Remember me

Winfield and Evans

June 25, 2020
Dave Winfield was born on October 3rd, 1951. Dwight Evans was born a month later, on November 3rd. Same year, if you skipped the class on how calendars work.
There is scant information about what kinds of baseball players they were in elementary school.
Dwight Evans was drafted out of high school, selected by Boston in the fifth round of the amateur draft. He was the 109th player selected. Dave Winfield, a St. Paul native, decided to attend the University of Minnesota on a full baseball scholarship. Winfield was drafted, way down near the bottom of the draft, but it was a flier in case he remembered what winters were like in Minnesota.
Dwight Evans had a strong Triple-A season in 1972, winning the International League MVP award. Dave Winfield, an elite pitching prospect for the University of Minnesota, was the MVP of the 1973 College World Series.
Just to break into the narrative a little: Winfield pitched a shutout in the first game of the College World Series playoffs, striking out fourteen Oklahoma Schooners. He started the semi-final against defending champs USC, and dominated again…through the first eight innings Winfield had struck out fifteen batters while allowing just an infield single. The Gophers built up a 7-0 lead… you know where this is going.
The first batter of the 9th singled, but Winfield coaxed a double-play grounder from the next hitter. The umpire at first missed the call and the runner was called safe. And then, in spectacular fashion, the wheels came off the Golden Gophers’ bus. The wheels and axles and most of the seats. Winfield was relieved after a few more batters reached and sent out to play left, where he watched USC come all the way back, winning 8-7. Fun way to cap off your college career.
Winfield was the drafted fourth overall by the Padres in 1973. He was also drafted by an NBA team, and an ABA team: he was a star cager for a team that won the Big-Ten championship. And…just for kicks…the Minnesota Vikings drafted him in the seventh round. Winfield hadn’t played college ball, but what the heck. I think the national bobsled team tried to sign him up, but he didn’t fit in the sled.
Winfield jumped right to the major leagues…not as a pitcher but as a leftfielder. By the summer of 1973, Evans and Winfield were both starting outfielders in professional baseball.
This was at a moment where there was a lot of uncertainty about what the best path to success in the majors was. College and the minors were in a kind of war to decide which one would shake out to be the better road to the majors, with advocates on both sides. Winfield was a great college player, and Evans was a great minor-league player: no one really had any real idea which one was more likely to succeed.
Winfield succeeded first, out-hitting Evans in 1973 and popping twenty homers in 1974, his first full stint in the majors. Winfield made the All-Star team in 1977, finished 10th in the NL MVP vote in 1978, and probably deserved to win the MVP in 1979, when he paced the NL in RBI while hitting .308 with 34 homeruns.
Dwight Evans made the 1978 All-Star team, but he was the third-best outfielder in Boston, a defense-first player eclipsed by AL MVP’s Fred Lynn (1975) and Jim Rice (1978).
I am as big a fan of Dwight Evans as there is in the world, and I strongly suspect that Evans would have broken out as an elite hitter earlier than he did, but for a Mike Parrot beanball that sidelined him at the tail end of 1978. Rushed back too quickly and likely suffering from a concussion, Evans hit .161 down the stretch with diminished defense as Boston punted a pennant lead to the Yankees.
Evaluating them thought 1980, because 1981 would be a transitional year for both men:
Evans had the lead in Gold Gloves, three to two, but Winfield had four All-Star appearances to Evans’ one, and Winfield had placed 10th and 3rd in consecutive MVP votes. Evans hadn’t ever received an MVP vote.
And the raw numbers undersell the differences between them. Both Evans and Winfield were right-handed hitters, but they were right-handers in very different parks. Dwight Evans benefited tremendously from the friendly confines of Fenway Park, which had a multi-year park factor for hitters of 112. Winfield endured the much tougher San Diego Park, which had a factor of just 89 for hitters.
The fans and writers at the time understood this. Dave Winfield was a star player: he was viewed as one of the game’s best player, right there with Dave Parker and Jim Rice and Mike Schmidt and Tom Seaver. Dwight Evans was a good player, but he was no one’s idea of a star.
And then things changed. Strike-year, everything has to change.
Dave Winfield, a free agent, signed one of baseball’s greatest blockbuster contracts prior to 1981: a ten-year deal with the Yankees for $23 million dollars. This was the highest contract ever given to a player, and no one thought Winfield was the wrong guy to get it. If you’re going to sign anyone to a ten-year deal, Winfield’s a decent bet.
Signing with the Yankees meant that Winfield was going from one of the toughest park for a right-handed hitter in the NL for one of the toughest park in the AL. That didn’t matter: hitters hit.
Up in Boston, Dwight Evans would see the Red Sox lose stars Fred Lynn and Carlton Fisk on some late-mailed contracts, an event that increased Evans’ visibility on the team. Dewey capitalized on the chance, posting an out-of-nowhere, MVP-level season. Tying the AL lead in homeruns, Evans paced in total bases, walks, and OPS, and finished third in the AL MVP vote.
Dewey followed that breakout with a 32-homer season in 1982, a 1984 season that saw him pace the AL in runs scored and OPS while collecting his first 100-RBI season, and then a three-year kick to close out the 1980’s with century marks in RBI, along with his only +.300 season as a hitter (.305 in 1987). Dewey would win five Gold Gloves before shifting to first base in 1987. Evans jumped from a defense-first ‘good’ player to a legitimate offensive powerhouse.
Winfield continued his efficiency, receiving MVP votes every year between 1981-1985 and also in 1988. All this while trying to avoid wrath of Yankee owner George Steinbrenner, who was hell-bent on making Winfield the goat for the Yankees playoff failures. Winfield popped a career-high 37 homers in 1982, nearly won a batting crown with a .340 mark in 1984, and made every All-Star team through 1988. He won five Gold Gloves, along with one citation for murdering a seagull. That counts as a capital offense in Canada.
Winfield missed the entire 1989 season with back problems. Despite this, Evans and Winfield were comparable players through the decade. 1981-1989:
Jumping to the 1990’s…
Dwight Evans didn’t last into the 1990’s, getting one last season in Boston as a DH for a surprise Red Sox team that won the division and then got buried by Oakland in the ALCS. He followed that up with an effective platoon year in Baltimore, posting an on-base percentage of .393 during his last season. Then he was done.
Winfield came back from a year on the DL with something to prove, and he proved it. Trading pinstripes for an Anaheim halo, he had a couple solid years with the Angels before signing a free-agent deal with the Blue Jays. At forty, Winfield had a surprisingly excellent season, hitting .290 with 26 homeruns and 108 RBI for the champion Jays, while finished fifth in the AL MVP vote.
Winfield followed that with an inevitable homecoming in Minnesota, collecting his 3000th hit for the Twins in 1993 before finishing out the string with Cleveland.
1990’s totals for both men:
And their final tale-of-the-tape:
Gold Gloves
All-Star Gms
AS Starts
MVP ballots
*            *            *
Wait…why am I doing this?
Dwight Evans is my favorite player. As my favorite player I am willing to entertain arguments about his greatness that I might not stake for anyone else. I believe Evans was a greater outfielder than Rice or Lynn. I think he was better than Parker. I think he deserved the 1981 MVP. He is a deserving candidate for the Hall-of-Fame.
When it comes to a comparison with Dave Winfield, there is a reasonable argument that Dewey was a more valuable player than Dave Winfield during Evans’ peak in 1980’s. It’s very close, but I can entertain that notion. Evans had an extra year, after all, and he did walk a bit more often than Winfield.
But if you consider their careers in total, Dave Winfield was a superior player. Winfield was a star in the 1970’s, playing in a gigantic ballpark, while Evans was still trying to get on his feet as a hitter. And Dave Winfield was still productive in the 1990’s, posting All-Star seasons as a DH-for-hire after Evans had left the game.

