Great pitchers (Justin Verlander, Ron Guidry) really are the most valuable players

As strong as my pro-Yankee bias is, I think Justin Verlander should be the American League Most Valuable Player this year.

Yankees Curtis Granderson and Robinson Canó had outstanding years that merit MVP consideration. So did some non-Yankee position players: José Bautista, Miguel Cabrera, Jacoby Ellsbury and Adrián González.

When baseball writers (MVP voters) discuss the MVP contenders, you hear one of the dumbest statements and one of the strongest biases in baseball, almost as strong as the anti-Yankee bias: Pitchers shouldn’t be considered for the MVP.

That notion — and the fact that it persists so strongly — reveals why baseball writers as a group are too stupid and too biased to decide anything meaningful.

The dim bulbs making this argument generally cite two points: Pitchers have their own award and pitchers aren’t as valuable as every-day players. If the first point really matters, let’s just rename the award Most Valuable Position Player. Batters have their own awards, Silver Slugger and the Hank Aaron Award. MVP should be for the most valuable player of any position.

The second point just reveals the writers’ deep ignorance of the game they cover. Everyone in baseball knows that great pitching is the most dominant force in baseball. So on that level, it’s easy to argue. It’s easy to see the impact a pitcher has: When a great pitcher (or lousy pitcher) is on the mound, the team’s performance varies much further from its normal performance than when any position player is in the game. Verlander’s winning percentage this year is .828, about 200 points higher than any everyday player’s winning percentage.

Yes, it’s hard to compare pitchers to hitters, partly because what they do is so different and partly because of the different impact. A starting pitcher impacts only every fifth game, but he has a huge impact on it. A position player plays every day but only hits four or five times a game. To say that the award should go to someone who plays every day is as ridiculous as it is to say that the award should to to someone who has a huge impact on dozens of plays every game he plays.

Pitching and hitting are equal forces dominating baseball (each much bigger than baserunning and fielding, the other components of offense and defense). It’s crazy to limit consideration of the game’s most valuable players to only the players doing one of the game’s two dominant pursuits.

It’s not an overstatement to say that the pitcher and hitter are the dueling lead players in nearly every baseball play. Base runners and fielders occasionally play a leading role, but usually play a secondary role to the pitcher and hitter.

So let’s look at how many plays Verlander has played one of these leading roles, as compared to each of the other MVP candidates. Verlander faced 969 batters this year. Jacoby Ellsbury had the most plate appearances of the MVP contenders, with 729. Verlander pitched more times than anyone in the A.L. hit. We should be asking whether they played enough to be considered against him.

I’m no sabermetrician. But I think I can show how mathematically that a pitcher has as much impact on his team’s season as a position player who plays every day.

Let’s start with an obvious premise: The outcome of every play rests 50 percent with the offensive team and 50 percent with the defensive team. (There are exceptions, such as balks or wild pitches, but baserunners certainly influence balks, and the hitter may influence some wild pitches.) Each play during a baseball game will count for one play share. So the defense and offense each get 0.5 play shares for each play (even balks and wild pitches, just to be consistent). A play is a plate appearance or anything else that results in an out or a runner advancing.

Obviously, the 50 percent offense at times rests entirely with the hitter (homer, strikeout, walk, no one on base). Other times, the 50 percent offense rests entirely with the baserunner (stolen bases or caught stealing, ignoring that perhaps a busted hit-and-run might cause a caught stealing). Other times, the offensive share is split between the batter and the baserunners (batter contributes most to the outcome by hitting a single, for instance, but the runner makes a notable contribution by scoring from second base). Similarly, a pitcher takes the full defensive credit for the homer or strikeout (yes, the catcher called the pitch, but we’re giving the pitcher full credit here; other aspects of the formula will favor position players). And when the ball is in play, the share is split between the pitcher and the fielders.

