Premium and Super Premium Subscribers Get a 20% Discount at MLB.tv!
February 23, 2012
Inside The Park
Ode to a Terrible Stat
"My choice for the front-runner is Welch, but I know a lot of people say Clemens. I know what Clemens has done for Boston, but now is not the time to change the rules. The guys who won it the last three years won the most games and had good stats. If Bob Welch continues to win at this pace, and he doesn't get it, something is terribly wrong with the judging."
Bob Welch had just won his 20th game when his Oakland teammate was asked about the voting, and it was just Aug. 17. It was his 13th season and the first and last time that the 33-year-old Welch would win 20 games.
"It means something special to me because it took so long," Welch was quoted as saying in an old Baseball Digest. "I talked to a lot of pitchers, great ones like Don Sutton and Tommy John, and it didn't happen to them very often, either. Having to wait so long makes it extra sweet."
Welch kept on winning, adding seven more to his tally over the last six weeks of the season. His 27 wins were the most in the American League since Denny McLain's 31 in 1968. He pitched well, but the win total was a fluke. Welch had 10 different seasons in which he produced at least 2.0 WARP, and 1990 wasn't one of them—he had just 1.1 that season. His 4.55 FRA was easily a career-worst and put him on a career plateau that he never escaped, at least until he retired four years later.
There were 112 pitchers across baseball who had more than 1.1 WARP in 1990; Roger Clemens had 6.3. Nevertheless, Welch easily outdistanced Clemens in the Cy Young balloting. That was just two decades ago.
"If you have always assumed that the number of runs scored in support of various pitchers or various pitchers on the same team evens out over the course of a year, I have news for you. I've been studying pitcher offensive support for six years, and it hasn't evened out yet. The designation 'hard-luck pitcher' is not merely an excuse; it is, in many cases, an accurate reflection of the evidence."
It's hard to say what is the first lesson learned by baseball fans who embrace quantitative research. If you're younger and became aware of the discipline because of Moneyball and its aftermath, that first epiphany might have been that on-base percentage was more valuable and informative than batting average. Others may have realized that teams bunt too often for their own good. I suspect for many, it was the observation that evaluating a pitcher based on his win total is a fool's errand.
It's an easy lesson to digest. There are so many things wrong with pitcher wins that even a saber-novice can easily explain the faults to non-believers. Pitcher wins are utterly absurd. Once you realize that, you can't wait for that first Cardinals fan to get excited when his team signs Mark Mulder, for example, because he won 17 games the year before.
"He gave up 4.75 runs per nine innings," you exclaim. "In a pitcher's park!"
The practice of using pitcher wins as an evaluative tool has been under assault for at least 30 years. By now, understanding of the limitations of the win statistic has moved well beyond the more esoteric circles of baseball analysis. This has been best evidenced by the selections of Zack Greinke and Felix Hernandez as AL Cy Young winners in back-to-back years, seasons in which they combined to win just two more games than Welch did in 1990 alone.
Nevertheless, wins are still everywhere—in box scores, pitching matchup tables, and Who's Who in Baseball, the brontosaurus of preseason annuals. Decisions are posted on scoreboards at the ballpark—even, fittingly, on the old manual board at Wrigley Field. We still refer to a player as a 20-game winner if he indeed has won 20 games. Old-school baseball writers do it, but so do allegedly new-school types like Joe Sheehan, Joe Posnanski and, gulp, Bill James.
Me? I do it too, even though I became aware of the limitations of the wins statistic way before I saw my first Baseball Abstract. It's not that I was precocious as much as I happened to buy a game called Statis Pro Baseball. Great game, if not as nuanced and accurate as Strat-O-Matic, and loads of fun to play. You could tear through a game in 15 minutes, easy. If you ever played Statis-Pro, then you remember that the number of games a pitcher won in real life didn't factor into his table-top effectiveness at all.
The key stat in Statis Pro was ERA. For every batter, a rating based on ERA determined whether the at-bat's result came from the pitcher's or hitter's card, which made it a huge part of the game. In a draft league, you could take an 11-win, low-ERA pitcher and put him on a good-hitting team and he might well win 20 games, real life be damned. Or he might win eight, if his luck was bad.
