05 March 2020

Ken Griffey Jr. and NL Gold Glove Voting Trends

One of the more fun aspects of Joe Posnanski's run-down of the 100 greatest players of all time over at The Athletic is his tendency to go down into "rabbit holes".  He'll chase something that takes his fancy and the next thing you know you've learned about all the notable players named after US Presidents, or about the Negro Leagues, or about whether Warren Spahn really threw a screwball.  Or whatever.

Also, I like that these articles often send me off into some rabbit hole of my own, usually due to some throwaway line in the article. Thrown away only because (I imagine) - with 100 of these articles to write in 100 days - Joe simply does not have time to chase down every one of these esoteric little tidbits, not because he doesn't want to.

A few weeks ago it was about how/why Robin Roberts somehow did not win the 1952 NL MVP Award.  The answer to that, the three or four of you who may have read my blog post will recall, was that the sportswriters were dumb and inexplicably voted for relief pitchers with unusually high Win totals on the merits of they'd never seen that before.

Today's rabbit hole comes from the article on player #48, Ken Griffey Jr.

Griffey was traded to Cincinnati by request just 10 or so weeks after he turned 30. He was good in his first year with the Reds — he hit 40 homers and slugged .556 — but for the first time since he was a rookie, he did not win a Gold Glove.*

*The voters, oddly, gave a Gold Glove to Steve Finley instead. Look, over his career, Griffey won several Gold Gloves that, in retrospect, look questionable, but it’s entirely unclear how he could have lost the 2000 Gold Glove to Finley, who was playing with a bulging disc in his back and had well-below-average range that year.

So this got me to looking into the history of NL outfield Gold Glove Awards.

The first ones were given in 1957, but these were for all the major leagues.  Willie Mays won one, of course, but so did Al Kaline of the Tigers and Minnie Minoso of the White Sox.  The next year they split them up by leagues, and the NL winners were Mays, Hank Aaron and Frank Robinson.

In those days they actually assigned ones to Left, Center and Right field, though that would change in 1961 when they just started giving them to any three outfielders, which meant (usually) Roberto Clemente and two center fielders.  If you were assigned to Left, it was usually because you were not a good defender in the first place.

From 1961-68, the NL Awards went to Mays, Clemente and someone else, usually Curt Flood.  Those three won them six years in a row, the longest such stretch in history.  Fifteen years earlier, White America would have never even see those three play.  Think about that for a while. 

By 1969, Mays was 38, and played in only 117 games, so Pete Rose won "his" award.  Then Curt Flood got himself embroiled in a legal controversy you may have heard of, so he was no longer playing in 1970, and Tommie Agee won that award instead, along with Rose and Clemente.

In 1971, Rose and Agee were replaced by Bobby Bonds and Willie Davis.  And that is the last time until 1995 that two new players will the NL Gold Glove in the same year, almost a quarter of a century later.

You see, I think Griffey's failure to win the 2000 NL Gold Glove comes down to what you might call "institutional inertia".  The NL GG voters - managers and coaches - evidently rarely deviated much from whomever they had voted for last year. In the AL, though the number of different players winning Gold Glove awards in the outfield is nearly identical overall, there was a lot more year-to-year change than in the NL, for some reason. 

But for the NL, between 1962 and 2005, a span of 44 seasons, only three(!) times did two of the three outfield Gold Glove awards go to people who had not won it the previous year.  The first was the one I just mentioned.  The other two:

1995: Finley, Grissom, Mondesi (Grissom, Barry Bonds and Darren Lewis had won in 1994)
1997: Bonds, Mondesi and Larry Walker (following Bonds, Finley and Grissom in 1996)

And that's the end of the list.  In the other 41 seasons, no more than one change was made from the previous year, and seven times, no changes happened at all.

I don't know if this means that those players really were that dominant or if it means the managers and coaches were just lazy and generally voted for whomever they listed last year, unless that player changed leagues, or positions or died or something.  It's just an observation of a trend.

Anyway, as for the NL Gold Glove situation in Y2K...

In 1999, Finley, Andruw Jones and Larry Walker had won the NL Gold Gloves.