Their batting numbers are a little deceitful. Just looking at their career home runs and triple-slash lines:
What this misses is that Evans was significantly advantaged by playing in a favorable home park, while Winfield was significantly disadvantaged by his home fields. Both players at home: 
In more neutral road contexts, Winfield’s superior hitting shows up in better contrast:
While Evans retains a slight edge in on-base percentage, Winfield gains thirty-points in batting average, and fifty points in slugging percentage. Remove their park contexts, and Winfield’s triple-slash line comes into parallel with the many other places where he shows as being a better player than Evans.
Switch their parks…imagining that Dave Winfield spent his career in Fenway while Dewey bounced around between San Diego and the Bronx… it is very likely that Winfield would’ve reached 500 homeruns. Stuck in bigger stadiums, I don’t know that Evans would’ve even reached 300.
Winfield was a perennial All-Star almost from the start, and he was paid like a star. Evans was a good defensive player who turned into a very fine hitter later in his career, but he wasn’t ever viewed as one of the elite players in the game. Winfield stole a decent numbers of bases in his younger years; Evans never crossed double digits. Winfield nearly won a batting title; Evans never came close. Winfield collected 3100 hits and 465 homeruns; Evans never reached 2500 hits or 400 career homers. Winfield hung around forever and was still a force at 40; Evans dropped off quicker. Winfield was elected to the Hall-of-Fame during his first year of eligibility; Evans fell off the ballot after three years, never gaining any traction with the voters.
There is a strong consensus among the people who watched both men play that Dave Winfield was the greater player. That consensus is supported both by the broad arc of both of their careers, and by the statistical record they left behind.
*            *            *

And then there is WAR.
WAR tells us that Dwight Evans was a better player than Dave Winfield.
Understand, first, that I very much want this to be true. I would very much like to rest in the knowledge that Dwight Evans was a greater player than Dave Winfield. I’m pleased at the reconsideration Evans’ career has received: he has gone from extremely underappreciated to a-little-appreciated to very-likely-to-get-elected to the Hall. I have been incredibly happy to witness that movement.
But these numbers are broken. They’re wrong. They do a disservice to both players, forcing an unreasonable comparison, and they do a disservice to the field of sabermetrics.
Dwight Evans wasn’t a better player than Dave Winfield. Considering the totality of their careers, it wasn’tclose. I want Dwight Evans’ career to be lauded as much as anyone, but I don’t want it this way. This is bullshit hidden in numerical chaos.
So let’s dive into that chaos.
The gap inWAR lies in the way the metric measure of both players’ defensive abilities. Baseball-Reference counts Dwight Evans as a slightly negative defensive player, and Dave Winfield as a more-than-slightly terrible defensive player. FanGraphs goes further: Dwight Evans is counted as a pretty darned horrible, and Dave Winfield was what you’d get if you made Adam Dunn play outfield in a blindfold.
Here’s how FanGraphs credits Winfield and Evans:
‘Off’ – short for ‘Offense’ – is an attempt to combine a player’s context-neutral value at the plate and on the bases. It is a combination of park adjusted batting runs above average. Dwight Evans, according to the math, was worth a context-neutral 348.4 runs as an offensive player.
I have some quibbles with ‘Off’, but it’s not an awful metric. It isn’t perfect, but it isn’t terrible. If you want to tell me that Dave Winfield’s value over Evans as offensive players is somewhere in the neighborhood of 58 runs above average…well, that feels a little light to me, but I don’t need to die on that hill.  
But ‘Def’ is terrible…it is a flawed, damaging statistic.
Def stands for ‘Defense’, and it is a metric that tries to do two things: calculate the number of runs a defensive player saved against the league average (Mookie Betts is a better right-fielder than Bryce Harper, for example), and make a positional adjustment on the value of those runs (a shortstop has more defensive importance than a right-fielder).
This is juggling a chainsaw, and then adding a beaker of anthrax. It is, from a statistics perspective, a risky move.  
The first component is measured by a stat called Total Zone, or TZ. This is an attempt to measure a player’s fielding runs above average.
TZ credits Evans as being +80, or eighty runs above average. And TZ credits Dave Winfield as being worth -94 runs. Ninety-four runs below average.
Let us stop right there and ask: does that make any kind of intuitive sense? Does that feel logical, in any way?
It feels logical for Evans. I don’t know what +80 looks like in context, but Evans was a good defensive player, who set a bunch of records while playing in a tough park for a right-fielder. He received 8 Gold Gloves and had a canon of an arm…if the number for that is +80, I can buy it.
But how in the world is Dave Winfield so significantly worse? Winfield isn’t just zero: the evaluation of his defense goes from Evans-to-zero-to-Anti-Evans. How is that possible?

Winfield won seven Gold Gloves. Winfield played in big outfields and had a cannon of an arm. Winfield was a gangly player, as tall people often are, but he was fast and tall and he could climb the fence.
Here are the ten outfielders with the most assists since 1961. Both men make the list:
R. Clemente
Barry Bonds
Jesse Barfield
Jhn Callison
Tony Gwynn
Dwg Evans
Larry Walker
K Griffey Jr.
Roberto Clemente, missing seven years of his career, is still ahead of the pack. Dave Winfield, credited with having a terrific arm, ranks third. Evans ranks seventh.
This seems like a perfectly reasonable list. Barfield is there: that man had a canon. Bonds is a surprise, but Gwynn could throw. Walker was incredible in his Montreal days. This is a good list.
Evans made thirty-five fewer errors than Winfield, in 400 fewer games. Evans notched a few more double plays. Let’s imagine that Winfield makes an extra error a year, and Evans nets an extra double play every other year. Let’s also assume that Evans got better jumps on the balls, and ignore that Winfield was tall enough to snag a few more dingers. I’m fine with all of those allowances.
Does that in any way justify a gap of +80 to -94? Of course not. Whatever gap on defense exists between Evans and Winfield, it’s not that big. It can’t be, given all that we know about both men.
And is that gap so significant that it makes up for Winfield’s edge as an offensive player? Does that difference on defense make up for Winfield’s extra eight seasons as an elite hitter?
Again, of course not.
Think it through. Just use your head.
How many more runs, on offense, was Winfield worth in 1978, when he hit .308 with 24 home runs and 21 steals in a Death Valley park, while Evans was hitting .247 in Fenway? How many more runs was Winfield’s bat worth in 1991 when he 28 homers in Anaheim to Evans’ six in Baltimore? What about 1992, when Winfield drove in 108 runs for the Jays while Evans was out of baseball?
And then there is the positional adjustment. Certainly, you have to adjust for position. A shortstop who creates 100 runs with a bat is a rarer thing than a right-fielder who creates 100 runs. I understand.
But this is ridiculous:
If your combination of defensive runs counting and positional adjustment negates most of the impact of an elite hitter, and when that negation flies in the face of a broad consensus of observers and analysts, you’ve got to admit you made a mistake.
Dave Winfield might’ve been an overrated defensive player, but there is no way that his defense was so terrible that it knocks out more than half of his value as a hitter. There is no way his defense was so bad that is closes the gap.
*            *            *
I do not generally attempt to write anything that risks venturing in an area that Bill is already writing about. This is because I don’t want to make the mistake of imagining I know what the heck Bill is thinking. If he wants to talk about something like ‘Runs Saved Against Zero’, I’m going to assume that he knows something I don’t know, and then I’m going to back out of the room slowly.
But I’ve been meaning to write this article for a while, and Bill’s article reminded me about it.
The two versions of WAR both conclude that Dwight Evans was a better player than Dave Winfield.
That conclusion rests, as far as I can tell, on a defensive measurement that is not calibrated to zero. It is calibrated to one thing that isn’t zero (defensive runs above average), and then it is calibrated again to something else that isn’t zero (positional adjustment). Those acts of calibration lead, in the case of Winfield and Evans, to a conclusion so improbable that the inventors should have recognized a flaw in their system and made an adjustment.
They didn’t. They haven’t. They’ve dug their heels in and let it stand.
Dwight Evans was a great baseball player. He was better than anyone thought at the time, and it is a great fact that his career is getting more attention than it received during his playing time. As a member of the sabermetric community, I am happy about it. As a Dwight Evans fan, I am over the moon about it.
But that re-evaluation shouldn’t happen on an architecture of bad math. Dwight Evans was a great player, but he wasn’t a greater player than Dave Winfield, and the comparison doesn’t do him any favors. It doesn’t do our community any favors, either: if our best statistic wants to tell us that Dwight Evans was a better player than Dave Winfield, we have work to do.
David Fleming is a writer living in western Virginia. He welcomes comments and questions here and at 