I won’t bog this piece down further her by explaining the formula. Go to the end of the piece for that. But here’s the bottom line: Verlander has more play shares than any of the MVP contenders this year: 430.9 to 418.62 for González (the formula gives a slight advantage to first basemen over other position players). In fact, Verlander had more than 100 play shares more than Bautista.

Of course, you can slant the discussion quickly toward the pitcher if you start counting successful plays versus unsuccessful, because baseball isn’t a 50-50 proposition. Pitchers get outs more often than batters get hits, or even get on base.

But dig a little deeper into the numbers. Verlander struck out 250 batters this year. On that many plays, he was completely successful, beating the batter as badly as a pitcher can by himself (a double play is a bigger success, but shared with the fielders). Because of baseball’s imbalance favoring pitchers, it would be unfair to compare this to the homers (completely successful plays for the batters), though that’s a rout for Verlander, 250 to 43 or less (Bautista led the A.L. in homers). So instead compare his complete successes with the hit totals of the batters, representing a range of success beyond walks and sacrifices: Verlander’s strikeout total far surpasses even the largest hit total of his competitors: 213 by González.

Flip it around: Verlander had 24 plays that were complete failures: giving up a home run. By contrast, each hitter had at least 89 plays that were complete failures: striking out (for that matter, you could add the double plays they grounded into).

If you want to go sabermetric here, Verlander had 8.4 “wins above replacement” (a formula to project how many more games his team won that it would have with a replacement-caliber player from the bench or the minor leagues). Only Bautista (8.5) topped that.

With no dominant hitter — you could make a case for any of six every-day players — it’s really an easy call for Verlander. He has been involved in more meaningful plays for his team than the hitters. And no one, pitcher or hitter, has been anywhere near as dominant.

Interestingly, the last starting pitcher to win the MVP, Roger Clemens in 1986, was actually dead even with the best hitter that year in play shares, rather than having a solid advantage. So why did Clemens win the MVP? Because that year put the most powerful bias of the baseball writers into play: their anti-Yankee bias.

Don Mattingly‘s 1986 season would blow away this year’s offensive contenders. He led the league in hits, doubles, total bases, slugging and OPS. Though he wasn’t known as a homerun hitter,

To see clearly how the anti-Yankee bias works, compare the Clemens-Mattingly race to Ron Guidry and Jim Rice in 1978. When it comes to play shares, Guidry blew Rice away, 467 to 394. Guidry pitched nine shutouts, a level of dominance an offensive player can’t achieve (no hitter can ensure that his team won’t lose in nine innings). Rice had a great season that year, with 213 hits, but that didn’t approach Guidry’s strikeout level. Guidry was completely unsuccessful — giving up a homer — in only 13 plays. (Rice grounded into 15 double plays, if you want a similar measure of failure.) On the other hand, Rice was completely unsuccessful — striking out — 126 times.

Rice and Mattingly had roughly comparable years. Each led his league in hits, total bases, slugging and OPS. Mattingly also led in doubles while Rice led in homers, RBI and triples. Mattingly had a notably higher batting average, .352 to .315, a higher on-base percentage and way fewer strikeouts. They both were lousy at stealing bases, but at least Mattingly never tried (Rice was caught five times). Both scored a lot of runs (117 for Mattingly, 121 for Rice). Each was easily the best offensive player in the American League, Rice with the power edge and Mattingly with the edge in contact. Mattingly was far superior on defense, winning the second of his nine Gold Gloves at first base. Rice was a mediocre outfielder who played designated hitter for 49 games (Mattingly DH’d once). Mattingly was so solid defensively that he played three games at third base, even though he was left-handed. If you discount defense entirely, you can make a case that Rice had the better offensive year, but they were pretty even.

Though Clemens had a dominant year in 1986, it fell short of Guidry’s 1978 in every respect. Each led his league in wins, winning percentage and ERA, though Guidry’s numbers were better in every respect (three-quarters of a run better in ERA). Each was second in the league in strikeouts, but Guidry had 10 more. Guidry was way better in shutouts, with a league-leading nine to one for Clemens.

