, ,

Bob Gibson set the all-time record with a 1.12 ERA in 1968 in over 300 IP. But he didn't lead the league in Wins.

Bob Gibson set the all-time record with a 1.12 ERA in 1968 in over 300 IP. But he didn’t lead the league in Wins.

I will get right to the point:

The statistics assigned to pitchers called “Wins” and “Losses” are nearly worthless for evaluation. Not because actual wins and losses are not valuable (they are the ENTIRE point of the game), but because of the way we define it for the box score and how we assign it to pitchers. In other words, Wins and Losses would make a good stat, if they actually reflected a pitcher’s performance. But they don’t. So called “Wins and Losses”, as a measuring statistic, fail because there are too many factors outside of a pitcher’s control, there are too many arbitrary parameters, player to player comparison using the stat only works in the broadest terms, and arguments like “pitching to the score” and “providing what their team needs” are good only in poorly thought out abstractions.

Factors Outside of a Pitcher’s Control, and Arbitrary Parameters

It would be very nice and convenient if we had an awesome and accurate way to sum up a pitcher’s performance by giving him a Win or a Loss. However, the actual process of defining what earns a Win or a Loss is much more difficult. Here are the biggest, most obvious shortcomings of Wins and Losses (on which I think most of us will agree):

  • A pitcher cannot control how many runs his own team scores, even in the NL (where a starting pitcher usually only gets 2-3 plate appearances per game).
  • A pitcher that allows one run through 9 innings might get a Loss, whereas a pitcher who lets off seven runs in just 5 innings can get a Win.
  • A starting pitcher might throw 8 shutout innings, but if his team also didn’t score until the 9th, the relief pitcher who comes in to to pitch just that one inning is awarded the Win.
  • A starting pitcher might have 5 games in a row where he pitches 7-8 innings and let’s off no more than one run per game, but never get a Win because the team’s relief pitchers keep blowing the lead.
  • A starting pitcher must pitch 5 complete innings to earn a Win, while a relief pitcher might get a win (or not) based on this arbitrary rule:

“the official scorer shall credit as the winning pitcher the relief pitcher, if there is only one relief pitcher, or the relief pitcher     who, in the official scorer’’s judgment was the most effective, if there is more than one relief pitcher.” (from MLB official rules, “10.17 Winning And Losing Pitcher

and it gets stranger:

Rule 10.17(c) Comment: The official scorer generally should, but is not required to, consider the appearance of a relief pitcher to be ineffective and brief if such relief pitcher pitches less than one inning and allows two or more earned runs to score (even if such runs are charged to a previous pitcher).

Sounds like someone just made it up, and everybody thought it sounded good at the time, right?

In 1992, Jack Morris had a 4.04 ERA, but he lead the league in Wins.

In 1992, Jack Morris had a 4.04 ERA (league average for that season), but he lead the league in Wins.

So what is the point of giving out these Wins and Losses, if there are so many significant (and common) ways to skew the results? Well, I will grant this much:

  • It is highly unlikely that a really bad pitcher would qualify to Win 20 games in season.
  • It is pretty unlikely that a great pitcher would Win fewer than 10 games in a season.
  • Most of the time, pitchers who consistently get 14-18 Wins per season are at least good (if not necessarily very good).
  • There are no “average” pitchers with 300 or more career Wins.

Are Broad Generalities Enough?

Are these broad generalities enough to help us figure out who has pitched better when we are discussing Cy Young Award candidates, or “Big Game” pitchers, or who the ace of a specific team might be? Does an 18-game Winner definitely beat out a 12-game Winner? How about a 24-game Winner over a 19-game Winner? You can probably find many legitimate differences (like ERA, strikeouts, walk rate, WAR) showing that a 20-game Winner is better than an 8-game Winner. But, on average, once you get into the category of guys with 13+ Wins, the numbers tend to even out, and the differences in “Wins and Losses” come more from those shortcomings we mentioned above (the ones the pitcher cannot control).