COMMENTS (71 Comments, most recent shown first)

Looking back at this article, let me summarize Winfield's fielding career as seen by Rfield (the portion of dWAR that compares him to others at his same position):
Ages 21-28, the Padre years - 13 runs better than average
Ages 29-36, Yankee years before the back injury - 58 runs below average
Age 37 - out all season with a back injury
Ages 38-39 - 40 runs below average in just two seasons! Maybe not a good idea to play him in RF still.
Ages 40-43 - 6 runs below average because he was mostly at DH.
That's how we get to a total of 91 runs below average for Winfield's career. It seems a good deal more plausible when we break the narrative down into stages like this.
4:28 PM Feb 23rd

Winfield, imo was vastly overrated as a defensive player mainly because of his incredible ability to rob hitters of home runs with his great leaping ability (and height of course) at the wall. However I cannot count the numerous times he was let I believe fly balls that could be caught drop in front of him for singles.

He had horrible range. You would figure a man who once he got going had excellent speed would have better range in the outfield but he did not. He got bad jumps on the ball. You could finish a book by the time Winfield got started going for the ball in the outfield.

12:16 PM Jul 3rd
Thank you, Charles
3:53 AM Jul 1st
I did the same analysis for Dwight Evans. Evans had 192 more putouts than expected. Prorated to Evans's innings, Evans's teams gave up 148 more singles to RHB but 56 fewer singles to LHB, 60 fewer doubles to RHB but 106 more doubles to LHB, and 31 fewer triples to RHB and 56 fewer triples to RHB. So while the Red Sox (and 523.1 innings as an Oriole) didn't have good defense either in this era, they weren't as bad as the Padres/Yankees/Angels of Dave Winfield, and the issues were more concentrated in the middle due to the heavy concentration of singles. Evans was +33 runs through 1981, -12 afterwards, +21 overall. Given about 10 runs from good fielding percentage, this is about what Total Zone has for him career as an outfielder (28-29 runs) before you add in his arm.

I'm sure Dave Fleming has moved on from this discussion, but if he hasn't, let me put it this way to him: Yes, traditional defensive statistics have their issues. But set aside that Dave Winfield was -206 putouts as an outfielder, career, while Dwight Evans was +192 putouts, career. Dwight Evans's teams weren't great at preventing hits, giving up 51 more hits than you'd expect, but Dave Winfield's teams gave up 470 more hits than you'd expect. Dwight Evans's teams gave up 41 fewer 2B/3B than expected, while Dave Winfield's teams gave up 229 more 2B/3B than expected. That's a lot of hits and bases, and somewhere in the order of 300 runs of difference. Don't you think WAR should account for those runs somewhere?
8:46 PM Jun 30th
@Brian: yeah, what MWaddell said.

I made a small error when inputting Winfield's team batting stats affecting RHB/LHB totals, and upped the run value of hits and outs slightly to accommodate for preventing a hit not allowing another batter to come to bat (I'm sure Tango will be pleased, though I haven't checked his blog since yesterday morning). The upshot is Winfield is a mere 206 PO below average, but is also now 79 runs below average.

Regardless, the basic issues remain: Winfield didn't make many PO given his team statistics, and, more importantly, his teams were bad defensively. Not all of that was Winfield, of course—as an example, his teams allowed an excess of 132 singles to RHB, a stat that doesn't have a ton to do with the RF but isn't inconsequential to them either—but he definitely had something to do with this. They were 470 hits worse than expected, prorated to Winfield's time afield, and that's way too many hits to ignore, especially given that Winfield also wasn't making outs relative to the other fielders.

Unless someone crunches the hit location data that exists for most of Winfield's games but isn't tabulated and shows some weird anomaly, I think it's pretty safe to assume that Winfield was not a good outfielder. His arm likely won him those Gold Gloves; the Gold Glove voters always did like the guns.
7:04 PM Jun 30th
Brian, pitchers do have largely persistent abilities to influence the groundball / flyball balances, and flyball are more likely to produce extra base hits than groundballs are.
10:23 AM Jun 30th
Question that comes to mind from this discussion: I know that the prevailing wisdom based on studies is the a pitcher's role in batting average on balls in play is very limited. Is anyone aware of any studies that would show the pitcher's role in extra base hits on balls in play?
8:01 PM Jun 29th
MichaelPat: you have to be careful comparing RF (or putouts) by today's players to those in Winfield's time. Rising K rates have changed the benchmark. Cody Bellinger's 2.06 RF is well above league average (1.98), while Winfield's superficially better RF of 2.13 was slightly below average (2.15) for his time.
1:02 PM Jun 29th
@MichaelPat: Winfield's Range Factor is only a part of it. First of all, it's poor after adjusting for strikeouts, groundouts, and opponents' batting hand. That's not going to be true of all of your list.

Second, Winfield played on subpar defensive teams. The Padres were comically bad while Winfield played for them, to the point of actually being a punchline in Mad Magazine (IIRC). My quick estimates have the Padres giving up an extra 161 hits while Winfield was playing. The Yankees were better, only giving up an extra 87 hits while Winfield was playing, but even then unusually bad for a generally good team. (Same story with Derek Jeter's years with the Yankees, especially with Joe Torre managing.) Overall, Winfield's teams gave up 282 extra hits, prorated to Winfield's playing time. That's an extra 19 hits a year.

You have to start with the assumption that the fielders on bad defensive teams are bad fielders. From there, you can use things like Range Factor to move estimates off of the starting assumption. In Winfield's case, we start with his teams being a bit below average defensively, then look at Winfield's own defensive statistics. The big one, Range Factor, is below average, even when taken in context, thus lowering our rating of him even further. For example, Winfield was 40 outs below average for the 1980 Padres, which were 8 hits below average adjusting for Winfield's playing time. By contrast, I assume that the shortstops, who were about 35 assists above average, were pretty good, negating the knock the hits would give them.
8:44 PM Jun 28th
Am I right to say that Winfield's poor defensive showing in WAR is due almost entirely to his relatively poor range factor?

I looked at the BRef 2019 data for right fielders who started 80 or more games, and it leads me to question whether range factor alone is enough to properly evaluate Winfield.

The average range factor per nine innings in 2019 was 2.06.(43,423 innings, 9,953 chances)

Here are six right fielders who had a range factor lower than that, yet rated positively on Total Zone's total fielding runs above average, Rtot. Rtot (correct me if I am wrong) is a broader analysis of fielding results, using details that we just don't have from Winfield's playing era.

Name GS Rtot RF/9
Bryce Harper 152 7 2.03
Randal Grichuk 83 5 1.95
S. Piscotty 90 5 1.92
Christian Yelich 121 3 1.91
Cody Bellinger 102 2 2.05

This suggests to me that analyzing Winfield's fielding prowess solely on the basis on range factors may well be missing the mark.
Add to the argument the man did win seven Gold Gloves, albeit during an era when range factors were given the same wight by most GG voters as the effects of passing marsupials on balls in play.

