All alone, as just another statistic to consider when evaluating pitchers, there’s nothing really wrong with crediting pitchers with wins.
It certainly isn’t what it used to be, when pitchers were expected to finish their starts, and batted for themselves, and there was far more reason to claim that the pitcher, all by himself, won the ballgame. Even with his teammates providing most of the offense and fielding most of the batted balls, you could see where the pitcher controlled the action and therefore controlled the game.
That particular utility of the pitcher win statistic hasn’t existed for a long, long time, but that doesn’t mean it’s an irrelevant statistic. In modern baseball, we immediately know several things about most pitchers with a lot of wins.
They were healthy.
They were trusted by their manager (or had a terrible bullpen, I suppose).
They went deeper into games than most starting pitchers.
They had to be relatively effective, otherwise they would have been pulled from the game before they could get the win.
Those are important things to know about a pitcher, so there’s value in the stat. It’s just not the value it once was, back in the days when a starting pitcher just put in a lot more work. It’s inarguably a less important stat than it used to be.
That said, its elevated importance in the minds of fans, writers, players and even talent evaluators lingered for much longer than it should have. In some circles, that feeling lingers even now, and it was certainly still firmly in place in 1990, when most voters for the American League Cy Young Award essentially said, “Screw it, let’s just give it to the guy with the most wins.”
That was the year the Bob Welch of the A’s had the most fortunate year of his career. Note that I didn’t say it was the best year of his career, because it wasn’t. In fact, it wasn’t particularly close to being Welch’s best year. That was probably 1987, his final season with the Dodgers. Compare that season with his stats in 1990.
1987: 35 starts, 251.2 IP, 196 Ks, 21 HR allowed, 3.47 FIP, 1.152 WHIP, 7.1 WAR
1990: 35 starts, 238.0 IP, 127 Ks, 26 HR allowed, 4.19 FIP, 1.223 WHIP, 2.9 WAR
Welch’s 7.1 WAR led all pitchers in 1987, and he tied for the league lead with four shutouts. He finished just eighth in Cy Young voting that year, though, a victim of the fact that he didn’t lead the league in any of the Triple Crown stats and had a record of “only” 15-9 despite his team being dead last in the league in scoring runs. So that was pretty unfortunate for him.
That luck changed in 1990. Now on the A’s, Welch got to have the third-best offense in the league on his side. Oakland averaged 4.52 runs per game in 1990, and, even more fortunate for Welch, averaged 5.21 runs in games he started. So while Welch didn’t pitch as well as he did in 1987, he pitched well enough, in front of a good enough offense, that the A’s went 29-6 in his starts.
Welch was credited with 27 of those victories, the second-highest win total for an American League pitcher since World War II, topped only by Denny McLain’s legendary 31 wins in 1968. The Cy Young voters couldn’t resist that gaudy total, and gave Welch the award.
And it was absolutely the wrong decision.
Without knowing anything else about them, which of these American League pitchers had a better season in 1990?
Welch: 238.0 IP, 125 ERA+, 1.223 WHIP, 4.8 K/9IP, 1.54 K/BB, 4.19 FIP, 2.9 WAR
Pitcher B: 236.0 IP, 158 ERA+, 1.233 WHIP, 6.8 K/9IP, 2.19 K/BB, 3.30 FIP, 5.5 WAR
Well, it’s pretty obvious that Pitcher B was better, right? He struck out more hitters, and prevented runs better, and was less reliant on his defense.
If that point isn’t sufficiently clear, let me add that he got 4.63 runs per game of support compared to 5.21 for Welch, and he played for a team that was 22-10 when he started (.688) and just 58-72 (.446) when he didn’t start. That’s a gap in winning percentage of 242 points. For reference, the A’s were 29-6 (.828) when Welch started and 74-53 (.583) when he didn’t, and nearly identical gap of 245 points.
Pitcher B was Chuck Finley of the Angels, when went “only” 18-9 that season and therefore finished just seventh in Cy Young voting despite the fact that he:
A) Pitched better than Bob Welch.
B) Got less run support.
C) Played for a worse team.
D) Lifted their winning percentage by the same amount as Welch when he pitched despite B) and C).
But Finley didn’t deserve the Cy Young that year, either. That honor goes to Roger Clemens, who was better than Welch in every way.
Clemens: 1.93 ERA, 211 ERA+, 2.18 FIP, 8.2 K/9IP, 3.87 K/BB, 10.4 WAR
Welch: 2.95 ERA, 125 ERA+, 4.19 FIP, 4.8 K/9IP, 1.65 K/BB, 2.9 WAR
The Red Sox were 22-9 (.710) when Clemens started and 66-65 (.504) when he didn’t, meaning that Clemens, like Welch and Finley, raised his team’s winning percentage over 200 points when he was on the mound. Only he did it with less run support than the others. The Red Sox scored just 4.34 runs when Clemens pitched, nearly a run per game less than the support Welch received.
Eight voters were savvy enough to recognize that Clemens’ 21-6 record was achieved under much more difficult circumstances than Welch’s 27-6 record, and that he otherwise out-pitched Welch in every measurable way. Those eight gave Clemens their first-place votes for the Cy Young. Voters for the MVP Award were even more savvy, because Clemens finished third in MVP voting that year, and got the most MVP votes among all American League pitchers. Welch was ninth in MVP voting, receiving fewer votes than not only Clemens, but also Bobby Thigpen, Dennis Eckersely, and Dave Stewart among pitchers.
That’s right…two fellow members of the A’s pitching staff received more MVP support than Bob Welch, and yet he won the Cy Young Award anyway. Had the voters decided to award his teammate Dave Stewart instead, that would have been perfectly justified.
Welch: 2.95 ERA, 125 ERA+, 4.19 FIP, 4.8 K/9IP, 1.65 K/BB, 2,.9 WAR
Stewart: 2.56 ERA, 144 ERA+, 3.33 FIP, 5.6 K/9IP, 2.00 K/BB, 5.2 WAR
Despite playing for the same team, Stewart got about a half-run less in run support each game, 4.72 to Welch’s 5.21. But the A’s were “only” 25-11 in his starts, and his own record was “only” 22-11, so the voters overlooked the reasons for that and gave Welch the award. Stewart finished third, getting just three first-place votes.
If these are my four candidates for an award that is supposed to go to the best pitcher of the year, I know who I would be voting for:
More importantly, I know who I would not be voting for. In 1990, voters were only allowed to vote for three pitchers. Under that system, Welch wouldn’t have been on my ballot.
Lucky for him, that one number, his win total, was all most voters looked at it those years. That cost him a Cy Young he might have deserved in 1987, but the voters sure made up for it a few years later.
Run support and bullpen performance are so critical to wins and losses for a starter, that it’s hard to fathom the “best” pitcher is the one with the most wins.
Clemens seems like the right choice when you’ve got the data. If you’re effective (FIP and WHIP) for more innings than the average starter, how could you not be the right guy?
Great stuff! This could easily be part of a series showing how the best player isn’t always awarded, or how even a player's best season is t where he gets the accolades.
I think I would have voted for Clemens but not certain of it. Welch would be my 2 at worst. 27 wins and an ERA under 3 on a team headed to playoffs would be hard to not support. I think if it happened this season, even with the attention advanced metrics get now, he would still win. It might be closer but 27 wins would be such a big deal that I think he would win