Jones, was, of course, AMAZING in CF with 3.8 dWAR, not that anyone knew this at the time, since that statistic did not exist yet, but they knew he was awesome.  That easily led all NL outfielders, which is to say that he deserved the award.  Finley was 3rd in dWAR (1.9), a good distance behind Mike Cameron (2.6), but deserving, nonetheless.  Larry Walker actually had the worst defensive season of his career in 1999, -1.2 dWAR, but he also threw out 13 runners in just 114 games, and those are the kinds of things voters remember, I suppose.

In 2000, well, Finley and Jones were both still healthy and productive but Larry Walker was injured (because of course he was...) so who should the third OF award go to?

If you look at dWAR among NL outfielders in 2000 (min 130 games), the top 3 were Jones (2.7), Tom Goodwin (1.4) and Griffey (1.3), with Richard Hidalgo (1.2) hot on their heels.   Edmonds was 11th, at 0.4 dWAR.  Finley was 18th (!), at exactly zero.  But voters can be fooled by recency bias and by SportsCenter, so here we are. 

The trouble, I think, was that Griffey joined the NL in the same year as Jim Edmonds - who simply *looked* a lot more exciting out there in CF, made more SportsCenter highlights, etc.  So, even though Griffey covered more ground, made more plays, etc., Edmonds had one more Assist, made one less Error, and that's as far as anyone probably looked at the numbers at that time.  They likely never even thought much about Goodwin or Hidalgo of the fact that The Kid made 24 more Put Outs in CF than Edmonds.

In 2006, the voting started to change in character.  With the advent of better measuring sticks for defense, voters started paying more attention, doing more homework as it were, and the voting became less of a popularity contest.  In 2011 they went back to awarding Left, Right and Center-fielders separately.  In 2013 they incorporated a sabermetric element to account for 25% of the vote weight.  And it's all helped.

In the 14 most recent seasons, i.e. since 2006, only three times have the NL Gold Gloves been awarded to two or more of the previous year's winners.  Just as many times, all three winners have been someone who did not win the previous year.  And every year since 2006 has seen at least one brand new name appear on the list of awards, including eight seasons with two new names.  Not just two who didn't win last year, but two who had never won before. 

But 20 years ago, none of that was happening.  Steve Finley was winning the award despite being exactly average on defense.  Rafael Palmeiro was winning the AL GG award for first base despite playing only 28 games there.  Jermaine Dye won an AL Gold Glove that year, and he was straight up terrible! (In fact he only had a positive dWAR in a full season once in his 14-year career!)  It was chaos!

So yes, Griffey probably deserved that award in 2000.  But since the same system had allowed him to win the award with a negative dWAR of his own twice (1992 and 1999) as well as in a season in which he played only 72 games (1995) perhaps the system does not owe him anything?

Stumble Upon Toolbar

21 January 2020

Notes on the MLB HoF Voting before the 2020 Announcements

Modern technology is great.

For the last decade or so, a small group of devoted fans, led by Ryan Thibodaux, has been collecting the publicly and privately confessed Cooperstown voting results from BBWAA members  prior to the announcement of the actual vote, which is anticipated tonight at 6PM.  Far from eliminating any tension or wonder before the official announcements, this serves only to increase it, or perhaps just change its nature.  

Since we only know a portion of the voting results (they have a little over half of the ballots accounted for at this point) they also calculate the percentage and number of remaining ballots needed to make enshrinement, or to meet the 5% minimum requirement for staying on the ballot. They track not just which sportswriters voted for whom, but who they didn't vote for, if they did last year, and who they may have added to their ballot.  Additions are coded in green, retractions in red.  

Pete Abraham, for example, added Todd Helton and Billy Wagner to his ballot.  Filip Bondy removed Gary Sheffield (in his 6th year of eligibility) but added Helton, Jeff Kent, Manny Ramirez and Larry Walker, in his 10th and final year.  Bondy used all 10 of his available votes, so you can see why he would think Sheffield expendable, if it meant giving Walker one last chance at enshrinement.  

Peter Gammons, on the other hand, removed Helton and Walker from his ballot, but added Kent.  Gammons has also consistently voted for Barry Bonds, Roger Clemens and Curt Schilling, so I guess it's OK to be a loudmouth who jokes about reporters being killed or a cheater who uses drugs to improve his results, but being an awesome hitter in Colorado is somehow no longer acceptable behavior.  Man, having his face on the $20 bill has really gone to Gammons' head.  