Here are the all the RFs who started 80 or more games ranked by range factor....

Name GS Rtot RF/9
Franmil Reyes 86 -11 2.27
Mookie Betts 126 8 2.25
N. Costellanos 136 -7 2.21
M. Conforto 110 0 2.21
Kole Calhoun 146 4 2.2
Aaron Judge 90 20 2.14
Adam Eaton 137 6 2.09
Cody Bellinger 102 2 2.05
Bryce Harper 152 7 2.03
Nomar Mazara 100 -5 2.01
Josh Reddick 106 -1 1.97
Randal Grichuk 83 5 1.95
Dexter Fowler 80 -2 1.93
S. Piscotty 90 5 1.92
Christian Yelich 121 3 1.91
Yasiel Puig 143 -2 1.85
Nick Markakis 98 -2 1.83
C.Blackmon 134 -14 1.81
Adam Jones 117 -6 1.81
Avisail Garcia 89 -3 1.72

8:22 PM Jun 28th
I went ahead and used the 3-year component park adjustments from Seamheads, and not surprisingly (San Diego Stadium and Yankee Stadium were both pitchers parks), Winfield's rating goes down to -59 runs, career. That's -43 RF, -9 LF, -7 CF, or -24 Padre, -25 Yankee, -11 Angel/Blue Jay/Twin.

@Brian: This rating is only 55% his putout totals, using about 0.14 runs/PO. Prorated to his innings, Winfield's teams had 15298 singles allowed, 3698 doubles, and 580 triples. Based on park and league averages, they should have allowed 15209 singles, 3540 doubles, and 545 triples. The rating also weighs these three kinds of hits by lefties and righties differently for each position. Winfield played for bad defensive teams and didn't make a big percentage of the outs on them.
12:04 PM Jun 28th
If the primary source of Winfield's low rating is his low putout rate, isn't it a little bit like rating a batter based on batting average versus slugging percentage? There are various references to Winfield playing deep, having a great arm, and anyone who has seen him play can attest to his height and jumping ability. It a skill set coupled with positioning that is likely to cut off extra base hits. He played in mostly large outfields during his career and between playing deep, having good speed and that powerful arm you would expect that he was reducing the amount of doubles and triples hit against his team. Also, at least in he early 1980s, he had a good deal of success leaping over the wall and preventing home runs. Even 3-4 home runs saved per year could save between 5 and 10 runs.

Now I would also think that his playing deep, combined with a slow first step or two before he could get to full speed would lead to a below average putout total. That is borne out by his stats. He might even had been a better overall defender if he did play a shallower outfield. I am not saying he made the right choice. But is he getting credit for the fact that the putouts he does make and the runners he is holding at first base with his arm are likely to save a significant amount of bases versus the average outfielder?
10:22 PM Jun 27th
@ Charles Saeger. Thank you for your work.
6:35 PM Jun 27th
Addendum/correction: would you believe that Winfield's assist rate was below average? I tried computing it two different ways (per PO-SO, per H) and yeah, it's below average, though (accurately, IMHO) shows Winfield as having an above-average arm in RF. It goes to show how important having the advances is when evaluating the arms.
6:14 PM Jun 27th
I broke down and said to hell with it, and computed Winfield's runs saved without any park adjustments. This is just based on Hits (broken down Singles, Doubles, Triples by LHB and RHB) and Putouts. I used 1/7th of a run for all PO (which is what it was for both LF and CF). Runs saved from hits varies for each position. For example, if a LHB hits a triple, I divided by 10 to assign any runs resulting from this to the LF, by 3.5 to the CF, and 4.5 for the RF. This is from a massive project I undertook months ago with the Statcast data from 2017-2019, neutral defensive alignment, the field broken down into 6° segments, breaking apart GB/LD/FB/PU … it was complex as hell. I'm not 100% sure on the results, but they're pretty good for the most part.

Anyways, doing all this says Winfield is -35 runs as a RF, -7 as a LF, -6 as a CF. As a Padre, he's -17 runs; as a Yankee, he's -19 runs; as hitter for hire, he's -11 runs. His fielding percentage in RF is above average, in LF is below average, in CF is pretty much average, but errors by outfielders aren't costly events. Obviously he had a good arm; the traditional defensive stats don't do a good job a on outfield arms, but his assist rate was above average.

A note about Total Zone pre-1988: it allocates blame for hits by hitter, based on the hitter's out distribution. That is, if Eddie Murray made 6% of his outs to 2B, then the 2B gets blame for 6% of Murray's hits. This works alright for infielders since hits and outs on the ground are pulled at about the same rate. However, this is not true for hits and outs in the air—batters tend to pull hits in the air but send outs in the air to opposite field (though the tendency isn't strong on the latter). This causes all kinds of problems for corner outfielders, so I'd suggest taking their totals under TZ pre-1988 with a grain of salt.
4:31 PM Jun 27th
I just looked at the putouts for Winfield's whole career in the outfield; I ignored his scant time in the infield. I analyzed adjusting for team assist rate, number of outs by lefty batters (who make more outs to left in the air, since outs in the air are not pulled), and the hitting statistics of Winfield's team (meaning, the 248 PO the 1974 Padres batters made to LF are not counted in the league average). This is not analyzing the team defensive quality; I'd have to put park adjustments in my spreadsheet for that. I don't have park adjustments for the number of outs at each position; I'd love to have them.

In right, Winfield had 3657 PO when he should have had 3840.
In left, Winfield had 838 PO when he should have had 855. A lot of his time in left came early in his Yankee career and I would think there's a big park adjustment involved, but the Yankee batters made more PO to LF than their left fielders made in LF in that time frame.
In center, Winfield had 485 PO when he should have had 511.

As a Padre, Winfield had 2248 PO when he should have had 2237.
As a Yankee, Winfield had 2250 PO when he should have had 2421.
After his time with the Yankees (Angels/Blue Jays/Twins), Winfield had 482 PO when he should have had 548. By 1991, it was apparent to everyone that Winfield couldn't play the outfield anymore, so this isn't surprising.

One thing that really sticks out is that Winfield was a regular outfielder on a team that faced anywhere close to a normal number of left-handed batters only twice: the 1974 Padres and the 1985 Yankees. (I'm defining this by saying that within 100 AB-H-K of what you'd expect given league average opposition is normal.) These were always below expectations. I wasn't surprised to find this as a Yankee, given how much Yankee Stadium was a left-handed pitcher's park, but it happened as a Padre and Angel, too.

So, Winfield recorded 4980 PO in the outfield over his career and an average outfielder in the same conditions would have recorded 5206, a shortfall of 226 PO.​
1:51 PM Jun 27th
What seems to be missing from this conversation is the recent focus of Bill's articles about Defense. Whether his approach will prove to be better or not is unknown at this point. But, much of the approach is driven by questions he has
about the DEF part of WAR and the fact that it is average based. And, results that are wildly hard to accept for many. Without going into many examples that can drawn of these types results and discussed ad nauseum. The dispute mostly comes down to chalking it all up to bias perception and being overrated, underrated with the model numbers and the make perfect sense for the large majority cases. This can be a bit hard to swallow for many and especially in light of observer evidence over a long period of time. It certainly is worth further investigation and different approaches. I mean it could be that the spread is too wide.

A separate question, not entirely unrelated to this point is Phil Birnbaum's study of park factors perhaps being to extreme
12:25 PM Jun 27th
But the Defense + Position Value *doesn't* negate half of Winfield's value. It moves Winfield's value halfway towards *average*. A large share of Winfield's value was the difference between "average" and "replacement", and that value isn't touched by his Defense + Position Value.

Baseball Reference lists Winfield's career WAR as 64.2. It lists his dWAR (Defense + Positioning) as -22.7. I know that their WAR isn't simply oWAR + dWAR, but as a rough estimate, I might guess that his WAR with average defense/ positioning would be 86.9.

