This article reports on a very short piece of research; in fact, I am sure it will take me much longer to write the article than it did to do the research.  (it took me about six times as long.) This has to do with a "Hey, Bill" exchange two or three days ago in regard to Roger Maris:

There are two variables in MVP seasons, and a lot of people don't get this.  There is the variable of whether a player actually deserved the MVP Award or the Cy Young Award, as for example with Justin Morneau in 2006, but there is also the issue of whether it was a good crop of candidates or a weak crop of candidates.   Some guys win an MVP Award, like Dick Groat in 1960 or Elston Howard in 1963, because there just really isn't anybody there having an MVP season.   It would be pretty easy to do a ranking of the strongest and weakest MVP candidate crops of all time. . .I probably should do that.  

This is that research; the statement about Dick Groat may not turn out in a few pages to be absolutely accurate.  It kind of seems to me like I undertook this project once before, although I’m not sure whether I did or didn’t.  Doesn’t matter; I go back to the same issues over and over because I figure out how to improve the way I studied it before.   This is the system I set up to measure the strength of a year’s MVP candidates.

First, I ignore anyone with 20 Win Shares or less, since a player with 20 Win Shares or less would not ordinarily be a serious MVP candidate, and a player with 21-25 Win Shares would not ordinarily be a strong MVP candidate.   For players above 20 Win Shares, I:

1)     Subtracted 20 from their Win Shares, and

2)     Squared that, and

3)     Added up the league total. 

Just for a little context, in 2019 there were six players who had exactly 20 Win Shares.

Jose Abreu, first baseman, led the American League in RBI with 123, but hit .284 with a bad strikeout/walk ratio (152-36), only two stolen bases and led the league in grounding into double plays with 24.   That was 20 Win Shares. 

Yuli Gurriel, first baseman, hit .298 with 31 homers, 104 RBI—similar to Abreu.  He drew only 37 walks. That’s 20 Win Shares.

Corey Seager, shortstop, hit .272 with 19 homers, 87 RBI, .817 OPS, led the National League in doubles with 44, missed about a month of the season with injuries.  That’s 20 Win Shares.

Justin Turner, the red-headed stranger, hit .290 with 27 homers, 67 RBI, .881 OPS, had about the same playing time as Cory Seager.   That’s 20 Win Shares. 

Michael Conforto, Right Fielder, hit just .257 but with 33 bombs, 90 runs, 92 RBI, .856 OPS.   That’s 20 Win Shares.

Paul DeJong, Shortstop, hit just .233 but with 30 homers, 97 Runs Scored, .762 OPS.  Low average, but he was a full-time shortstop on a good team, and that was 20 Win Shares. 

So 20 Win Shares is a really good player, a good season; championship teams need several players at that level, and championship teams are mostly made up of players not quite THAT good.  But 20 Win Shares is not normally an MVP season. 

At 25 Win Shares there were also six players:

Josh Donaldson, third baseman, hit .259 with 37 homers, 94 RBI, also 100 walks, .900 RBI.   That’s 25 Win Shares.

Mookie Betts, right fielder, hit .295 with 29 homers, 80 RBI, also 97 walks, .915 OPS, led the majors in runs scored with 135, was 16-for-19 as a base stealer and as good a defensive outfielder as there was in baseball.  That’s 25 Win Shares.

Xander Bogaerts, shortstop, hit .309 with 52 doubles, 33 homers, 110 runs scored, 117 RBI, .939 OPS, although he does not/did not have the quickness of the best defensive shortstops.  That’s 25 Win Shares. 

Anthony Rizzo, first baseman, hit .293 with 27 homers, 94 RBI, .405 on base percentage, .924 OPS.   That’s 25 Win Shares. 

George Springer, outfielder/mostly center fielder, hit .292 with 39 homers, 96 RBI, .974 RBI.  That’s 25 Win Shares.

Matt Chapman, third baseman, hit 36 doubles and 36 homers, drove in 91 runs, had an .848 OPS, but that was in a pitcher’s park and he was generally regarded as the best defensive third baseman in baseball.  That’s 25 Win Shares.

Players with 25 Win Shares can and do win MVP Awards sometimes, but generally don’t, because generally there are a few players in the league with 30+ Win Shares, and usually one of them will win.  

In the system that I set up to evaluate not individual MVP candidates, but groups of MVP candidates, a player of 25 Win Shares counts 25 times as heavily as a player of 21 Win Shares:

(25 – 20) ^2 = 25

(21 – 20) ^2 =  1

A season of 25 Win Shares counts 25 times as much as 21 Win Shares, but one-fourth as much as a season of 30 Win Shares, one-ninth as much as a season of 35 Win Shares, and one-sixteenth as much as a season of 40 Win Shares. 