Other curious things to note here:

  • Derek Jeter is the only player named on every ballot so far.  With Mariano Rivera having finally broken the longstanding trend of non-unanimous voting results, writers have no good reason not to vote for an obvious Hall of Famer like Jeter except spite.  Nobody can stand any longer on the logic of, "If Joe DiMaggio didn't get in on the first ballot then nobody should!"  or, "If Babe Ruth wasn't unanimous then nobody should be!"  That ship has sailed.  There is a precedent now, and it's doubtful any of the BBWAA members want to deal with the backlash from perhaps being the ONE writer who inexplicably refused to vote for Jeter.  

Whatever his shortcomings, and this is not to say that there were none, Jeter checks EVERY box for a Hall of Famer:  He was a 14-time All Star shortstop who hit .310 over a 20-year career on the game's biggest stage, led the team to five World Championships, won five Gold Gloves, five Silver Sluggers, a Rookie of the Year award, and the MVP awards for the All Star game and the World Series.  His postseason career is like another All-Star caliber season unto itself, as he is the career leader in almost everything except homers and RBIs, not the stats you expect from a shortstop who usually hits leadoff.  (He's 3rd and 4th all-time in those, BTW.)  He's the total package, and his defensive shortcomings or perceived overratedness are no reasons to pass him up.  

  • Three voters - Dan Shaughnessy of the Boston Globe, Anthony Reiber of New York Newsday, and former Newsday writer Steven Marcus all voted for ONLY Derek Jeter.  
    • Marcus either did not vote for anyone last year or did not have a vote, as his line shows neither greens nor reds, including the Jeter vote.  
    • Shaughnessy is a well known curmudgeon who's not above making something like this about himself and his own sense of indignation.  Last year he voted only for Mariano Rivera.  
    • The real curiosity is Reiber, who actually had voted for Bonds, Clemens, Pettitte, Ramirez, Schilling and Vizquel last year, but took those away so he could make a statement, I guess.  Maybe in the blinding light of the awesomeness that was Derek Jeter's career he saw that nobody...OK, never mind.  He's an idiot.  
  • Barry Bonds and Roger Clemens are getting closer, but are not there yet.  Both have over 70%, but experience suggests that the writers who do not publish their ballots in advance or email the trackers privately about them tend to slant more conservatively in their voting, so those percentages are likely to drop a bit in the final tally.  Since both finished with about 59% last year, though, they're still poised to make a significant jump.  
    • Curiously, two voters, Jon Heyman and Christina Kahrl, have given the nod to Bonds (a 7-time MVP award winner who ostensibly started using PEDs after having already won three of them.) but not to Clemens (a 7-time Cy Young Award winner who ostensibly started using PEDs after having won three of them.)  Clemens even has the better postseason resume, often something of a tiebreaker for tossups like this.  Heyman's logic was picked apart by Deadspin a few years ago, but as far as I can tell, despite all the questions about her ballot on Twitter, Kahrl has not explained her ballot. 
  • Curt Schilling (78%) and Larry Walker (83%) are the only other players polling over the 75% required for enshrinement at this point.  
    • Walker's vote total, like those of Bonds and Clemens, dropped a little more than 11% in the final tally last year, which would put him slightly below the threshold.  Again, the more old-school voters who don't know (or care to know) how to Internet tend to discredit players whom they think received a disproportionate benefit from their home ballpark.  (Jim Rice excepted, evidently.) However, players often get a bit of a bump in their final year, and this being a less crowded ballot than in recent years, Walker may still make it.  
    • Schilling probably will not, as he's just barely over the threshold to begin with, and if he drops as much as he did last year (about 9%) he'll end up well below it, but still close enough to likely achieve enshrinement in 2021, his 9th year on the ballot.  
  • The only other players polling even close to 50% are Omar Vizquel and Scott Rolen.  This is their third year on the ballot, but they've followed very different paths here.  Vizquel started out strong, with 37% and then 43% last year.  He may even end up over 50% this year, as he and Andy Pettitte were the only two from last year who did better in the final count than the pre-announcement polling had showed.   Rolen, however, started at about 10%, then got 17% last year, so if this result holds - he's currently at 47.7% - it would be a ~30% jump in one year, which is pretty rare, I would imagine.  Evidently there is a pretty serious Scott Rolen is better than you think campaign going on somewhere, and it's working.
  • Several players appear likely to fall off the ballot for not receiving at least 5% of the vote.  In fact, several of them have only one vote to date: Cliff Lee, Eric Chavez and Jason Giambi.  
    • Not that I expected him to get elected, or even think he should be, but it surprises me that Giambi isn't getting a little more support.  He has the sort of resume that might have kept him on the ballot in the Old Days - 440 homers, over 2000 hits, an MVP award (and he probably should have won another) - but players like this have gone once-and-done on the ballot several times recently: Carlos Delgado, Lance Berkman, Mo Vaughn, Andres Galarraga, Jim Edmonds, Paul Konerko (who currently has two votes and is epected to fall off the ballot), etc.  
The unfortunate side effect of the so-called Steroid Era and all the wonderful hitters we got to watch at the time is that we don't know whose stats to take seriously, and the BBWAA tends to err on the side of caution.  Plus, Giambi has his sniveling press conference in which he apologized - sorta - for using steroids.  