For example: What if there is a pitcher with an 18-7 record and a guy with a 16-9 record, both up for Cy Young Award discussion. They have pitched the same number of innings over 35 starts. The guy with the 16 Wins has a lower ERA, more strikeouts and fewer walks. His team scores fewer runs than league average, and his bullpen is awful. Who should win the Award as the best pitcher? Since ERA, strikeouts and walks are fantastic indicators of a pitcher’s value, I think many of us would vote for the 16-game Winner.  What is 2 Wins, anyway?

2 Wins difference actually helps prove my point:

  • If each pitcher had 35 starts, then they each had 10 no-decisions.
  • If we agree that the pitcher with the lower ERA, more strikeouts and fewer walks (in the same #IP) is better, then we are disregarding those 2 Wins that separate him from the other guy.
  • This means that 12 of their 35 starts (10 no-decisions, 2 Wins we decided weren’t enough of a difference) don’t “count” towards deciding who is better, if we go by Wins and Losses. That’s 34% (more than 1/3) of the season.
Felix Hernandez won the 2010 AL Cy Young Award, despite  a 13-12 record for the last place Mariners.

Felix Hernandez won the 2010 AL Cy Young Award, despite a 13-12 record (and 9 no-decisions) for the last place Mariners.

Those “Other” Games

What in the world are “No Decisions” about, anyway? The performances in these “erased” games only show up in a pitcher’s other stats. If a pitcher gets 10 no-decisions, that could add up to 60-80 strikeouts, and totally change his ERA (for better or for worse). A pitcher with 200 strikeouts and a 2.50 ERA is significantly different in our eyes than a guy with 160 strikeouts and an ERA of 3.00. Those 10 no-decisions could create that difference.

And what if the team ended up winning most of those no-decision games? What if in 5 of those no-decisions, our pitcher let off 3 or fewer runs, and his team came back to win them all? Shouldn’t the pitcher get credit for keeping the other team to 3 runs or less, thus giving his team a chance to come back in the later innings? No. Instead, we discount it in the Wins-Losses stat.

The Issue of  “Pitching to the Score”

This concept is about as legit as a promise from ARod. The idea that a pitcher would allow or not allow base runners and scoring based on what his team needs is absurd. This is an argument I have heard from some who say that Wins are a better way to judge a pitcher than ERA. Jack Morris recently became the poster boy for this argument. “He won 18 games. So what if his ERA was 4.28? When his team would go up by 6-7 runs, he would start pitching to the score (meaning not trying to strike guys out, but letting them hit it to allow his fielders to do their job)”. Here are the reasons why this angle holds absolutely zero water:

  • In how many innings per season would this situation present itself? A lead of anything less than 10 runs before the end of the 5th inning is not a safe enough lead to intentionally allow guys to hit the ball. So we’re talking only innings 6,7,8 and 9 of games where his team has a “safe” lead of what? 6-8 runs? How many times does that happen in a season?
  • Why would a pitcher ever not try to get batters out? Even with a big lead, if a pitcher has the skill to get guys out, why would he not? The game goes by quicker when you get guys out. Sure, striking them out takes longer than getting them to ground out. But if we’re talking about pitching that affects ERA, then we’re at least talking about singles, if not extra base hits, right? If the other team is scoring, the game goes by less quickly. If you’re really good, just get them out and go home.
  • Perhaps Jack Morris was an exception, but how many pitchers out there don’t care about their own ERAEverybody wants to win first and foremost, sure, but if your team is up 8 runs in the 7th inning, why would a guy not also try to keep his own ERA down? 
  • Does this kind of thing work late in the season? In other words, after 200+ innings, in late August in 90 degree heat, why would a pitcher want to throw more pitches? Again, the reasoning for Wins over ERA is that “pitching to the score” adversely affects ERA. This means guys are getting on base and scoring. That means more time on the mound and more pitches thrown. Does a pitcher or manager want that going on late in the season? Iff not, then we have to restrict the occurrences of this event to “when your team is up by at least 6 runs in the 6th inning or later in a game in April, May, June or July”. I say again: How often does that happen?
  • If a pitcher (like Mr. Morris) pitches 200-250 innings during a year, and ends up with a high (or at least unimpressive) ERA, then how many of these “pitching to the score” innings would he had to have pitched to bring his ERA that high (in the range of non-Cy Young Award and non-Hall of Fame)? During Jack Morris’s three 20-Win seasons, he averaged 267 IP and allowed an average of 104 Earned Runs. His ERA’s during those seasons: 3.34 (10th in the AL), 3.27 (6th in the AL), 4.04 (league average). So he had the most Wins, but was not among the top 5 run preventers. The three years combined average an ERA of 3.53.  So let’s see: if Morris had the opportunity to “pitch to the score” 5 times per season, then what would we have to expect those innings to have looked like in order to keep his ERA from getting into the “golden zone” of sub-3.00 ERA’s? In order for Morris to have kept his 3-year average (of his “best” three years, when he won 20+), above 2.99 (to its actual 3.53), then he would to have intentionally allowed 3 runs in the final innings of a game five times in each of those years. And when you look at it like that, then we have to wonder: what kind of lead do you have to have for that to be ok? I had said 6 runs or more, but allowing 3 runs in that instance means you are cutting your lead to just 3. Even in the 9th inning, who would chance that? No, I think to safely allow 3 runs, you’d have to have an 8 run lead. And in that case… How many times can that situation possibly have come up?
  • The final knock to the “Pitching to the Score” supporters is this: If it is a legitimate tactic, and Wins (the stat) are more important than ERA, then why did all-time greats like Sandy Koufax, Pedro Martinez, Greg Maddux, Bob Gibson, Christy Mathewson (and a whole bunch of others) end up with excellent-to-amazing career ERA’s? Why did they bother? All of those guys played on championship teams that probably had large leads in the late innings. And yet they all had seasons with an ERA of under 2.00.