(25 – 20) ^2 = 25

(30 – 20) ^2 = 100

(35 – 20) ^2 = 225

(40 – 20) ^2 = 400

Seasons of 40 or more Win Shares don’t happen every year.  Normally, only guys like Ted Williams, Stan Musial, Willie Mays, Mickey Mantle, Barry Bonds, Albert Pujols and Mike Trout ever get to 40 Win Shares, although occasionally an outsider sneaks into the group like Al Rosen (1953) or Norm Cash (1961) or Andrew McCutchen (2012). 

I add up those values to represent the strength of the GROUP of candidates.  But, of course, there are several wrinkles to the system.  After we find those totals, we multiply that by the number of players in the league who had 30 or more Win Shares, plus one.

A good MVP competition requires multiple players competing at a high level.  In the season that we will identify as the toughest MVP race ever, there was no player who had 40 Win Shares, but there were a whole bunch of guys in the neighborhood of 35.  I don’t want a system where one Barry Bonds steroid shot eclipses a whole raft of great competitors (although that does still happen.)  We have to add one to the count of 30+ seasons, because if we didn’t add one, then any season in which there was NO player with 30 Win Shares would rank at zero, and we would lose the ranking system because there would be multiple zeroes in the system.

So that’s wrinkle 1.   Wrinkle 2 is the seasons, like 2020, when a full season is not played.   2020 is not actually in the system yet, not actually in the data, but let’s take 1994, when the average team played only 113-114 games because of the strike, or 1981 or 1995, which were also strike-shortened seasons.  What do we do about those?

            I didn’t want to make a "full" adjustment for the playing time, because that would be adjusting out of existence something that actually happened.  You can’t say that 1981 or 1994 was a year when a lot of players had great seasons, because that didn’t happen.  113-game great seasons are not as great as 162-game great seasons.   On the other hand, I did not want a system in which all of the strike-shortened seasons ranked as the worst MVP groups ever, because that’s just dumb.  So I made a partial adjustment for the shortened season.  I treated 16 Win Shares as 20 (that is, the weight of each season was Win Shares minus 16, rather than Win Shares minus 20), and the "count" of the number of top-level MVP candidates was set at 24 Win Shares, rather than 30. 

            That is for 1981 and 1994, which were seasons of similar length.  For 1995, which was shortened to 144 games, I changed "20" to "19" and "30" to "29".  That way, the strike-shortened seasons are not all grouped as the worst MVP races ever, but they’re effectively prevented from ranking at the top of the chart, either.  They are mostly near the bottom.

            For the years of World War II, when Ted Williams and Bob Feller and Joe DiMaggio were in the war (1943-1945), I adjusted the system by changing "20" to "21" and "30" to "31".  That may seem like a little thing that wouldn’t have much effect, but it isn’t; just that one-point adjustment to the system has a meaningful effect.  If the system returned results saying that the toughest MVP races ever occurred in the 1943-1945 era, when most of the best players in the world were off fighting the war. . .well, I wasn’t trying to jigger the results; I was trying to let the data speak for itself.  But that would not have been the "real" answer, so I did not want it to be the answer that I got. 

            The issue of expansion is tougher.  There are multiple issues of expansion, two tough and one not.  The three issues having to do with comparing seasons post-1960 to seasons pre-1960 are:

1)     The number of teams competing in each league,

2)     The length of the schedule, 162 games as opposed to 154, and

3)     The quality of the competition in the first couple of years after expansion. 

1) The number of teams competing in each league.    There used to be 8 teams in each league, now there are 15, so should we adjust the "tough competition" scores to be fair to the years before expansion?

Very clearly we should not do that.   That would be adjusting reality out of existence.   The fact is, winning an MVP Award in a 15-team league IS much more difficult—and much more impressive—than winning an MVP Award in an 8-team league.   It would be obviously wrong to adjust that away, and give the modern player no more credit than if he were competing for the award with half as many players.

2)  Schedule length, 162 games versus 154. 

That’s a tougher question, but I ultimately decided not to make an adjustment.   The schedule length is offset by the forces of history, which constantly reduce the extent to which the best players dominate the average players, thus reducing the number of players who reach the relevant plateaus (relevant to this study—20 Win Shares and 30 Win Shares.)  From 1931 through 1940, the first ten years of modern MVP voting, there were 160 teams playing 154-game schedules, but there were 409 players who earned 20 or more Win Shares.  That’s 2.6 per team.  From 2010 to 2019 there were 300 teams playing 162-game schedules, but there were only 535 players who earned 20 or more Win Shares.  That’s 1.8 per team.  The number of players reaching the relevant levels not only has not increased due to the length of the schedule, it has very significantly DECREASED over time.   That’s why I decided not to make a schedule-length adjustment. 

3) The quality of competition in expansion years. 