And he has recency bias going against him.  That is, the way in which his career just kinda petered out over a drawn out time after his Yankees contract probably hurts him a bit, too.   He hit .212 over parts of six seasons at the end of his career, with about as many homers (44) in his last 410 games - more than half in a Rockies uniform - as he did in his MVP season alone (43).  The voters tend to frown on "padded stats" preferring guys to go out closer to their peak.  Take away those six seasons and Giambi's career (.286 batting average, 396 homers, 146 OPS+ in almost 8000 plate appearances) looks a lot like Duke Snider's or Orlando Cepeda's, but also Frank Howard's and Albert Belle's.  Not a slam dunk or anything, but maybe more than one person would have thought him worthy of a check mark.
  • Two players - Rafael Furcal and Alfonso Soriano - have not received a single vote yet.  Furcal I get: He was a pretty good player for a few years, but didn't amass the counting stats the voters like to see.  His defense bumps his overall WAR total to about 40 (Vizquel has about 45, for reference, in about 10 more seasons) which is impressive for a short career, but just not enough.  
  • Soriano, though?  He got some MVP votes, finishing as high as 3rd one season.  He hit over 400 homers, had over 2000 career hits, had a 40-40 season and just missed a second one by a single homer. You'd think someone would give him a vote.  I mean, Danny Tartabull got a single vote, for crying out loud, and Soriano hit 150 more homers!  Danny Tartabull!  
Anyway, it'll be interesting to see how this goes tonight.  I fully expect Jeter to be unanimous and Walker to just make the cut, but that should be it.  

Next year though, with all of these guys carrying over and nobody who seems particularly like a Hall of Famer being added to the ballot (Mark Buehrle, Tim Hudson, Barry Zito and Torii Hunter are the best of the new nominees next year) it could be Clemens, Bonds, and Schilling giving speeches on the dais in Cooperstown next July.  

It may be a good year to get cheap accommodations and tickets to the ceremonies.  To paraphrase  Yogi Berra, people will be staying away in droves.  

Stumble Upon Toolbar

16 January 2020

How did Robin Roberts Lose the NL MVP in 1952? (Hint: Not the Ladies' Fault)

The latest in Joe Posnanski's series on the top 100 baseball players of all time, over at The Athletic, is Robin Roberts.

Among the other compelling stories he tells about Roberts is this little tidbit:

The Baseball Writers of America gave out the first Cy Young Award in 1956, one year after Roberts’ historic run. So, one of the greatest pitchers ever never won a Cy Young.
He also didn’t win an MVP award, though it’s hard to see how he lost the award in 1952. They gave it Hank Sauer, who led the league in homers and RBIs. Here’s how Oscar Fraley of the United Press International responded to that vote:
“Anybody who knows the difference between a bunt and punt must be completely flabbergasted at the selection of Hank Sauer in the National League. Most of the voters obviously never heard of Robin Roberts … one theory is that they were all on vacation and the ballot was filled in by the editor of the women’s page.”
Yes, there was always time for a little misogyny in 1950s baseball writing!
But the main point was sound: Roberts went 28-7 with a 2.59 ERA, and his last 23 starts the Phillies went 21-2; both losses came when Philadelphia was shut out. By WAR, Roberts was three wins better than Sauer. And it’s not like Sauer played a significant role in the pennant race; his Cubs were mediocre non-contenders.
The rest of the story on Roberts is of course excellent and well worth the pittance you need to invest to break down the Athletic's paywall, but this got me to wondering how that could happen.  How does a pitcher who so clearly outclasses the rest of the league, who wins 28 games for a decent team,* not manage to win, or at least come closer to winning, the MVP? 