So his below-average defense at a less-demanding-than-average position negates ONE QUARTER of his value, more or less.

On Fangraphs, I'm gonna simplify and say that 10 runs = 1 win. His WAR is 59.9. His Defense + Positioning is 24.4. With average Defense + Position, he'd have a WAR of 84.3. In that case, because he's a below-average right fielder (and eventually a DH) instead of, say, a good right fielder or a decent third baseman, he loses JUST UNDER 3/10 of his value.

That's not ridiculous, in my opinion. "Half" would be, but that's not what's happening.
11:25 AM Jun 27th
malbuff. Thank you, I'm not a lawyer but your direction to the legal statement mentioned is interesting food for though and sounds reasonable.
10:45 AM Jun 27th
"Dave Winfield might’ve been an overrated defensive player, but there is no way that his defense was so terrible that it knocks out more than half of his value as a hitter. There is no way his defense was so bad that is closes the gap."

That's the key observation here. We have a set of numbers that indicate Dave Winfield's defensive value negated 60% of his offensive value over the course of his career.

In law we have a rule called the "reasonable person" standard. It's not a faith-based statement, nor is it an expression of populist rhetoric, to say the application of these numbers present a value that is simply unreasonable on its face.

9:09 AM Jun 27th
About the leaderboard that ksclacktc posted in these comments below after which he concludes that Fangraphs WAR concludes that Winfield was the worst MLB defensive player during 1960-2000:

1) Realize that Winfield's Def of -243.9 runs consists of two components that are combined for presentation purposes. One is a career position adjustment of -150.9 runs and the other is -93 runs compared to the average fielder at Winfield's positions. It is really just the second figure that is evaluating Winfield himself.

2) The post says that 1960-2000 was a randomly selected period but of course that is not true. It includes all of Winfield's 1973-1995 career and results in leaving out portions of other players' careers. This doesn't improve the results (from the perspective of pro-Winfield posters) much, but a more fair leaderboard is career DEF for players whose careers partially overlapped the 1973-1995 time period. That leaderboard has:
Gary Sheffield - negative 300.9
Manny Ramirez - negative 276.6
Frank Thomas - negative 267.4
Dave Winfield - negative 243.9
So 4th worst, not worst.

3) So the top four on the leaderboard all were Hall of Fame caliber players with long careers who didn't play at premium defensive positions. DEF is in some ways a counting stat. Winfield had a 23 year career (or 22 omitting 1989 when he was injured) so he had many years to accumulate a negative DEF. If one were to compute a leaderboard of DEF per career games played, I doubt that Winfield would be ranked among the very worse. I haven't done the work to generate that leaderboard however.

4) If we look just for fielding runs for players whose careers overlapped with Winfield (so we comparing players to the average of whatever positions they played but this is still a career long stat, not a rate stat), the worst defensive players whose careers at least partially overlapped with Winfield's career are:
Gary Sheffield
Manny Ramirez
Bernie Williams
Derek Jeter
Ricky Gutierrez
Bobby Bonilla
Danny Tartabull
Juan Samuel
Frank Howard
Rick Monday
Dick Allen
Bill Madlock
Chris Gomez
Jeff Burroughs
Howard Johnson
Mariano Duncan
Toby Harrah
Jorge Orta
Dean Palmer
Bobby Murcer
Gerald Laird
Dave Winfield
So Winfield is now 22nd worst among players whose careers overlapped with him. We're still looking at a career long counting stat so the truly terrible fielders tend not to be here because their careers wouldn't have lasted long enough accumulate more than the negative 93 fielding runs that Winfield had.

I think this data presentation is more fair if we're trying to assess how Fangraphs' version of WAR evaluates Winfield's defense, and not just exaggerating the situation for debating purposes.
8:46 AM Jun 27th
Marc asked a question a few comments ago. Yes, I believe that WAR treats defense (including pitching) as equivalent to offense). They are both initially measured against average and then adjusted down to replacement level.
8:00 AM Jun 27th
What this discussion says to me is that, before Statcast, we don't know nuthin.

There are multiple reasons why Dave Winfield might make fewer plays than the average. One is that he gets to fewer balls, whether because his range is poor or he has been positioned badly or something about the ballparks he played in. One is that he had fewer opportunities to make plays, whether because of the pitchers he played in front of or the other outfielders he played with or something about the ballparks.

It's tempting to say that over a career as long as Winfield's, the factors not directly a function of his ability will even out. Might even be true, though I don't see how you could prove it.

Given the uncertainties of purely statistically based analysis of defense, it is pretty near a mortal lock that the numbers will misrepresent the abilities of some percentage, probably a small one, of players. Winfield might be one of them. Or he might not.

This is why we are now in a new era of defensive evaluation. Long live Statcast!
7:51 AM Jun 27th
4 pieces of wisdom I was taught by my high school math teacher, one of the smartest persons I've ever met, when it came to getting through life:

1) there are no absolutes about anything in life

2) no matter how much you prove something, there is always going to be some outlier that disproves it

3) if you don't believe piece of wisdom no 2, go back and read piece of wisdom no 1.

4) sometimes sh*t happens. don't stress it, because it it's not killing someone, or not saving someone's life, it's not really as big a deal as you think it is.

4:11 AM Jun 27th
The Win awards book gives Evans a B- and Winfield a D+.
9:44 PM Jun 26th
ksclacktc lists for us "All Defensive Players 1960-2000 (randomly selected time period)". And immediately presents us with a list of people widely acknowledged to be bad defenders. Plus Dave Winfield -- but the rest of the list is a strong argument *for* the broad legitimacy of Tango's data.

That Winfield has the highest negative, relative to average (not to replacement value) is a function of playing time. He's getting positional penalties every year except maybe the part-time center-fielder year, and he got them over quite a few more games in the field than most of the rest of the list. Many of them while in his late thirties.

We have seen arguments that the data could be unfair to him: that perhaps he was paired with a lot of really fast center fielders, who took putouts that he would've gotten to just fine. True? Maybe!

We have seen arguments that the data could be entirely fair. Position is *huge*. First step is a big deal. If he was positioned badly for most of his career and took a lot of wrong first steps, it's not hard to imagine that he would make lots of impressive plays while also missing a lot of balls that a different right fielder would have gotten to. True? Even Dave seems to suggest so. See also Eric Hosmer, says this K.C. Royals fan.

I'm too young to have seen either of them before they were well into their thirties. "WAR is wrong about Winfield" is a statement I have no opinion about. "WAR is otherwise generating a list that makes sense", though, is hard for me to not notice.
6:27 PM Jun 26th
How about for evidence, the fact that people who watched him play didn't think he was the worst defensive player of the era?

It is a huge leap of acceptance to say that it was all biased with false perception.

Why is it not plausible to check the system against information that is very contrary to what it is saying?

Didn't Bill James say something about surprising outcomes produced by models likely being incorrect?

You know there could be data missing from the calculations that wasn't available at the time that could change the outcome immensely.
6:21 PM Jun 26th
"And, we shouldn't question this because it is beyond reproach."

Just to be clear: you can question it.

This issue is rejecting it because you don't like the results. Rejecting something means you've come to a conclusion. And in order to reach a conclusion, you should have evidence to support it.

And so far, the only evidence provided is number of gold gloves. That level of evidence is sufficient to question the results, but not to reject it.
5:29 PM Jun 26th
All Defensive Players 1960-2000 (randomly selected time period)

Name Def
Dave Winfield -243.9
Don Baylor -233.5
Harold Baines -224.9
Rusty Staub -216
Frank Howard -206.9
Willie McCovey -204.8
Danny Tartabull -202.8
Greg Luzinski -201.3
Jeff Burroughs -192.1
Hal McRae -188.8

Who knew? Winfield was the worst defensive player over the 40 year period.

