April 23, 2014

Overrated Baseball Statistics – Pitcher Wins

In the first of what I hope to be a series of several articles, I plan to take a look at some of the more overrated and underrated statistics in both baseball and basketball. Today, I begin with the one that started this whole thought process for me – pitcher wins in baseball. For any feedback or input on ideas for additional metrics to explore as either overrated or underrated, feel free to drop me an email at kirk@waitingfornextyear.com

Corey KluberTwo weeks ago, when I was watching the Indians lose a couple of hard-luck games with Corey Kluber and Scott Kazmir on the mound, it only reinforced what stat-geeks (like our own Jon and Jacob) have been saying for a while now, pitcher wins are one of the most overrated statistics out there being heavily used. It got me thinking on a larger scale too. What are some of the most overrated and underrated statistics in baseball?

I’m as guilty as anyone of relying on these stats, what with my 4×4 fantasy baseball league that uses wins (and saves) as one of the four pitching categories. Let’s take a look at why wins are just not very good indicators of pitching success in small sample sizes.

This article from Glenn DuPaul gives a great feel for how some in the baseball community feel about pitcher wins.

“I dream that one morning I’ll wake up to a world in which the pitching win no longer exists as a baseball statistic. I can’t stand pitching wins. Pitching wins may not be the worst baseball statistic ever invented, but quite honestly it is the one I like the least. I listed some quotes below that always seem to make me cringe:  “He pitched well, but didn’t get the win” “He never pitches great, but he always pitches just well enough to get the win” When I read or hear quotes like that the first thought that always pops into my mind is an exasperated “Who cares?” But for some reason people still care and I’m slightly pessimistic that we may ever see the day when people don’t talk about pitching wins. The sabermetric community hates pitching wins, but not everyone listens to saber-heads. Pitching wins still alter the perceptions of many baseball fans, which is the main reason why I can’t stand the statistic. 

So much of it depends on the style of the ballclub. The Indians are often a late inning team. Just looks at how their bullpen boasts some win-heavy pitchers with Cody Allen (5), Joe Smith (4), and Chris Perez (4), Bryan Shaw (2), and Matt Albers (2) all having multiple wins. That’s 17 of the team’s 61 wins right there. Hill, Pestano, Langwell, and Martinez each had a single win this year out of the bullpen. Therefore, only 40 of the team’s 61 wins have come from the starters this season.

There’s the concept of the “vulture win” among bullpen pitchers who will record a single out and then become the pitcher of record, perhaps scavenging the win from a more deserving pitcher. Even worse is the scenario where a pitcher blows the save and then are the beneficiary of an offensive assist in the bottom of the inning or the next inning.

What do I think is probably the best indicator in place of wins? That’s a really difficult question, and if I had a really good answer, I’d probably have some cool nickname like #JakeyStats or something. To me, it is a combination of quality starts, what the game situation was when they left, and team record when a pitcher starts. Let’s look at the Tribe’s records when their primary five starters pitch.

Masterson – (15-8)
Jimenez – (14-8)
McAllister – (8-6)
Kluber – (10-8)
Kazmir – (10-9)

Doesn’t quite fit what you’d expect, does it? The main thing to point out of this is a guy like Ubaldo that has not been going six innings a lot, but IS keeping his team in the game, giving them a chance to win it in the later innings. That’s a half dozen games that the Indians have won where Ubaldo has started and didn’t earn the win (Ubaldo left ahead in two of those games and tied in another).

I’d love to see the concept of “legacy wins” and “scorer wins” emerge. The idea being that for historical purposes and the letter of the baseball law, you could and should maintain the way pitcher wins are determined now with “legacy wins”. Then, as a supporting statistic, you put the pitching decision in the hands of the official scorer with “scorer wins” that gives the official scorer the ability to determine who is deservedly the winning pitcher. This already occurs when a starting pitcher fails to go five innings and his team leads when he exits and maintains that lead. This would basically eliminate the vulture win and help adjust win totals more appropriately to starting pitchers as the wins statistic is much more visible and relevant for starters as opposed to relievers.