Guidry had the advantage over both Rice and Clemens with 8.5 wins above replacement, compared to 7.0 for Rice and 7.9 for Clemens (Mattingly had 6.9, confirming my assessment of him as pretty even with Rice).

The comparison is pretty clear: Each time a dominant pitcher was clearly more valuable than the dominant hitter. Usually, the anti-pitcher bias swings that in favor of the hitter. But when the hitter was a Yankee, that was enough to seal the MVP for the pitcher. Even when the pitcher had a bigger advantage over the hitter, both in dominance and in actual meaningful plays, when that pitcher was a Yankee, he had to settle for the Cy Young Award.

Unlike the baseball writers who choose award winners, I recognize my bias. I have a heavy Yankee bias. But not heavy enough to say either Granderson or Canó deserves to be the MVP this year. Verlander should be the runaway winner (like Guidry should have been in 1978).

Below the videos, I explain the play-shares formula.

Play shares explained: I’ve devised a formula to reflect a player’s impact on the outcome of his team’s plays. This is a measure of quantity, not quality: just an effort to count the number of plays in which the player made a significant contribution to the outcome, not whether that outcome was positive or negative for his team.

We’ll recognize that this formula does not fit every play perfectly. For instance, it will not recognize the batter’s contribution to a caught stealing that was a busted hit-and-run. It won’t recognize the fielder’s impact on an out that the centerfielder makes by reaching over the fence to catch a potential homer. But I think it reflects on average the contributions that players make to a team’s success or failure.

For homers, strikeouts, walks and hit batters, pitcher will get .5 play shares and batter .5. Any contributions by baserunners and fielders are negligible. For stolen bases and caught stealing, the runner gets .5 play shares. With no runners on base, the batter also would receive the full .5 play share for the outcome of the play. With any runners on base, the batter still gets most of the credit, .4 play share. Runners would get .1 play share. I’m using year-end statistics, not daily score sheets, so I am trying to simplify here. I will give the batters .45 play share for plate appearances that result in balls in play. This splits the difference. They also will get .1 play shares for every time they are on base. This will give them probably more credit than they deserve. If they reach base with two outs and the next batter strikes out or pops up, they really aren’t involved in the play. If they are on base when someone hits a homer, they aren’t really involved. But other times, they may reach base and be involved in a couple of plays, advancing quickly on a bunt, so the defense can’t go after the lead runner, then scoring on a single. I think .1 for every hit (excluding homers), walk and HBP is generous.

My formula gives the pitcher .4 credit for balls in play, with the other .1 going to the defense. However, defensive statistics make it difficult to apportion that .1. If a first baseman makes a spectacular catch of a line drive, he should get the full .1, but it’s just a putout in the stats. On the other hand, if the shortstop makes a great grab in the hole and throws a strike to the first baseman, he still gets a putout, even though he probably deserves only .01 of the credit. That’s OK, though, I want to reflect the pro-position-player bias of the baseball writers, so I am going to give players .07 for every defensive chance — putout, assist or error, more than splitting the difference. This gives first basemen a slight advantage because of the large number of putouts they get.

Baseball’s conventional wisdom says you steal bases on the pitcher, not the catcher, so I’ll credit pitchers with .3 play share for every stolen base or caught stealing. A pick-off belongs more to the pitcher, though the fielder still has to put on the tag. So I’ll give .4 play share for pickoffs.

Wild pitches and balks probably aren’t a 50-50 proposition, but I think the formula has to keep each play equally divided between offense and defense. I don’t have offensive stats for individual runners advancing on balks or wild pitches, but I have already probably over-credited offensive players for the plays that happen when they are on base. Wild pitches and balks count as .5 play shares for pitchers.

For play shares and other stats on the players mentioned in this blog post, check my spreadsheet (all stats from Baseball Reference).