An expansion does somewhat dilute the talent for a few years, and some of the expansion seasons have had an unusual number of players having "great" seasons (although some of them have not.)   I could have chosen to treat those years like the War Years, and adjust the standards to diminish slightly the "toughness" scores of those competitions.   Ultimately I decided not to do that, mostly because I don’t believe that the expansion effect was really very significant.  The number of hitters having great years in 1961 was mostly due to the two new parks brought into the AmericanLleague, which were the two best hitters parks in the league, or maybe 1 and 3.  The Win Shares system automatically adjusts for that.   The quality of "expansion pitching" didn’t really have that much to do with it, I don’t believe. 

So that was the system, but then I had to make an after-the-fact adjustment as well.  When I first figured the results, the National League in the steroid seasons completely dominated the results, with the Barry Bonds-Mark McGwire-Sammy Sosa seasons.  

Is that right, or is it a blind spot in the method?  I think it is a blind spot in the method.  Steroids allowed SOME players to pull way ahead of the AVERAGE players.  I’m not a fanatic about that; it was just something that happened.   It was regrettable, but it happened; let’s move on. 

But I don’t really want a list of steroid seasons to show up as all of the toughest MVP races of all time, either.  I think that allowing that to happen was a blind spot in the system.  I applied the "war-time adjustments" to the years 1996 to 2005, as well—that is, I treated value above 21 as value above 20, and a season of 31 as a season of 30.   Which, I repeat, actually make a pretty big difference in the scores. 

OK, so. ..take it for whatever you think it is worth.   We will have seven levels of MVP races here, which are:

The 10 toughest MVP races ever, which we will treat as the top 5% although it is actually more like 6%,

The "extremely tough" MVP races, which is the next 10%,

The "tough" MVP competitions, which is the next 20% after that,

The "normal" MVP competitions, which is the middle 30%, from the 35^th percentile up to the 65^th,

The "soft" MVP competitions, which is the 20% below that,

The "weak" or "very soft" MVP competitions, which the 10% below that, and

The 10 weakest MVP races ever, which we will treat as 5% although it actually closer to 6%.  From 1931 to 2019 there have been 88 seasons, so there have been 176 MVP competitions. 

We’ll start at the bottom, the ten weakest MVP seasons of all time:

10.  1981 NL Award, won by Mike Schmidt.  Schmidt was having an absolutely fantastic season in 1981, interrupted by the strike but otherwise headed for the greatest season of his Hall of Fame career.   The other top MVP competitors were Andre Dawson and George Foster, who were 2-3 both in Win Shares and in MVP votes.  Without the strike, it would not have been a weak MVP competition, but when you take a third of the season away from each player, it’s just not the same. 

9.  (Meaning WORSE than the 1981 competition)  American League in 2009, won almost unanimously by Joe Mauer.    Mauer had a great year, hitting .365 with 28 homers, 96 RBI.  He led the league with 32 Win Shares, so he was by that standard deserving of the MVP Award.   (WAR has him third in the league, behind Greinke and Zobrist.)  I would go with Mauer, but 32 Win Shares is on the low side for a league leader, and the fact that he was one vote away from a unanimous selection is a clue here.   Nobody else in the league had 30, which is unusual.  Although it won’t seem unusual compared to the other seasons on this list, all of which had weak competitions. 

8.  1960 American League, Roger Maris.   I said in the post that provoked this research that it was a weak MVP class, and the research confirms that it was.  Maris and Mantle were the only two guys in the league who had anything remotely resembling an MVP season; Maris won the MVP Award because he had a higher batting average (.283-.275) and more RBI (112-94), which were the 1-2 elements of an MVP season in that era.  No one else in the league had more than 24 Win Shares. 

7.  2018 National League, Christian Yelich.   Yelich was the only guy in the league with 30 Win Shares.  He had 34.   Like Joe Mauer, he was one vote away from a unanimous selection.   And it’s a legitimate MVP season, but it wouldn’t have been anywhere NEAR unanimous in 19 seasons out of 20. 

6.  1962 American League, Mickey Mantle.   Mantle won the MVP Award despite missing 40 games with his leg problems, winding up with 30 homers, 89 RBI.   By Win Shares, Mantle had SIX seasons better than 1962 in which he did not win the MVP Award.  He said on winning the Award that he was pleased and honored, but disappointed that he had taken the Award away from Bobby Richardson, who finished second in the voting despite having only 22 Win Shares. 

5.  1981 American League, Rollie Fingers.   Again, it would not have been a weak class if it were not for the strike.  Rickey Henderson (27 Win Shares) or Dwight Evans (26) would have been vastly better choices for MVP than Rollie Fingers (17)—as, for that matter, would have been Cecil Cooper, Eddie Murray, Bobby Grich, Gorman Thomas, Robin Yount or Dwayne Murphy.  Or others.  Fingers was way down the list of people who should have won it, but the popular fiction of the moment was that it all depended on your closer. 