And, we shouldn't question this because it is beyond reproach.
4:34 PM Jun 26th
Guy...buddy: I've done the work. Believe me or not, but I've gone down the rabbit hole to figure out how WAR works. I'm not jumping in blind. I'm not some Luddite arguing about the damned machines.

I'm asking you to do the other work. Leave the WAR aside. Start from scratch. Start in 1974. Who you gonna take, the twenty-two year old who hit ten homers in Fenway, or the 22-year old who hit 20 in San Diego?

Go to 1978. Who do you want? The twenty-six year old who is hitting .308 with a 20-20 year in San Diego, or the twenty-six year old who is hitting .247 in Boston? Look at their track records...offense and defense. All of it. Whatever you want. Who ya taking?

Go to 1981. Who are you trusting? The one who has a breakout year in Boston, or the one who had the same good year he always has?

1983. Boston guy hits a slump...his average goes back down to .238. The other guy, now a Yankee, hits .283-32-116. We know those aren't the only numbers, but still. The Stadium is tough on right-handers. That's pretty impressive.

The Boston guy comes back in 1984, big year. The Yankee guy hits .340. Who are you going to take going forward?

WAR says we should've been taking the other guy. Reputation be damned. Common sense be damned. Take the other guy.

The case rests on two metrics: an assumption of outfield defense against a league average that makes imaginary fly balls into real fly balls, and a positional adjustment that makes a run different if that run comes from an outfielder or a shortstop.

I'm not arguing that those aren't BAD ways to think through a problem. They're fine ways to think through a problem. But....shouldn't we be skeptical?

The people who watched Winfield bounce around the outfield thought he was capable enough. They gave him awards for his defense. They let him play center and left and right. They praised his arm, which was very good. Maybe he played deep. Maybe he wasn't great on tracking the balls.

And maybe the math doesn't have it right. Winfield is a weird case...he bounced around, and he sometimes had speedsters in center or left or right. Maybe the math is punishing him a little too much.

That's possible, too, right? You can't be SO wedded to TZ and RF and the positional adjustments to think they're absolutely perfect.
2:48 PM Jun 26th
So Dave, I take it that is your way of confessing you know very little about
WAR and how it works, but nonetheless felt qualified to write a critique. But instead of fessing up, you frame it as a populist criticism of alleged elitism on the part of those people who are willing to make the effort to understand it. Sorry, but this man of the people shtick doesn't impress.

The WAR system is complex, no doubt, but it is also 100% transparent. Read the glossary items at either website. Or take a deeper dive and read some articles. Everyone can decide for themselves how much they care about the details.

In the end, your position is that A) "it's your job to convince me to set aside my totally subjective and data-free opinions," and also B) "I refuse to learn the methods underlying your argument." It's a Catch-22, designed to protect your faith-based views. And that's entirely your right, of course. But then you shouldn't expect anyone to take your "argument" seriously.

The consensus opinion of every person who watched baseball for twenty years is wrong.

You have to love someone making this sardonic statement to dismiss a data-based conclusion -- unironically! -- on Bill James' website. Truly marvelous. Dude, do you even know where you are?
1:02 PM Jun 26th
Marc Schneider
The statistical analysis here is beyond me but my question is more based on observation than analysis. And that is, are defense and offense equally important? I take it that WAR does treat them as equally important. Does this make sense? I understand that a run saved is just as good as a run driven in, in most cases. But, does that even out in terms of putouts? I mean, a guy misses some fly balls but that doesn't necessarily correlate to additional runs. This isn't really so much about Evans and Winfield, but just generally about how to evaluate the importance of a player's defense vis a vis his offense. I assume it depends on context-position, who else is on the team, etc.
12:56 PM Jun 26th
Also, the blindfolded Adam Dunn line was a joke. Please: no one feel obligated to post a table of his defensive metrics.
12:47 PM Jun 26th
The defenders of WAR always...always...hide behind the line 'You should understand what you're talking about before you talk about it."

That's nonsense. The people who have created WAR have an obligation to have it make sense. If you're going to say 'We have a great system that takes everything a player does and gives it a numerical value,' you should be able to show your work clearly.

What WAR says is this:

-Dave Winfield is worth X. His twenty-year career is this two-digit number.

-Well, we know that's true because his Offense is this number, and his defense is this number.

-How'd we get that? Well, the offense is based on these calibrations. The defense is based on these. With the defense we've added some other stuff about trying to adjust for the position where he played.

-More detail? Well...these calibrations are made by taking x, y, z, and multiplying it by r, p, and q, and then measuring them against an invisible player Pi, who looks good in batting practice but can't hit the curve. Nice guy, though.

-That's not clear? Well, you have to find Pi by measuring the circumference of a baseball and divide it by the radius of a bat handle, but you have to use a different baseball for each year, and the bat handle data can be found here. don't like our conclusion? That's because you didn't understand it. Don't talk about it until you understand it. Just accept it as right. The consensus opinion of every person who watched baseball for twenty years is wrong.


WAR starts with a conclusion, and demands US to work back to the start.

Say what you will of traditional stats, at least they asked us to do the work going FORWARD. The traditional method has been: "Player A was a shortstop who hit .290 with 120 HR and won six Gold Gloves and made eight All-Star teams. Player B was a left fielder who hit .275 with 90 walks a year and 400 home runs and pretty crummy defense, but he won an MVP. Who is better? You tell us."

WAR starts with the answer. "Player X is worth 72. What does that mean? Figure it out yourself. Better yet: shut up and take our word for it."

That isn't a knock on the people who worked to create it, or even the metric itself. But it's a big problem. WAR has been granted significant weight to how we evaluate players, and most of us don't understand it. Most of us don't have the capacity to look under the hood and understand what the heck is going on. A lot of us just shrug and say 'OK...I trust it."

I don't trust it. I don't think it's accurate, but that doesn't really concern me. I'm fine if it's inaccurate: I don't need perfection. What worries me about WAR is that it's's top-down, here's-the-answer-but-don't-look-under-the-hood design leads towards a mode of thinking that is reductive and narrow, a mode of discourse that is rigid.
12:45 PM Jun 26th
FanGraphs goes further: Dwight Evans is counted as a pretty darned horrible, and Dave Winfield was what you’d get if you made Adam Dunn play outfield in a blindfold

Again, it would really help the discussion if people familiarized themselves with these metrics and what they actually say. None of this is true.

WAR says that Dewey was an above-average RF (excellent in his youth), but a bit below average defensively among *all* players. All that means is that RFs are worth much less than most other positions on the diamond. Do you disagree? If so, why? And if WAR is punishing corner OF unfairly, why did 6 of the top 10 modern players in WAR play there?

Nor does WAR equate Winfield and Dunn (never mind a blindfold). It says that Winfield was 7 runs per season better than Dunn -- I don't know if that's exactly right, but seems plausible. It's true that a below-average corner outfielder does not have much defensive value in the WAR system. But isn't that true? Or if these skills are actually scarce, why do players have to hit so well to keep a job there?

And can we stop with the tribal nonsense in which everyone is assumed to be in a "pro-WAR" or "anti-WAR" camp? Of course WAR has flaws -- no one is saying otherwise. One can point out that Dave's specific criticisms here happen to be almost entirely without merit without making a blanket endorsement of WAR.
12:08 PM Jun 26th
I think some folks should read what the article originally said and the point that was made. I now see people trying to defend the logic behind the system at a micro level to say it makes sense. Go back and read the original point which was that offense is trustworthy as a measurement, and the advantage gained hitting by the perceived superior player (Winfield) is wiped out by a very flawed statistic - DEF. And, that the player, perceived to be a good fielder, throughout his early scouting and his career (earning accolades) was actually one of the lowest scoring players in Def, at a level with Adam Dunn. Now people will say that perceptions are biased, and they are, but stats have anomalies too, and this is one. Maybe he was overrated, or just maybe he wasn't Adam Dunn or perhaps the system needs to be improved-alot.

