings both ways. Guess who was the LEAST valuable player in the league last year according to WPA? It was Brad Lidge, the Phillies closer, who single-handedly cost his team the equivalent of nine wins. Most closers don't rack up that kind of damage because if they do, they aren't closers any more. But the Phillies stuck with Lidge and it cost them time after time.

(Apropos of nothing, the lowest WPA on the Twins pitching staff belonged to…Liriano. He cost the Twins the equivalent of 4.5 wins.)

One can argue a lot of things about a closer. One can argue that they're more replaceable than a starting pitcher. One can argue that they're overpaid. One can argue that if you had to choose between making a player an ace starter or a closer, you should choose the ace starter.

But you cannot argue that a closer – or other late-inning arms – are not extremely valuable. Indeed, they are likely to be more valuable than all but your very best ace starter. The role is critical, and it is critical that the person who takes it has success.

Personally, I don't think that person is Liriano. But if the Twins do, they won't get any argument from me about him being more valuable someplace else. When you go beyond the counting stats, you find there are precious few places any pitcher can impact a team more than anchoring a bullpen.

For yesterday's comment, I'm going to run with one from obtusebanter, who responded to Nick's quote:

"If he can't take a legitimate step forward this year, I suspect that the Twins will seek to move in a different direction."

Agreed. The mental distractions are gone. I can't think of an excuse should he fail this year.

And I think I'll disagree. If he doesn't breakthrough, Young will cost the Twins about $4.5M in arbitration next year, and I think they'll go another round at that cost. It's a tough call, but Young is starting to remind me of Luis Rivas – skilled enough to tease, disappointing enough to be affordable. And so the Twins just kept bringing him back year after year.

I've already given you one link to WPA, but if you're looking for more, I'd highly recommend checking out FanGraphs.com, which is the source I used for all of these WPA values.

Elsewhere….

  • Nick has been reviewing the Twins various positions as spring training approaches
  • In case you missed it, an acclaimed baseball book by a local author came out in paperback this week. You can find out more about it here.
  • And if you're fed up with the NCAA tournament and want to talk baseball - get over it. Even the baseball players are obsessing about their brackets this weekend. I'll be doing so solidly for the next 48 hours. If you'ld like to watch my hopes and dreams slowly crumble yet again this year, feel free to follow me on Twitter.