4.  1931 National League, Frankie Frisch.   The first year of the BBWAA awards, 1931.    Like Fingers, Frisch was a terrible selection from a weak class.  Wally Berger was the only player in the league with 30 Win Shares, and he had only 31.  Frisch won the Award although he was tied for 17^th in the league in Win Shares.  I don’t know where he ranked in WAR, but he had 4.0 WAR and the top ten ended at 5.2, so I would guess that he was even lower than 17^th.  I don’t doubt that Frisch had qualities as a player which have not been documented by statistical analysis.   I mean that sincerely, not ironically.   Just because we TRY to measure everything doesn’t mean that we DO measure everything. 

3.  1994 National League, Jeff Bagwell.   Another strike-shortened season.  Bagwell obviously deserved the award, and won it unanimously.   Craig Biggio, second in the league by Win Shares, was 16^th in the voting. 

2.  1963 American League, Elston Howard.   In the original post I had observed off-hand that this was an extremely weak MVP class.  I didn’t realize at that time HOW weak it was.  Other than the strike-shortened 1981  AL season, it is the only league since 1931 which had NO ONE with 30 Win Shares.   Bob Allison led the league in WAR, Carl Yastrzemski and Tom Tresh tied for the lead in Win Shares (29), but Howard seems as deserving of the Award as anyone.   No one did anything that would normally be taken seriously as an MVP season.

The American League at that time was just very much star-deprived.  This is not an original observation; they lost the All Star game with the predictability of mosquitoes in a swamp, and that league at that time had very, very few Hall of Famers.   This is just another piece of evidence there—1960/1962/1963 AL MVP races all ranking as among the weakest ever. 

1.      2004 American League, Vladimir Guerrero.   

The memory of the American League in 2004 is very special to me, and it hurts me to put down the league in any way, but facts are facts.   Vladimir hit .337 with 39 homers, 126 RBI, a .989 OPS, and those are very good numbers and deserving of an MVP Award; he also had 206 hits, and led the league in Runs Scored.  But the steroid era abounds in big, BIG numbers, so it took a very large number of runs to win a game.  If you compare Vlad’s numbers in 2004 to the other MVPs of that era, you can see that those are actually not fantastic numbers for that era.  He had 27 Win Shares, 5.6 WAR, both pretty weak totals for an MVP; the other MVP candidates (Gary Sheffield, Miguel Tejada, A-Rod, David Ortiz) all had about the same kind of numbers.  I wish David had won it, but I wouldn’t want to take anything away from Guerrero.   I’m just saying that 2004 was no better than any other season he had between 1998 and 2002, but it won the MVP Award because the competition wasn’t as strong as it had been in the National League in the late 1990s. 

OK, that finishes the bottom 10.  These 17 MVP competitions rank as "weak" or "very soft", which is not to suggest that the winner was undeserving, but rather, that the winner’s competition was not as strong as it usually is.  The player himself may have been absolutely fantastic, and in some cases was.

1939 National League (MVP:  Bucky Walters)

1941 National League (Dolph Camilli)

1943 National League (Stan Musial)

1945 National League (Phil Cavaretta)

1946 National League (Stan Musial)

1951 American League (Yogi Berra)

1953 American League (Al Rosen; Rosen was superman)

1955 American League (Yogi Berra)

1956 National League (Don Newcombe)

1958 American League (Jackie Jensen)

1959 American League (Nellie Fox)  Again, this is part of the American League’s "down" era, when they had a real shortage of star players. 

1980 National League (Mike Schmidt)

1990 American League (Rickey Henderson)

1998 American League (Juan Gonzalez)

2003 American League (Alex Rodriguez)

2006 American League (Justin Morneau)

2008 American League (Dustin Pedroia)

These 35 leagues had "soft" MVP competitions, but not "weak" MVP competitions:

These 52 MVP races were of normal strength, neither stronger than usual nor weaker than usual:

I said before, in my original comments on this subject in "Hey, Bill", that the 1960 National League race was weak.   This turns out to be not true; it actually was a competition of normal strength, although the winner, Dick Groat, had only 25 Win Shares, which is a very weak number for an MVP.  I made that mistake because (a) I used the MVP, Groat, as a basis for my evaluation, (b) I compared Groat first to Roberto Clemente, explanation later, (c) the two best players in the league, Willie Mays and Hank Aaron, both had down years by their own standards, which were still very impressive seasons in absolute terms. 

Sorry for going on about this; maybe one of you will find it interesting.  The 1960 NL MVP race involves these players:

1)  Eddie Mathews, Milwaukee Braves, 38 Win Shares.  Probably Mathews should have won it.  He hit just .277, but with 39 homers, 124 RBI, and also 111 walks. . . .a season that anyone would like to have.  But he wasn’t going to win it because his batting average and his home runs were both way down from his 1959 season, and his team didn’t win the pennant.   