This of course is all old news, to me and to certainly to you as a Baseball Prospectus reader. It's been old news for many years. In 2011, however, I came to realize that not only was I still was fixated by pitcher win totals, but I no longer cared to break the habit. I understand what wins mean and what they don't and would never in a million years use them as the basis of any kind of debate or serious piece of analysis. I'm hard-pressed even to think of a situation in which I'd refer to them as a data point. Nevertheless, I love when a pitcher puts up big-win totals. Absolutely love it. And I'm not ashamed to admit it.
Why did I finally decide to come clean about this fondness for wins? It's not a why but a who, and his name is Justin Verlander.
"You know for over 100 years won-loss was used to help judge a pitcher. Then some stathead guy comes in and comes up with all these new ways ... sorry, you can sit there and tell me won-loss is meaningless all you want, but when I see a guy win 18-plus games with an ERA below four, I am going to say he is pretty dang good. Even your new stats will reflect that."
I could have pulled any number of old Murray Chass snippets to quote, but I thought this anonymous guy from an Angels blog summed up the sentiment nicely enough. This is a common attitude that I encounter even today, in the 21st century. Not just from casual fans, but from players, team personnel, and mainstream media types. As I suggest, it's an easy debate to win, if I'm feeling contrary.
Welch was 27-6 with a 2.95 ERA in his Cy Young season. That seems pretty unassailable, but we know it's not. He was good that year, no doubt about that. He was a major-league pitcher who put up 238 innings on a pennant-winning team. That's nothing to be ashamed about. But he also put up a .255 BAPIP—22 points under his career mark—and allowed a career-worst 26 homers. He struck out fewer than five batters per nine innings and walked nearly three. Oakland scored 5.2 runs per start when Welch was on the mound and scored five or more runs for him 18 times. Indeed, he had a lot of help and luck in getting those 27 wins.
You can say the same thing about Verlander, but no rational person would question the value of his contribution. His worthiness for the MVP award is another debate, but he was a fairly clear-cut choice as the AL's Cy Young Award winner. A great argument: Who was baseball's best pitcher last year, Verlander or Clayton Kershaw? I'd go with Verlander, but it was close enough to be a classic baseball debate. They were baseball's two best pitchers last season.
As good as Verlander was, it was his mounting win total that first caught the attention of the general baseball public. Well, it was his no-hitter, then it was the win total. He went into Memorial Day weekend with a modest four wins, but then won seven straight starts. He was just getting warmed up. From July 26 to Sept. 18, Verlander reeled off an unreal stretch of perfection—a team-record 12 wins in 12 starts.
When the dust settled, Verlander was sitting on 24 wins. The Tigers had won 40 of their last 57 games and run away with an AL Central race that was close late into August. Verlander's magical streak showed why a big-win season is so compelling, but also why even 24 victories can only tell you so much.
I covered the second win in Verlander's streak at U.S. Cellular Field. He pitched well but gave up a two-run homer to Adam Dunn in the first (no small feat last season), then another two-run shot to Paul Konerko in the sixth that tied the game. Verlander's pitch count was mounting when in his team's half of the eighth, Wilson Betemit stroked a two-run singled to break the tie. Verlander got through the bottom of the inning and departed after 125 pitches. Jose Valverde got the Sox in order to preserve the win, a common theme in a season in which the Tigers didn't blow a single save in one of Verlander's starts.
Good fortune and clutch performances defined the streak. In his next outing, Verlander was up 3-0 after seven against the Angels but gave up two unearned runs in the eighth before striking out Torii Hunter with the tying run on second to end the threat. (The unearned runs were due to Verlander's own error, which always drives me nuts, but that's another story.)