Here are just a few of the games where the Indians starters lost what should or could have been wins.

May 18th – McAllister pitches 7 1/3 innings against the Mariners, allowing just two runs and departs up by two. Chris Perez blows the save in the top of the ninth and earns a win for his trouble as the Indians win it in the bottom of the ninth. ((This is the number one scenario I’m looking to avoid happening with my new system.))
June 26th - Kazmir pitches seven innings, allowing one unearned run and exits ahead by one run. The bully surrenders the lead against the O’s but goes on to win 4-3 by scoring the two runs back in the top of the ninth. Joe Smith earns the win after being the one that blew the lead.
July 7th – Kluber leaves with a runner on second and one out ahead by five runs in the seventh, the bullpen blows the lead. The Tribe goes on to beat the Tigers 9-6 with Cody Allen retiring the only batter he faced and earning the win.
July 26th – Kluber left after six innings ahead by three runs. Cody Allen and Joe Smith blow it, but the Tribe wins it in 11 innings and Bryan Shaw gets the W with an inning of scoreless work.
July 31st – Kluber exits one out short of nine innings with the game tied. Tribe goes on to win in the bottom of the 10th. Chris Perez gets the W after a scoreless inning of work in the Top of the 10th.

I still think wins serve a purpose. They tend to indicate the success of a pitcher over an entire career. In a single season, however, or certainly in a smaller sample size, run support, how good the team is overall (defense, bullpen, etc.) can all have drastic effects on whether a starter gets the W in his column, regardless of whether his outing is really deserving of one or not. I thought this piece from Joe Posnanski puts pitching wins in the proper perspective.

“So you can see that a winning starting pitcher is almost twice as likely to pitched fewer than six innings as throw a complete game. How absurd is it to say that a pitcher who threw 5 1/3 innings WON a game? Not too absurd, apparently. We say it all the time.

But, today, I come today to praise the win not to bury it. Because in addition to the pitcher win being fun to argue about, it also connects us to the game’s history — which I think is a good thing. Bill James was probably the most outspoken critic of the whole “error” concept in baseball. It irritated him to no end that baseball, which as a game is set up for such a clean statistical record, would allow fielders to be judged and catalogued by how they looked from the press box rather than how many plays they actually made. But Bill has also told me that he would not want to get rid of errors because they such are a part of baseball. That’s how I feel about the win too. I like referring to Steve Carlton as a 27-game winner in 1972. It’s a common language in a time when common language is becoming rarer.

I certainly think wins have SOME correlation with a pitcher’s ability or performance. Not nearly as much as The Mighty Win people want, but good performances obviously tend to lead to wins more often than bad performances. Winning 300 games in a career is a remarkable thing, and winning 20 games in a season is an accomplishment. There have been 72 20-win seasons since 1990, and all of them were good seasons. Some were a lot better than others. Randy Johnson’s 20-win season in 1997 was a whole lot better than Bill Gullickson’s 20-win season in 1991.”

So, for now, wins stay a major part of the statistic that we share and analyze on a daily basis. They probably will for a long time, but just remember all of the things that have to go right for a pitcher to get that W and how the man on the mound isn’t in control of so many of those.

(Photo: Chuck Crow/The Plain Dealer)




  • The_Real_Shamrock

    Kirk all you have to do is watch the show with Brian Kenny and Harold Reynolds on MLB Network at 4 pm ET where they argue over your topic daily. Sorry to say I’m old school (Reynolds) and prefer the sacred stats including wins. There are always instances and games where one could argue whether the pitcher did well enough to win or not bad enough to lose. This doesn’t mean we need another statistical category.

  • mgbode

    when you can give up multiple runs in 1/3 inning and “earn” the win, something is wrong with the statistic.

  • hans

    “There are always instances and games where one could argue whether the
    pitcher did well enough to win or not bad enough to lose.”