2) Willie Mays, San Francisco, also 38 Win Shares.  Mays hit "just" .319 with "only" 29 homers and "only" 103 RBI, and for Willie Mays, that looks like a significantly sub-par season.   But it’s misleading, because Candlestick Park opened that year.  It was a cold year in San Francisco; they played a bunch of games in 58 degree weather with the wind whipping around.   All of the hitters on the team had terrible years by their own standards.   Orlando Cepeda, who always hit .300 with 100+ RBI and usually 30-40 homers, hit .296 with 24 homers, 96 RBI.  Willie McCovey hit .238 with 13 homers, a dreadful season.  Eddie Bressoud, who had really nice seasons with the bat for Boston from 1962-1964, hit just .225 for San Francisco in 1960, and Don Blasingame, who hit .289 in 1959 and .281 in 1962, a regular both years, hit .235 in 1960.  In the context of the place and time, Mays’ 1960 season is actually really impressive.   He led the league in WAR by a pretty wide margin (9.5 to 8.0). 

3.  Hank Aaron, Milwaukee Braves.   Again, his numbers were way down from 1959, but he hit .292 with 40 homers, 126 RBI.   You CAN win an MVP Award with those numbers.   35 Win Shares.    If you compare him to the American League MVP that season, who hit .283 with 39 homers, 112 RBI, you see the point.

4.  Ken Boyer, St. Louis Cardinals.  1964 MVP.  By any modern analysis, he had at least as good a season in 1960 as he did in 1964.  31 Win Shares. 

5.  Ernie Banks, Chicago Cubs.  Significantly off from his MVP seasons in 1958-1959, but you still have a shortstop here hitting 41 homers.  29 Win Shares. 

6.  Orlando Cepeda, San Francisco.   26 Win Shares despite numbers well below his usual. 

7.  Don Drysdale, Los Angeles.  Finished just 15-14 despite phenomenal peripheral stats.   WAR (7.0) was the second-best of his career.  25 Win Shares, the same as the MVP. 

8.  Joe Adcock, Milwaukee.  Hit .285 with 35 homers, 108 RBI.  25 Win Shares. 

9.  Lindy McDaniel, St. Louis.  Fantastic year in relief.  Also 25 Win Shares. 

Dick Groat won it, with 25 Win Shares.  After the season, Roberto Clemente (then usually called "Bob" by the American press). . .Bob Clemente groused openly about Groat winning it.  This is against the unwritten rules of the game; when you win the World Series, you’re not supposed to complain about your teammate winning the MVP Award.  It isn’t done. 

Also, Clemente was wrong.   Groat WAS more valuable, that season, than Clemente.  Clemente would be a great player later on, but in 1960 he was just better than he had been before 1960; he wasn’t really great until 1961.  Because of Clemente’s complaining, which was well known at the time, I tended to resolve the issue in my mind by saying, "Well. . .Groat or Clemente, I guess I have to go with Groat,  It was just kind of a weak crop of MVP candidates."

But actually it wasn’t; actually it was a normal crop of MVP candidates; it was just a poor choice.  But there were good candidates out there. 

The following seasons had "tough" MVP competitions, meaning stronger than average, but not strong enough that we would want to make a big deal out of it:

And now we get to the genuinely tough MVP contests, the "Very Tough" competitions.

1931 American League (MVP:  Lefty Grove.  Lou Gehrig drove in 184 or 185 runs, and wasn’t the league MVP.)

1932 American League (MVP:  Jimmie Foxx.  Three players with 36 Win Shares, both in 1931 and 1932.  I’ll explain the significance of that at the end of the article.)

1934 National League (Dizzy Dean)

1941 American League (DiMaggio.  Ted Williams hit .406; DiMaggio hit in 56 straight.  Both men had 40+ Win Shares, which is unusual, but it has happened three other times.  There is a lot of mythology that revolves around this competition, but those are magic numbers—magic numbers closely allied to value, yes, but still magic numbers.)

Well, let me tack the opposite direction on that issue.  There are only four seasons since 1931 in which two players in the same league had 40 Win Shares each—the 1941 AL, with Williams and DiMaggio, the 1961 AL, with Norm Cash and Mickey Mantle, the 1962 NL, with Willie Mays and Frank Robinson, and the National League in 2001, with Barry Bonds and Sammy Sosa. 

The 1941 American League is the only one of those without an invisible asterisk next to it.  The 1961 AL and 1962 NL are both expansion years with expansion pitching, and in both cases neither players with 40 Win Shares won the MVP Award.  In 2001 the two players were steroid monsters. 

Going back the other way, now. . .1941 AL is the only such season SINCE 1931, but there are many such seasons before 1931.   There are also several seasons in the late 1940s in which both Musial and Williams had 40 or more Win Shares, but they, of course, were in different leagues. 

1961 American League (Maris.  Norm Cash and Mickey Mantle.)

1964 National League (Ken Boyer won the MVP Award, because of RBI, but the top five in Win Shares were Dick Allen (41), Willie Mays (38), Ron Santo (36), Frank Robinson (33) and Hank Aaron (33).  That’s a lot of MVP candidates. 