"The gap inWAR lies in the way the metric measure of both players’ defensive abilities. Baseball-Reference counts Dwight Evans as a slightly negative defensive player, and Dave Winfield as a more-than-slightly terrible defensive player. FanGraphs goes further: Dwight Evans is counted as a pretty darned horrible, and Dave Winfield was what you’d get if you made Adam Dunn play outfield in a blindfold."

Here’s how FanGraphs credits Winfield and Evans:


But ‘Def’ is terrible…it is a flawed, damaging statistic.

11:22 AM Jun 26th
Michael hits on another challenge of adjusting to a league average/replacement level. Different leagues have varying levels: if Winfield was a bad defensive outfielder compared to his peers, it's possible that some of that gap was less about his net abilities as a defensive player, and more about where the line of 'average' happened to be when he was playing.

I think of Gil Hodges: he was not a great first-baseman by any historic standards, but he was the best first baseman of his era...he was the best first baseman at a moment when all the really great players were outfielders or middle infielders or catchers.

Do we measure Hodges contribution against his peers, where he shows as the person at the top of a historically low mountain, or do we try to understand his contributions in a wider historical context?
11:11 AM Jun 26th
Average isn't just an abstract construction, it's a measure of the midpoint of all performance at the position by everybody playing there. Ignoring any math for the moment, just looking at names...

Winfield as an NL RF was in a group with Parker and Griffey, but also Staub, Murcer, Jack Clark, Jeff Burroughs... hard to believe Winfield was below average among that group.

As an AL LF, he had Henderson and Wilson, but also Rice, Oglivie, Kemp. Again, he should be OK here.

As an AL RF, he's with Barfield, Brunansky and Evans, so that's tough. Kirk Gibson, Cory Snyder and Rob Deer were in there too, not sure how they rated at the time, but there's also Tartabull and Baines(!). Still, this is tough company for a 35-ish year old big guy.

The Padres (managed by Roger Craig) did something weird with Winfield in 1978, where he played a bunch of games in his usual RF but also a bunch in CF, with the total GP way over 162. Not sure if this messes up his numbers significantly.
10:37 AM Jun 26th
I had an email exchange with Bill at the time. He wrote about Parker here:
9:25 AM Jun 26th
[i]Both lines of thought...both narratives...are absolutely consistent within themselves, but they are very much at odds with each another. I'm interested in that disconnect, and what it reveals about the limits of both stories.[/i]

I confess that really wasn't my takeaway from your piece. But maybe I misread you?

[i] these numbers are broken. They’re wrong. They do a disservice to both players, forcing an unreasonable comparison, and they do a disservice to the field of sabermetrics. [i/]

The gap between these two stories is not only about defensive perceptions failing to align with statistical data. Dewey was basically the player you would create in a lab if you wanted a player to be undervalued (while Winfield was to some extent the reverse).

Compare them on BA vs OBP:
DE .272 / .370
DW .283 / .353

Compare them on RBIs vs. RC (per 650 PA)
DE: 85 RBI / 99 RC
DW: 96 RBI / 95 RC
(Winfield batted 3rd or 4th in the lineup 73% of the time, Dewey just 14%.)

They were basically equal hitters, but Winfield did the things that both fans and "experts" valued most at that time, while Dewey was best at the things we came to know (thanks to Bill) matter a lot.

It's possible the truth lies somewhere in the middle of these two stories. It sounds nice to say so. But I seriously doubt that these two narratives have equal explanatory power.
9:06 AM Jun 26th
I think that Guy123 gets close to what interests me about this specific pair.

I can follow all of the logic Guy123 and Tom are taking us through: I've done it myself. We start with Evans getting to one or two balls a month that Winfield wouldn't get to...we go from there and figure that's 10 or 12 runs a year...we go from there to saying we have to make a positional adjustment for outfielders...and we incorporate that into offensive statistics that have been adjusted for park and league contexts. I get all of that. I can follow that line of thinking...that narrative...and believe it accurate.

But there is another line-of-thinking I can also follow. That line starts with players as minor leaguers or college stars...and it goes through their first years trying to deal with being in the majors, to their best years, to their declines. It is a narrative that accumulates in daily box scores and articles in newspapers and photos on magazines and highlights on the television. There is some math to it: it is tabulated in the counts for All-Star games and MVP votes and HOF votes. It is a broad line of thinking, one that takes into account multiple perspectives.

That second line of thinking is what I was trying to give expression to, so that we could set against the line of thinking that Guy123 and Tom have both explained.

Both lines of thought...both narratives...are absolutely consistent within themselves, but they are very much at odds with each another. I'm interested in that disconnect, and what it reveals about the limits of both stories.

8:23 AM Jun 26th
I couldn't find the article that Tom Tango refers to. There are articles where Bill James compares Dwight Evans to Dave Parker, but I didn't see Tango's name in the comments.

One of those links refers to a Grantland article. Here it is but there aren't comments on it:
8:04 AM Jun 26th
Thank you to Tanner, MWeddell for reading my comments in their spirit: not defensive.

Tanner: yes, I presented the facts so it could START a discussion, not end it.
- I started with putouts for Winfield for a single season and his career.
- Then rtayatay provided more facts on Angels outfielders.
- Charles follows it up with a different angle in addition to showing that Win Shares also has Winfield below average.
- CCornell notes that an independent system has Winfield as below average fielding.
- Guy provided more facts based on Dewey, and the context of what happens when you just add it up over a career.

Of those of us providing facts and other data-driven viewpoints, there's a consensus that Dewey is above average and Winfield is below average. Even those based on anecdotes are willing to concede that.

The question is the MAGNITUDE of that. How much better than the average RF can Dewey be? And how much worse can Winfield be? Can each be +/-5 runs per season x 20 seasons? Well, that's a 200 run gap.

And I had that same discussion with Bill James with regards to Dewey and Dave Parker several years back. If you can find that article in the archives.

7:24 AM Jun 26th
Getting to “one less ball per month” does not 100% equate to 1 run per month. And over the course of his career DW was creating far more than 1 run per month more than DE. Sorry, not buying it. I don’t understand how intelligent people can refuse to acknowledge an obvious weakness of WAR yet turn around and attack Bill James constantly for systems he’s trying to develop to improve our understanding of defense.
5:09 AM Jun 26th
I’ve been staying away from the Readers’ Posts because of the prevalence of hatred there, but the air seems cleaner on these comment boards so I thought I’d raise a point.

Winfield, if I read it correctly, had a higher-than-average number of assists but caught fewer flyballs than could be expected. Could these things be related?

In other words, was he accidentally generating more assist opportunities? And if so, does it impact how we evaluate his overall defense?

Scenario: Runner at second, two outs. A short fly is hit in front of Winfield but he can’t reach it in time. However, he gets it on one hop and nails the runner at the plate. Result: Winfield gets an assist but no putout. Still, the outcome is the same as if he had caught the flyball off the bat.

3:50 AM Jun 26th
I agree, Tango isn't being defensive. He's just stating his point.

However, he has only picked a few certain facts to prove his point. As with anything, anyone can pick facts to prove any point:

1 - batters who hit .300 reach base at a higher percentage than those who hit less than .300

2. the more games a team wins, the more 20 game winners they will have

3. the more errors a shortstop makes, the more runners he will allow on base

All facts, and all correct. However, if you let that one fact in each area become the final measure of anything, then you have not truly looked at the issue the best way possible, and need to add in more facts to find the true picture.

If I read it correctly, and Tango can please correct me if I'm wrong, that was his entire point, in agreement with what Bill is doing: that the facts we are using are not the only method to use and that there is more to be done. Apologies if I'm putting words into his mouth.