Next outing, Verlander gave up two runs to Kansas City in the seventh, leaving two on and two out for Alcides Escobar. It was a development year for the Royals, so Escobar hit for himself and fouled out to Alex Avila. A nice break. Then the bullpen threw two scoreless innings, and Verlander won by one. That was followed by another one-run win, this one over Indians, in which the Tigers held Cleveland scoreless over the last six innings, the final two by the bullpen.
The next two starts were drama-free wins over Minnesota and Tampa Bay. On Aug. 27, he had the Twins again and wasn't on top of his game. He allowed four runs and nine baserunners in six innings and was done after throwing 120 pitches, departing with the game tied at four. Then Delmon Young and Miguel Cabrera laced two-out RBI singles in the seventh, the bullpen put up three zeros, and Verlander got his 20th win.
Two starts later, Verlander was again done after six innings. He threw 113 pitches and coughed up a pair of two-run homers to Cleveland's Shelley Duncan. The Tigers were down 4-2 when he left. No worries, though—Victor Martinez hit a grand slam in the seventh off Joe Smith, and the Tigers bullpen came through again. I'll never forget the replay of Verlander in the dugout, with an expression on his face that said, "You've got to be kidding me."
That put Verlander at 22 wins, with a streak of 10 straight victorious starts. His next outing was slated for Chicago.
"But Bill (James) has also told me that he would not want to get rid of errors because they are such 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."
This is an essential point that Poz is making, and it underscores the problem that saber-friendly writers face when delving into the grand old game. Certainly there is a substantial audience for math- and acronym-heavy baseball analysis. There are probably not as many fans of advanced mathematics as there are of baseball, but I'd wager that the gap is narrower than you'd think, and there are plenty who have a taste for both. That's great.
I'm not formally trained in advanced statistical analysis, butI have tried to compensate for that by becoming an expert-level Excel practitioner. I lose blocks of days doing stats-based research. (These days, that tends to be truer of basketball than baseball.) At the same time, as a writer I've always viewed statistics as rhetorical, or even literary, devices. They help me communicate my observations about the game and give me firm footing when I'm doing commentary or analysis.
In a nutshell, I feel like stats and sports are inseparable, and if you're going to use stats, you might as well use the right ones. Since I am increasingly aiming my work for larger audiences, I am prone to look for easily-communicated, yet revealing, indicators. Nothing slows a narrative down more than throwing out a metric to an audience for whom you have to insert three paragraphs of explanation.
In that context, I have no problem with using Carlton's 1972 season as a reference point for something else and calling him a 27-game winner, as long as that’s not the only context you're offering. It's sexy. It's tantalizing. It's one of the most remarkable seasons a pitcher has ever had. But you have to be careful, because it's equally true to say that Bob Welch was a 27-game winner in 1990. Depending on your audience and the angle of your observation, you can easily open yourself up to lapses of reason.
When Verlander started to rack up the Ws, I found myself printing out a calendar-style Tigers schedule and marking off the games in which he might start. In my mind, I'd count off the maximum number of appearances. What if the race stays close? Will he go on three days' rest in September? After all, Leyland is an old-school guy. Could he win 30? No. He's not going to get enough starts, and he'd have to win every single one of them anyway. It's not possible. But what if he won 30? I couldn't shake the idea. I rued the lost wins from his first five weeks.
Reality is reality, so I abandoned those 30-win notions soon enough. But I still wanted Verlander to win every start. Twenty-six wins, 27, 28—that would be pretty cool, too. At a certain point, I began tracking a ridiculous statistic: wins per 250 innings, matching Verlander's number against those put up by McLain. McLain won his 31 in 336 innings, or 23.1 per 250. I figured Verlander would end up with around 250 innings (it turned out to be 251), so I decided 24 was the minimum number for Verlander to reach, and I wanted him to blow by it. It was as close to a 21st-century version of a McLain season as we can hope for.
After the season, I compiled the all-time leaders in this junkiest of stats for starters who have had at least 29 decisions, which I realize is totally arbitrary, but in this case, it doesn't really matter. I just wanted a list to look at.