* The 1952 Phillies started out just awful.  They were 10 games below .500, at 29-39 as late as June 23rd.  Manager Eddie Sawyer - who had perhaps gotten some grace for finishing 73-81 in 1951 because he had helmed the 1950 "Whiz Kid" Phillies to their second franchise pennant ever, had run out of rope with which to hang himself.  He was fired five days after that, ironically after the team had won 4 of  5 games.  

Steve O'Neill, who had managed the Tigers to the 1945 World Series championship, took over and the team immediately improved, going 59-32 the rest of the way, and finishing "just" 9.5 games out of first.  That doesn't sound so great until you notice that they were 17.5 back when O'Neill was handed the reins.  

This was, interestingly, the third time in his career that O'Neill had taken over a team mid-season and gotten immediate improvement from it.  He led the 1950 Red Sox to a 63-32 record after the great Joe McCarthy was forced to resign, with the team at 31-28, and he led the 1935 Indians to a 36-23 record after Walter Johnson (!) had stumbled to a 46-48 record.  He's one of two managers in history with 1000 or more career wins to his credit whose teams never played below .500 ball on his watch.  The other is, ironically, the great Joe McCarthy.   

Anyway, here is, I think, how Roberts lost the MVP in 1952:

The voting - and my understanding of the reasons for it - was as follows:

#1 Hank Sauer (226 points): Led the league with 37 HR (tied with Ralph Kiner, on the last place Pirates).  Only Gil Hodges (32) was even in the same neighborhood.  Nobody else in the NL hit more than 25 that year.  Also led the NL in RBIs with 121.  Second was Bobby Thompson with 108, only three others had over 100.  Remember the sports writers LOVED RBIs in those days.  He got *8* 1st place votes. This will be important later.   

#2 Robin Roberts (211 points): Had the amazing season noted by Posnanski above.  Led the NL with 8.5 bWAR, tied with Jackie Robinson, who somehow finished 7th.  (Robinson had by then led the NL three of the previous four seasons in bWAR, but nobody knew that at the time, and anyway  the writers tend to want someone new to vote for because it makes for a better story.)  Roberts outclassed all other NL pitchers by nearly two whole bWAR (Warren Spahn was 2nd with 6.6).  Roberts received seven 1st place votes.

#3 Joe Black (208 points), a rookie reliever with the pennant-winning Dodgers.  Went 15-4 with 15 Saves* in 142 IP and led the NL in games finished with 41.  Also won RoY honors.  He, like Sauer, received eight 1st place votes.

  • * The Save Rule was not codified until 1969, and was then applied retroactively, so nobody knew this at the time, but the writers must have been aware of how frequently someone like Black was used to save (lower case) a baseball game.  

#4 Hoyt Wilhelm (133 points), another rookie reliever, this one for the Giants, who finished a close 2nd in the NL pennant race.   They were as close as 3 games back on September 17th but went 4-5 the rest of the way, losing 4 of those games to the Phillies (Two to Roberts!) and finished 4.5 out.  Wilhelm led the NL in games, ERA (2.43 in 159 innings, all in relief), and winning percentage (.833, with his 15-3 record). 

Duke Snider got the other first place vote, though he finished 8th over all.  Nobody on a losing team finished higher than 13th that year.  The writers simply wouldn't vote for players on bad teams, almost ever. 

They would, however, vote for starting pitchers.

In the years before the Cy Young award was instituted, specifically since the sportswriters had been given charge for the decision in 1931 but prior to 1952, starting pitchers had won it nine times, not quite a quarter of the time.  Carl Hubbell and Hal Newhouser had each won it twice.  Bobby Shantz won it in the American League THAT VERY YEAR.  So how did Roberts miss out?  Was it the writers' fault for allowing (gasp!) the ladies' editors to vote Sauer, as Fraley intimated? 