1966 National League (Roberto Clemente won the MVP Award, because of RBI, but the top five in Win Shares were Willie Mays (37), Sandy Koufax (35), Dick Allen (35), Willie McCovey (34) and Juan Marichal (33).   That’s a lot of MVP candidates.

1967 National League (Orlando Cepeda won the MVP Award.  Top five in Win Shares are Ron Santo (38), Roberto Clemente (35), Hank Aaron and Orlando Cepeda (34 each), and Lou Brock and Tim McCarver (30 Each.)  If you have that many people with 30-point seasons, that’s a tough race.

1969 American League (Harmon Killebrew.   Top five in Win Shares are Reggie Jackson (41), Rico Petrocelli (37), Sal Bando (36) and Killebrew and Frank Howard (34 each.) 

1972 National League.  Top five in Win Shares are 27-game winner Steve Carlton (40), Joe Morgan (39), Johnny Bench (37), Cesar Cedeno (33) and Billy Williams and Pete Rose (32 each.)  Johnny Bench won the award, and perhaps he should, since perhaps the catcher deserves a little extra credit.

1973 National League (Pete Rose won the Award.  Top five in Win Shares are Joe Morgan (40), Willie Stargell (36), Pete Rose (34), Tony Perez (32), Darrell Evans (31) and Barry Bonds (31). 

1985 National League (Willie McGee (36) won the Award, and deserved it, but there were five other players who had MVP seasons as well:  Tim Raines (36), Pedro Guerrero (35), Dwight Gooden (33) and Gary Carter (33). 

1989 National League.  Will Clark had 44 Win Shares, a historic number, but it would require like five pages to explain why the system works that way.  He had some super-under-the-radar numbers.   Howard Johnson and Kevin Mitchell also had MVP-worthy seasons, 38 Win Shares each.  Kevin Mitchell won the vote. 

National League, 1996-1997-1998.   Beginning of the serious steroid era; best players pulling away from the league.  1996 was Bagwell (41), Bonds (39), Caminiti (38); Caminiti won the MVP Award.  1997 was Piazza and Tony Gwynn, 39 each, Craig Biggio 38, Barry Bonds, 36.  Anything above 35, you’re a strong MVP candidate.  Larry Walker (32 Win Shares) won the MVP vote.  1998 was the home run duel between McGwire (41) and Sosa (35).   Sosa won the vote; Craig Biggio also had 35 Win Shares, as well.    

2015 National League; Bryce Harper won the award.  Leading candidates by Win Shares are Harper (38), Paul Goldschmidt (35), Andrew McCutchen (35), Joey Votto (33) and Anthony Rizzo (32).  Leading candidates by WAR were Harper (9.7), Greinke (9.5), Arrieta (8.6), Goldschmidt (8.4), and Votto (7.8).

Now entering the top 10, the ten toughest MVP competitions of all time: 

American League, 1970 (MVP:  Boog Powell).   In the American League in 1970 there nine players—9, count ‘em—who are credited with 30 or more Win Shares.  30 Win Shares makes you a credible candidate for MVP; 35, I would say, makes you an MVP favorite in most seasons.  It may be 36, I don’t know; I haven’t actually studied that. 

Anyway, in the American League in 1970 we have Carl Yastrzemski (36), Roy White (34), Jim Fregosi (33), Tommy Harper (33), Boog Powell (31), Sam McDowell (30), Tony Oliva (30), Harmon Killebrew (30), and Frank Howard (30).  By WAR, the leaders are Yastrzemski (9.5), McDowell (7.9), Fregosi (7.7), Harper (7.6), Oliva (7.0) and Roy White (6.8), and then we diverge into a different group, getting rid of the three slow-moving sluggers (Powell, Killebrew and Howard) and replacing them with Reggie Smith (6.7), Jim Palmer (6.4), and Aurelio Rodriguez (6.4).   Boog Powell, the elected MVP, had 5.1 WAR, so he might make the top 20.

But I am not here to rail against the voting, or to denounce it or criticize it or anything.  This isn’t about who won and who should have won; this is about the strength of the class of candidates.  Nine players with 30 WAR is a strong class.   There have been nine which were stronger. 

            9.  National League, 1971 (Joe Torre).   Torre deserved and won the Award, with 41 Win Shares.   WAR ranks Torre first in the league in Offensive WAR (8.7), but regards him as a defensive nightmare, thus not even among the top 10 players in the league in overall WAR.   Ferguson Jenkins (37 Win Shares) is second in Win Shares, first in WAR.   The other strong candidates are Willie Stargell (35), Hank Aaron (33), Rusty Staub (32), Tom Seaver (32), Bobby Bonds (32) and Lou Brock (32).  

            8.  American League, 1946.  Ted Williams did not lead the league in Home Runs, RBI or Batting average—but had perhaps his greatest season, hitting .342 with 38 homers, 123 RBI, drawing 156 walks and leading Boston to 104 wins.

