8 IP and No Win

First, a brief quote from the New Bill James Historical Baseball Abstract (p364):

Analysis is about simplification; what all analysts essentially do is to try to figure out ways to simplify the data with the least possible distortion.

Whether I accomplish this in my writing is not for me to decide, but it is my goal. Everything I've done in my professional life (Kmart manager, GlaxoSmithKline pharmaceutical salesman) was undertaken with the intent to make the complex understandable or render order from chaos. I don't always succeed--there are times I re-read my stuff and ask how ANYONE understood what I wrote. When I run across excellent quotes like this, it reminds me of what I try to accomplish--to illuminate, not obfuscate. I will also try to remember this when I start writing some posts on the sacrifice bunt next week, because I got a whole lot of data.

MLB Network's Brian Kenny tweeted this Friday night:

Yes, Kris Medlen was terrific. 8 shutout innings. Did you know 13% of all pitchers this year doing so did NOT earn a treasured "W"?
— Brian Kenny (@MrBrianKenny) September 28, 2013

It was enough to pique my interest. I did NOT know that 13% of pitchers in these types of outings didn't get a win, but I'm not surprised. Part of the research I undertook for the #KillTheWin posts suggested that up to 30% of wins go to pitchers who had game scores below 50 or relievers who threw an inning or two. I won't rehash an argument that won't go anywhere, but since I'm always on the prowl for subjects to research and discuss, Kenny's tweet gave me something to work with.

This chart illustrates the well-known trend of the reduction in complete games, as well as tracking games of 8+ innings pitched since 1950:

It's very important to note the complete game didn't just die in the past 20 years or so but was well on its way out as far as 60 years ago. By the 1970s less than 30% of games were complete games. 8+ IP games march almost in lockstep with complete games, and both are becoming rare, with complete games dipping below 10% in the early 1990s and 8+ IP games in the 2000s. Neither are likely to breach that level in this age of pitch and inning counts.

This chart combines two elements--the number of 8+ IP games and what percent of those games the pitcher got the win, i.e., a just reward for a job well-done:

It's a two-axis chart--the left shows the number of games and the right the percentage of these games in which the pitcher was credited with the win. Especially in recent times pitchers simply don't make it to the 8th inning without having a decent game--what's the point of carrying 12 pitchers on a roster if they're not going to be used? This chart shows the ERA in these games:

Back in the man's man era of pitching where pitchers (didn't) finish what they started, the ERA in these games was still low, but in today's game pitchers are lifted at the first sign of trouble. This makes it especially egregious that 8+ IP pitching performances aren't recognized with more than a "good game, fella" since the reason they didn't get the win was beyond their influence--lack of offensive support, porous defense or a combination of both.

If you read either of my #KillTheWin posts (this link is to the first one) you should already have a good idea as to who will be on this list, which is the pitchers with the most no-decisions in 8+ innings pitched games (from 1950 on):

There's a whole lot of "conventional wisdom" upended in this chart. Don Sutton, who won "only" 324 games had either a no-decision or loss in 89 games in which he pitched his heart out, a 1.60 ERA that even in a pitcher's era in a pitcher's park was stellar--what would people say about Sutton if his team had scored runs in those games and he was over the 400-win threshold for his career? Gaylord Perry had 142 games with no help from his offense, and Bert Blyleven, Phil Niekro, Nolan Ryan and Tommy John are there as well, pitchers that "weren't that good" and yet did everything they could do to influence the game. Blyleven, Niekro and Perry are the poster children for the pitcher that did his job but had the unfortunate experience of playing for teams that were incapable of coming through on offense.

I don't advocate eliminating the win for the simple reason that it will never happen, but I certainly believe we can better educate baseball fans by not focusing on it as much. I understand that a rigid analysis of the numbers suggests that Chris Sale or Hisashi Iwakuma should win the AL Cy Young Award but I'm okay with Max Scherzer winning it--this chart of the 2013 pitchers with the most 8+ IP games (through Friday, September 27th) gives a partial answer:

Part of a pitcher's WAR is, well, PITCHING, and these were the best pitchers in getting deep into games in baseball, none of which are surprising. Scherzer's on that list, suggesting that along with the second-best run support in baseball while pitching, he did what he was supposed to do when on the mound. Did Chris Sale pitch better? Sure, why not, but in the end, was Sale's pitching enough to do anything for a horrendous White Sox team? Apparently not. I understand it wasn't Sale's fault the Sox played so poorly (a quick peek at the Mistake Index shows it was very much an equal opportunity), but at some point winning has to count for something. I use advanced metrics to guide my decision process, not dictate it.

Like any other metric, this is just another way to view data to see if it matches up with our expectations. Like the Bill James quote, I try to simplify the data without OVER-simplifying it, or even worse, making grand pronouncements on the basis of one metric (you can read my thoughts on THAT here). Brian Kenny made a very simple point--pitchers who had pitched outstanding games had nothing to show for it in a significant number of games this year. All I did was flesh out that point with some historical context.