The writers who blamed other writers for electing Sauer were missing the mark, it seems.  Really, it was the writers who voted for relief pitchers who perhaps should have been shamed.

At the time, the relief ace was just becoming a Thing.  Prior to 1950, there had only been two pitchers to appear in 50 or more games and pick up 15 Wins without starting at least five times: Jim Konstanty - who won the 1950 NL MVP for the Whiz Kid Phillies - and someone named Mace Brown of the 1938 Pirates.

Those Pirates were a decent team, finishing in second place at 86-64.  They had a solid pitching staff (3rd in ERA) that lacked stamina (second to last in complete games).  So Brown, who was eminently mediocre (100 ERA+) but apparently perpetually available, vultured off 15 Wins from the starters.  It was more out of necessity than intent that he managed to win so many.  He also lost nine and had a 3.80 ERA.

With Konstanty, it was different.  He was intended to be the relief ace. They pitched him every other game, for a couple of innings, on average, and he led the NL in appearances, games finished and Saves (22), while racking up 16 Wins and pitching 152 innings.  He made the All Star team and won the MVP pretty handily over Stan the Man Musial, picking up 18 first place votes, while nobody else on the ballot got more than two (Granny Hamner and Eddie Stanky).  It was a great story.

So in 1952, when not one but two spectacular rookie relievers came along at the same time, and their two teams finished 1st and 2nd in the NL, with each picking up 15 Wins for them, I guess those guys vultured off a lot of the sportswriters' attention just as they vultured off credit for "Winning" games in which they had pitched two innings while someone else covered the first seven.

And Joe Black, being on the first place Dodgers, got more of the votes than Wilhelm, or, as it would happen, than the preposterously amazing Robin Roberts and his 28 (mostly deserved) victories.  Wilhelm, as a knuckleballer, may also not have been given as much credit as he deserved.  He got zero first place votes for MVP and finished a distant second in the Rookie of the Year voting to Black despite having, on paper, almost exactly the same season.

The MVP voting system follows a curious, top-heavy approach, in which a first place vote is worth 14 points, but second and beyond are worth 9, 8, 7, 6, etc. on down to 10th, worth just one point.  There are three writers for each city with a team assigned to vote, which at the time meant 24 total writers for the eight-team National League.  It's a feature, not a bug, as they want the players writers think are the best to have an advantage over everyone else.

So those eight first place votes that Black received were worth 14*8 = 112 all by themselves. His 96 other points meant that he averaged just six points each from the other writers, or a 5th place vote.  Roberts' seven first place votes netted him 98 points, which meant the average of his remaining voters was 6.65 points, almost a 4th place showing per writer.  But those five extra points for a first place vote are key, and that's where the issue lies.

If three - just three - of the voters who thought that Joe Black was the NL MVP had instead voted for Roberts, he would have gotten at least 15 extra points in the voting and tied Sauer for the award.  And that assumes that the voters who gave Roberts 9 points, for second place, voted for him first instead.  If any one of those who had given him a third place or lower vote changed his mind to recognize the absurdity of voting for Black over Roberts, Roberts would have won the award, as he deserved to do.

Instead, somehow Black received eight first place votes compared to Roberts' seven.  And this for winning slightly more than half the games (15 to 28), striking out slightly more than half as many batters (85 to 148) and walking almost as many (41 compared to 45) despite pitching far less than half of Roberts' innings (142 to 330).  Just bizarre.

Black, you may think, led the league in Games Finished! Which meant he was often on the mound for the deciding moment at the end of a game!  It's the timing, is what it is!

Which is fair...except that Roberts led the NL with 30 COMPLETE games, and also pitched in relief, finishing two other games.  Which means of course that he actually finished 32 games overall, while also starting 30 of those.  There is just no plausible way to suggest that Black was anywhere near as valuable a pitcher as Roberts was in 1952.

On the plus side, according to most sources, Roberts was such a humble and pleasant man that it probably never bothered him nearly as much as it bothers me. 

Stumble Upon Toolbar