            This chart compares the top candidates in that league, by Win Shares, MVP voting and WAR:

In this case, there is

·      88% agreement between Win Shares and MVP voting,

·      86% agreement between Win Shares and WAR, and

·      77% agreement between MVP voting and WAR.

This can be figured as follows, using Win Shares and MVP voting for illustration.  Ted Williams is at the top of both lists, so we score that as 10 X 10, or 100 points.  Johnny Pesky is second on one list (9 points) and fourth on the other (7 points), so we score that at 9 X 7, or 63 points.   Hal Newhouser is third on one list (8 points) and second on the other (9 points), so we score that at 8 X 9, or 72 points.  Mickey Vernon is fourth on one list (7 points) and fifth on the other (6 points), so we score that at 7 X 6, or 42 points.  Bob Feller is fifth on one list (6 points) and sixth on the other (5 points), so we score that as 6 X 5, or 30 points.  Hank Greenberg is sixth on one list (5 points) and eighth on the other (3 points), so we score that at 5 X 3, or 15 points.  Bobby Doerr is 9^th on one list (2 points) and third on the other (8 points), so we score that at 2 X 8, or 16 points.

Adding all of those up (100 + 63 + 72 + 42 + 30 + 15 + 16), we have 338 points.   If the two lists aligned perfectly, that total would be 385 points, so we divide the actual total (338) by the possible total (385), which gives us .877, which is 88%.  If no one who was on one list was on the other, that would be zero percent agreement.  If the lists were the same, it would be 100%.

One wrinkle is that, in the case of ties, I always interpret the tie in such a manner as to result in the maximum score.  In other words, Bobby Doerr and Dizzy Trout are tied, at 27 Win Shares.  Since Doerr is also on the MVP list, and Trout is not, then, in figuring the degree of agreement between Win Shares and MVP voting, I count Doerr as being in 9^th place (2 points) rather than 10^th place (1 point), since this yields a higher score.   But in comparing Win Shares to WAR, since Trout is on the WAR list and Doerr is not, I count Trout as being in 9^th place, since this yields the higher score.   But you can’t count BOTH of them as being in the higher place in the same calculation, since this would make it theoretically possible for a score to exceed 100%. 

Although eighth on the overall list, this ranks as the toughest MVP contest in American League history—not the closest, but the toughest.  From Ted Williams’ standpoint, he beat a list of 7 or 8 other guys who would have been decent MVP selections.  From their standpoint, they all had to compete with Ted Williams.   All seven seasons listed ahead of this one came from the National League.

7.  2004 National League (Barry Bonds).   Bonds drew 232 walks, giving him a .609 On Base Percentage.  Adrian Beltre hit 48 homers and .334.  Albert Pujols hit 46 homers and .331.  The National League had ten players with 30 or more Win Shares—those three plus Scott Rolen, Jim Edmonds, Bobby Abreu, Mark Loretta, J. D. Drew, Todd Helton and Lance Berkman. 

6.  1968 National League (Bob Gibson).  Gibson, of course, had his greatest season, with a 1.12 ERA (36 Win Shares.)  Nine other players had 30 or more—Willie McCovey, Hank Aaron, Pete Rose, Jimmy Wynn, Dick Allen, Felipe Alou, Lou Brock, Billy Williams and Willie Mays. 

5.  1963 National League (Sandy Koufax).  Koufax had his first Cy Young season, going 25-5 with a 1.88 ERA, but Win Shares sees him as being nowhere near as valuable as Henry Aaron (41 Win Shares) or Willie Mays (38).  

            Baseball re-defined the strike zone in 1963, pitching the game into a six-year offensive slump.   Aaron hit .319 with 44 homers, 130 RBI—about his usual numbers, but, because the offensive context shrunk, Aaron’s value increased relative to his context.  Thus, Aaron went from a measly, pathetic 34 Win Shares in 1962 up to 41 Win Shares in 1963, based on essentially the same numbers.   Same thing in WAR; Aaron went from 8.5 to 9.1 despite a small decrease in Runs Above Replacement.   Of course, the contemporary MVP voters did not see this, did not understand this, and I’m not even necessarily saying that it is right; I am just saying that this is the pathway that analytical statistics have taken.  Personally, I’m fine with Koufax being the MVP. 

Win Shares:

1.     Aaron

2.     Mays

3.     Koufax (tied, 32)

4.     Dick Ellsworth (tied, 32)

5.     Johnny Callison (tied, 32)

WAR:

1.     Willie Mays

2.     Dick Ellsworth (9.9)

3.     Sandy Koufax (9.9)

4.     Hank Aaron

5.     Juan Marichal

MVP Voting:

1.     Koufax

2.     Dick Groat (huh?)

3.     Hank Aaron

4.     Ron Perranoski

5.     Willie Mays

Groat had 31 Win Shares and 7.1 WAR, so he certainly had a better season in 1963 than in 1960, when he actually won the MVP Award.