There's old friend Bob Welch on top, which only underscores the unlikely nature of his Cy Young season. Still, this list means nothing in practical terms and is too obscure even to rise to the level of trivia. Yet when I look at it, it means something to me and helps me understand why I care about a non-essential counting stat like pitcher wins. Each of those pitchers is a story, and each year listed in the table represents an era. Lefty Grove? Well, he was truly dominant. So was Walter Johnson. John Burkett? Jack Chesbro, but not his 1904 season? Ah, old Dizzy, the last of his kind. And Steve Stone, the forerunner to Welch when it comes to unlikely seasons. I can stare at a list like that and lose an hour, easy.
And what is the portal for this daydreaming? Wins. Why? Because they've always been there. And really, you could say the same thing about other Paleozoic stats like batting average and RBI. Be careful how you use them, but don't get rid of them. They may have only metaphoric value, but that's worth something.
"Pitcher wins ain't what they used to be thanks to the rising offensive levels, deeper lineups, longer at-bats, and increased reliever specialization which have made the complete game a relic from the increasingly distant past. … Against this backdrop, the win has come to be understood less as the product of an individual pitcher's brilliance or intestinal fortitude on a given day, and more as the confluence of the right amounts of support from the offense, the defense, and the bullpen."
Jaffe's points are true enough, but nobody was in a mind to cast aspersions about Verlander's win streak when the Tigers paid their last visit of the season to Chicago in mid-September. Detroit was on a 10-game winning streak and sent its ace to the hill to blow an even bigger hole in a division race that never was to materialize.
The Tigers' clubhouse was understandably in high spirits. When a team is in contention in September, the music is a little louder, the jokes are a little bolder, and the card tables are a little more crowded. It's fun to be around, especially because you know the players are going to be that much more apt to chat with the media. Detroit's general enthusiasm extended even to Leyland, who seems to relish his crusty old baseball guy image. He popped out of his smoky office—fully clothed—and said hello to the media. The clothes observation might seems like an odd detail, but you have to keep in mind that you never know what state you're going to find Leyland in when he conducts his pregame chats. He just flat out doesn't care.
The fact that he was still wearing a sports coat and a shiny pair of shoes was a good sign. We knew he'd hold his presser soon, because he likes to get it over with soon after the clubhouse opens. Most visiting managers wait until the club goes onto the field to stretch, so Leyland in effect saves his media an interminable two-hour wait on a daily basis. Sure enough, a couple of minutes after Leyland returned to his office, the Tigers' media guy waived us in.
Leyland was tilted back in his chair, feet on desk, smoking a cigarette. His sports coat and dress shirt were gone, leaving only the wife beater he had on underneath. He'd removed his shoes and socks, so we had to place our recording devices on his desk at what was hopefully a safe distance from his feet, but still close enough to pick up his words. In his ashtray was the stub of an old cigar Leyland had been carting around from city to city. He'd started smoking it the day the win streak began, and Leyland doesn't fool with streaks. We didn't find out until the next day that he was also wearing the same crusty pair of unwashed underwear that he'd been wearing at the outset of the surge.
As you'd expect, Verlander was the main topic of conversation, which was great, because that's what I was interested in. I tried to focus but kept finding myself looking at Leyland's toenails. They needed a trim.
Verlander entered the game with the longest streak of winning starts since Prince Hal Newhouser won 11 straight in 1946, a season in which he finished with 26 wins. Newhouser completed all 11 of those starts, one of them a 10-inning affair, giving him 100 innings pitched as a starter during that stretch. (He also pitched once out of the bullpen.) So here's the part where you're tempted to write that Newshouser pitched when men were men and sheep were sheep, but you can't write that without writing this: Newhouser was finished as a frontline pitcher at the age of 29. Denny McLain's usefulness was extinguished after he started 82 games in 1968 and 1969. He was out of baseball at 29.
Verlander is 28, and despite his heavy workload in relation to the rest of the league, no one would doubt that he's got many years as a productive pitcher ahead of him. Like Doc Halladay, he's passed that threshold. The lighter load he had to carry in his youth may have given us many more years to enjoy his pitching. Maybe it wouldn't have mattered. We'll never know. These musings led me to ask Leyland about how he knew when Verlander had reached the point at which his innings total could really be pushed.