    That’s the problem, a pitcher doesn’t “win” or “lose” a game, a team does. A pitcher, pitches the ball. He has influence over the velocity and location of the ball (and deception and movement as well).

    It is perfectly fine to enjoy the metric, and to discuss if a pitcher did well enough or not to do his part in a team’s win, but the W-L metric unfortunately is used inappropriately with little or no validity to what people think it is measuring.

    ERA is far from perfect, but is at least solely related to the defensive aspect of the game. There is no reason to include the offensive aspect of the pitcher’s team when evaluating the pitcher’s ability to prevent runs. None.

  • The_Real_Shamrock


  • The_Real_Shamrock

    That’s part of the game some might say the beauty. Regardless I’m not in favor of creating statistical categories in order to micro manage every possibility. It’s the love affair with statistics that has helped foster the PED generation as far as players go. It’s not solely responsible $$$$ is a bigger part but it also relates to the statistics, IMO..

  • mgbode

    baseball players have always loved statistics. they are part of the game. we just keep refining them to the ones that we prove matters.

  • The_Real_Shamrock

    Much of what baseball is and has become is because of statistics heck it’s why I play fantasy sports in general. That being said I’m more of a historian and don’t feel that more statistics have to be created in order to micro manage every statistical possibility. Wins will never go anywhere. You have the magical 300 win mark and I believe an entire section in the HOF. If you want to redefine or tweak that’s one thing but I’m firmly against the elimination.

    Of course baseball has changed dramatically over it’s existence. Whether it’s the 100 pitch mark or the creation of bullpen specific jobs such as setup man and closer. I’m sure eventually something will change but for now nothing dramatic is needed. People have a tendency to go completely overboard with statistics.

  • nj0

    Did you just blame sabermetrics for PED’s in baseball?

  • The_Real_Shamrock

    No but I can see how you might think so. I was eluding to $ = statistics and to improve their stats these dopers took short cuts and cheated.

  • Steve

    “the love affair with statistics that has helped foster the PED generation”

    Of all the inane things . . .

    Though, its kind of true. MLB has been in love with statistics (just a lot of wrong ones) and involved in PEDs, since before 1900.

  • The_Real_Shamrock

    It’s Monday I’m well rested and ready for you Steven even after your misdirection of a comment. First calling my comment inane then agreeing. Your like some kind of Jedi master me thinks. You are my hero Steven. You are the wind beneath my wings. You complete me!

  • woofersus

    This is one of the more “captain obvious” moments, and it highlights the resistance of some old school baseball people to rational thought and consideration of new and different ways of measuring a player’s performance. Perhaps, once upon a time when pitchers usually pitched the whole game and probably played on offense in more than a token, ceremonial way (I”m looking at YOU National League!) you could say they held a greater stake in their own winning percentage, but to judge a pitcher on the production of the team’s bats is ludicrous at best. The W/L records of the current Indians pitchers is a good example, and the situation with other teams is undoubtedly the same. Just last night we saw a guy go 7 innings with 1ER and not get a win.

    Aside from evaluating pitchers on other, more direct measurables, like ERA, WHIP and average against, I agree there might be another way of looking at pitchers in a results-based way that is more fair. For example, Jiminez isn’t our best starter, but he doesn’t give up more than a couple of runs very often and generally gives his team a chance to win. You could argue that’s more important than his ERA or WHIP. Maybe he’s more valuable than a guy with a lower ERA and WHIP who throws shutouts and then blows up alternately. On the other hand, when the team wins he contributes less to the W because he doesn’t give you long outings.

    Maybe a combination of IP per start (since pitchers doing poorly tend not to stay out as long, and pitchers who can’t give you many innings in a good start contribute less to a potential win) and percentage of the time the TEAM wins when that pitcher starts. (not counting a no-decision against the pitcher if they did well) I also think there should be an adjustment for total runs allowed. I believe this new magical SABR stat needs some more data, but what I’m getting at is calculating the pitcher’s share in a team’s success on days that they start. I’m open to suggestions.