Dick Groat had a lot of seasons in his career in which, to be blunt, he wasn’t worth a crap.   He was a .300 hitter with no power, didn’t walk, terribly slow, grounded into a LOT of double plays, and played shortstop, but wasn’t really a GOOD shortstop; he wasn’t quick, and his fielding percentages were a little below average. 

However, in all honesty, he WAS a good player in 1960 and in 1963.   His game worked for him; the percentages were high enough relative to the league that he did have a lot of value. 

4.  1965 National League (Willie Mays).   Willie Mays had one of his greatest seasons, with 43 Win Shares.   Nine legitimate MVP candidates, although none of the others really close to Mays:  Billy Williams, Sandy Koufax, Dick Allen, Ron Santo, Jimmy Wynn, Hank Aaron, Joe Morgan and Juan Marichal.    Mays in 1965 had nearly identical statistics to Frank Robinson in 1966, although Robinson’s season is much more remembered because he happened to win the Triple Crown. 

 3. 1992 National League.  Barry Bonds (41 Win Shares) won his second MVP Award, beating out Terry Pendleton (35), Andy Van Slyke (35), Ryne Sandberg (33), Barry Larkin (32), Craig Biggio (32), Gary Sheffield (32), Darren Daulton (31), and Ray Lankford (31). 

            2. 2001 National League.  Bonds (73 homers, 177 walks and 54 Win Shares) won his 5th MVP Award over an entirely different cast of competitors—Sammy Sosa (42), Luis Gonzalez (37), Shawn Green (34), Rich Aurilia (33), Lance Berkman (32), Phil Nevin (31), Jeff Bagwell (30), Gary Sheffield (3) and Jim Edmonds (30).   It is almost as if he had found some wonder drug which kept him young while those around him grew older. 

            And the #1 toughest MVP contest of all time is:  the National League in 1969. 

            The late Tom Seaver was 25-7, 2.21 ERA, leading the Miracle Mets to the World Championship.   He would have been a fine MVP selection, honestly, but he was sixth in the league in Win Shares, and second in MVP voting.   The late Bob Gibson, realistically, was probably a better pitcher; Gibby went 20-13, but pitched 40 more innings than Seaver with a slightly better ERA and a much better strikeout/walk ratio.   Gibson led the league in WAR by a huge margin (11.3 to 8.4), but was 5^th in Win Shares.   The 11.3 WAR is in part because Gibson, who was a good hitter, had a good year with the bat, as well. 

            Anything above 35 Win Shares we can describe as a "strong" MVP.   Jimmy Wynn hit .269 with 33 homers and 87 RBI, numbers that won’t win you an MVP Award, probably, and certainly would not in 1969, but he drew 148 walks, which was a National League record until somebody invented Barry Bonds.  We credit him with 36 Win Shares.

            Pete Rose was ahead of Wynn, with 37 Win Shares.  Rose won the batting title at .348, and it was not an empty .348; he had 88 walks, 33 doubles, 16 homers and 82 RBI.  By Win Shares, Rose was better in 1969 than in 1973, when he actually did win the NL MVP Award.  WAR disagrees; WAR believes that Rose was rosier in 1973, but I don’t know that this is worth quarrelling about; the real point is that there are four MVPs running loose in this league, not counting Tom Seaver or Bob Gibson.

            Hank Aaron hit .300 with his usual 44 homers; he is second in Win Shares with 38.  The Award went to the league leader in Win Shares, Willie McCovey, who hit .320 with 45 homers, 126 RBI—easily the best season of his Hall of Fame career. 

            So that is four MVPs—McCovey (39), Aaron (38), Rose (37) and Wynn (36), before we get to Bob Gibson (33) and Tom Seaver (32).   And actually, we’re not done yet; after them we have Tony Perez, who hit .294 with 37 homers, 122 RBI (32 Win Shares), Bobby Bonds, who hit 32 homers and stole 45 bases, scored 120 runs (31 Win Shares) and Cleon Jones, who hit .340 with a .422 on base percentage (30 Win Shares).  Somehow, Tommie Agee doesn’t qualify for this list, although I am not sure where he falls short.

            Suppose we say that 36 Win Shares is a "strong" MVP season, don’t take the expression too seriously.   The 1969 National League is the only season, other than a couple of steroid seasons, in which there are four players with 36 Win Shares.   There are 176 league/seasons which have had an MVP vote. 

            In 60 of those leagues, there is no player with 36 Win Shares.

            In 66 of them, there is one player with 36 or more Win Shares.

            In 31 leagues, there are two players with 36 or more Win Shares.

            In 16 leagues, there are three players with 36 or more Win Shares—three players fully deserving of an MVP award.

            In three leagues, there are four players with 36 or more Win Shares, but two of those are steroid-era leagues.   So the 1969 National League season is the only season in history with no steroids and four absolutely solid MVP players.

            So it was a hell of an MVP contest; in fact, it was the toughest MVP contest of all time.   I didn’t intend to go on quite so long about this, and I appreciate your patience.