"I tried to be awful careful," Leyland said. "I'm not the smartest guy, but I could see that there was something special there. He gets mad at me sometimes because I take him out.
"I don't consider 110 pitches a horse. Justin Verlander's a horse because his 110th pitch is 101 miles per hour. That's a fuckin' horse. My own personal opinion (about pitch counts for younger pitchers)—and I don't stick my nose in and get involved—is that's one of the problems with the minor leagues. They don't let them throw pitches. If you look at the innings of Greg Maddux and all those good pitchers, they pitched innings in the minor leagues. I don't agree with it, but there is so much at stake with the big contracts and the agents. Let's face it, you're a little bit gun shy as a manager."
He didn't have to worry about riding Verlander too hard that day. Detroit broke open a tight game with on a three-run Victor Martinez homer in the sixth. Verlander departed after seven fairly economical innings, and the Tigers won 5-0.
"The last few starts, I've been just a hair off," Verlander improbably said after his 23rd win. "I really felt like tonight I was able to get back in a rhythm out there."
Ummm … didn't you just win your 11th start in a row?
"Just because the results work out doesn't mean you don't have something to work on."
Verlander threw eight shutout innings in winning No. 24 in his next start at Oakland before dropping his last outing of the season. That cost him the best 25-win season in the majors since Ron Guidry's incredible 1978 season. And, yes, I was angry when Verlander lost.
"It's definitely an accomplishment. It's difficult to win 20 games in any era. It'll be nice to see him go out there today and get it."
Carpenter gave me that quote before Wainwright got his 20th win in what turned out to be his last start until the day he makes his 2012 debut. Carpenter can be a little persnickety when it comes to dealing with reporters he doesn't know. He gave me perfunctory answers to questions I asked him about the wind having an impact on strikeout totals at Wrigley Field and whether he adjusts his game plan based on the conditions. But when I asked him about Wainwright, he brightened up. That text is the full transcript of his response, but for Carpenter, that's positively effusive.
Though the Cardinals were on the verge of mathematical elimination from the NL Central race that September, the mood was decidedly upbeat after Wainwright threw six good innings, hit an RBI double, and won No. 20.
"He got what he deserved," Tony La Russa said. "He's had such a great year. What a performance. The wind's blowing out at 20 mph. That's some serious pitching. Those of us that have watched him over the past two or three years, there isn't anybody in the big leagues better than he is. He's right there with the best starters that you want to talk about. You put Adam Wainwright's name on that list."
Wainwright seemed almost anxious to talk afterwards—for a Cardinal—though he did ask that we adjourn to the laundry room out of deference for his teammates. The year before, he finished with 19 wins after Kyle McClellan blew a 6-1 lead in Wainwright's last start of 2009. McClellan was one of the pitchers who helped preserve that day, which pleased Wainwright. He also enjoyed reaching the milestone against the Cubs.
"It's always fun to pitch at Wrigley, but to beat the Cubs for your 20th win is an extra treat," Wainwright said. "My team gave me a beer shower after the game. The team continues to impress me, to make me feel wanted. They've made me feel at home here. There's not a better bunch of guys to share this with."
It was just a number, and one that we all know shouldn't mean that much. Yet, that day it clearly did—to Wainwright, to his teammates, and to his manager.
I wish wins were a better stat. That would make things easier, wouldn't it? But you can't change the definition now, because without historical context, the stat means even less than it already does. And we don't want to re-write all those baseball books and encyclopedias. The baseball lexicon can change over time. The most obvious example I can think of is OPS. It's surprising from whom you hear that acronym slip these days. Which is nice, even as most members of the analytical world have sort of set it aside. Perhaps 100 years from now, everyone from the statheads to the old crusties will have a win-based stat to cite for pitchers that actually means something, something that catches on with the general public.
Until then, I'll be rooting for somebody, preferably a Kansas City Royal, to somehow win 30 of his 33 starts this season. That would be awesome.