In 2011, Derek Holland had a record of 16-5. His ERA was 3.95, but he received the highest run support (7.64 runs per game) of any pitcher in baseball that season.

“Providing What the Team Needs”

This is a close cousin to the Pitching to the Score idea. Some people say that a good pitcher will provide what a team needs that day, hence earning the Win or the Loss. If a team is facing someone like Felix Hernandez, then a good pitcher would be able to keep Felix’s team’s hitters to only a couple of runs, knowing the score will be low. A good pitcher will bounce back after a rough first inning and keep his team in the game by not allowing any more runs. 

All of those things above are good ideas, and have a kernel of truth to them. But the problem with it is similar to the ones I listed in the last section: Why would a pitcher ever not pitch at his best? Furthermore, and maybe even more importantly: At what point does a pitcher know what kind of game his own teammates will have? In other words, at what point would a pitcher realize that “Uh oh, I’ve been pitching with 60% effort, but now I realize the other guy is good today. Now I have to try harder!” Does that happen after the 3rd inning, when his team hasn’t scored yet? The 5th?

I will definitely acknowledge that some pitchers respond better to pressure than others. Some great pitchers can continue to pitch at ace level when the heat is on, while others seem to lose their focus. But that usually has far more to do with the game environment (late season, playoffs, etc.) than whether his own team is scoring or not. So a pitcher will pitch as best he can, no matter what, and if his hitters and bullpen can do their part, then the team will win. 

When Bob Welch won 27 games in 1990, people argued that he “kept his team in ballgames, giving them a chance to win”. This was apparently at least moderately true. But if the argument is that a good pitcher gives his team a chance to win, wouldn’t ERA be a better measure? If a player has an ERA of 2.30, then you know he usually gives his team a chance to win even if they aren’t able to score a whole lot. If a pitcher has a lot of Wins and a high ERA, then you know he pitched well enough for his team to win when it scored a lot of runs. The first description is the pitcher of actual quality.


If not Wins and Losses, then what? I’m not sure we need to assign credit or blame for a entire team’s game results to one player. If  any position would most deserve that role, it would be the pitcher, but there is no way to justify even that. Wins and Losses should be relegated to the Team as a whole, and individual players should be evaluated by their independent performances. In the case of pitchers, the best stats to do this are FIP, xFIP, ERA, K%, BB%, Swinging Strike %, WAR and RA9 WAR. By comparing those 8 statistics, you should have a pretty darn good idea of which pitchers actually performed best.