However, one other point: you can liken this to doing an evaluation of an employee, or a business, or whatever.

No one does this based on the facts (numbers) alone. Sure, they might be making quota, or exceeding it, or developed a new product, or whatever.

But each employee/company is also evaluated on other criteria:

Is the worker an office gossip, who causes problems at the office?
Does the company allow it's employees time to volunteer in the community?
Does the employee steal from the company?
Or miss too many days of work? Or is rude and arrogant to co-workers?

Most of that is not quantifiable. It is perception, opinion, and experience.

And to me, any evaluation that is done purely by numbers and not with perception, opinion, and experience is a very sterile evaluation, and you will never get the true picture of anything.

Just my opinion. I could be wrong.
2:30 AM Jun 26th
This probably seems churlish, but I have to ask: has anyone participating in this discussion really examined what WAR (FG) says about the defensive skill of these two outfielders? For Winfield, it says he was -93 runs over his career. Maybe that sounds like a big number, but Winfield had a long career. Over a full season in the OF (150 games), that means Winfield was -5.8 runs compared to an average player. Let me translate that for you: WAR's outrageous assertion is that Dave Winfield missed a ball that an average MLB player would have reached once a month. It doesn't claim Winfield was "terrible," just a bit below average.

WAR says Dwight Evans was +65 runs in the field, or +4.8 per season. So WAR's audacious claim is that Dewey *caught* one ball per month than an average OF would have missed. It's tempting to believe that our senses are so finely honed that we can detect such differences in skill simply by watching players, but I promise you that we cannot. (We certainly know Gold Glove voters cannot, since they gave a few to Derek Jeter.)

The rest of "defensive WAR" is just the position adjustment for playing RF (and for Winfield, some DH), which reflect the reality that RFs are less valuable than catchers, and SSs, and CFs, and...well, nearly everyone else. Feel free to argue that WAR treats RFs unfairly compared to other positions, but you will have a hard case to make (and of course, it doesn't much affect the comparison of Evans and Winfield).

Now, I don't know if Evans was really 10-11 runs better than Winfield per season. But it's certainly plausible. As Tango illustrated, you'd arrive at a similar estimate if you just compared their putouts per game (Evans made an extra +.076 plays per 9 innings, or 9.2 runs/season). And the notion that we should reject WAR (or any metric) just because it provides a rating that we don't find intuitive is not my idea of science. And when that rating is entirely reasonable, well, it's just silly.

10:27 PM Jun 25th
Enjoyed the article. FWIW Michael Humphrey's Defensive system on the Baseball Gauge also sees Winfield as pretty terrible defensively.
9:40 PM Jun 25th
Just one question: Oklahoma SCHOONERS??
9:32 PM Jun 25th
You nailed it. The notion that Winfield loses 50% of his offensive value due to his "terrible" defense, is mind-boggling. I can buy that Winfield was overrated defensively. I'll buy he had poor range. But so much to negate 50% of his value as a hitter? That is utter bullshit. And everyone knows it. And this coming from a Red Sox fan who despised Winfield and wished many horrors upon him. He was far better than Dewey, whom I love forever.
8:51 PM Jun 25th
Fine article and excellent comments. I don't think Tom Tango came across as defensive. He's trying to say, based on the statistical evidence (of course WAR can't take reputation into account), viewing Winfield as a below average outfielder for his career might be reasonable based on the traditional fielding statistics.
8:03 PM Jun 25th
Winfield was my favorite player in the 80s so I really love this article. BTW does anyone remember that Bill James predicted DW would be the first player to drive in 100 runs at age 40? Several years prior. Anyway, add me to the list of doubters on WAR’s assessment of him compared to Evans. The enormous difference in road stats cannot be denied or refuted, no way DW’s D was so bad as to negate that.​
7:02 PM Jun 25th
I grew up watching Winfield's Yankees. Well, first it was Munson's/Nettle's Yankees but by the time I was in middle school Winfield was in charge, later Mattingly of course. I saw A LOT of Winfield. As Dave points out, he was regarded as a huge star his entire career and had gaudy contracts and for a long time, no titles. It was inevitable that people would start to carp. I don't know if people ever *really* thought that Winfield was some defensive standout, what I think is that they thought he was an incredible athlete and (maybe) that his size was kind of a liability sometimes, he could be awkward with those big arms and legs churning everywhere.

I'm willing to believe he had some problems out there but I don't believe it is plausible that he was greatly below average; maybe slightly below average. Those assists mean something, after all, you have to run and get to the ball and cut off gappers and get them back into the infield. OK, he made some errors. The Yankees finished in 2nd place almost every year he was on the team. I don't buy that he was some horrible hemorrhaging problem out there.
6:21 PM Jun 25th
Agreed. Very good article. For me it's the same thing with Traynor--supposedly he is a minus 32 overall defensively. I donno-looks like he got to everything. made the plays and every one Saw this and knew it. But hey ignore those countin stats and them rubes from back then. They didn't know Anything. Oh and by the way Gene Tenace Was a good catcher. Sure.

Not buying it. Sorry. Same here. Loved Dewey but he was Not better than Winfield. And if Winfield was your idea of a bad OF I don't want you within a light year of a front office.
5:18 PM Jun 25th
I've considered writing a comparison of these two for years. I have been a Padre fan for nearly five decades, and Dave Winfield was my childhood hero. I was always reluctant to perform any stringent analysis for fear that Dewey would come out ahead. I am now satisfied that Winfield had the greater career. Despite having been a fervent sabermetrics devotee since the late seventies, there's no way I buy the defensive metrics that peg Dewey as mediocre and Big Dave as horrible. I know the eyes can lie, but my eyes could not have possibly been THAT mistaken.

4:27 PM Jun 25th
I have Winfield 11th and Dwight Evans 14th in my listing of Right Fielders that has both WARs and win shares in the formula as well as some other stuff. They are both well above 750 points which I consider as an automatic hall of famer.

I have 181 players with at least 750 points which is below 1% of all the players who have played the game.
2:42 PM Jun 25th
I think how Dave breaks it down for the segments of their career might make his point better if he does it for the WAR numbers and defense too.

For instance, from Dave's quoted raw numbers, there is no question that Winfield was a much better offensive player through 1980, so how do the WAR numbers compare through that date? Total bWAR for Winfield is 32.0. I am unsure what the fWAR numbers are. Just looking at the raw numbers, I don't see Winfield as a terrible LF, like the 1990 numbers paint him to be. It looks to me that he was, at worst, an average LF, (his numbers are bad when he was asked to play CF) and usually a little bit better than average LF. His OWAR and DWAR numbers for that time period are 30.3 and -3.1.

Evans total bWAR for the same period is 26.5. Broken down, he is 16.6 and 5.9 for the same time period.

So Winfield only builds up a 5.5 bWAR lead during the first 9 seasons, despite the fact that he is nearly twice the offensive player. That's the period of time we need to examine the defensive stats to see if someone is amiss, not when Winfield was a 38 year old statue in RF for the Angels.
2:38 PM Jun 25th
Dave nice job and argument, one I have made many times myself about different comps. I grow weary of the argument actually and the arrows that come back towards you. It was ballsy of you to take it on. Good show!

Tangotiger. You have been a great, great contributor to the Sabr community, you are a genius in many ways. Keep up the good work. But, I know, there's always a but....I wish you would be more willing to hear different takes about WAR however. WAR has been a revolutionary development, but can we please just back away from the defensiveness. You don't appear to like any challenge. Let us try to improve upon it, and accept with gratitude the kudos you deserve. I've observed this MO for years, and it just seems really petty. You're better than that. I know, when something is your baby, criticism can come hard. Please listen. I hope you don't take this wrong, but that may be hoping for too much.

Having said that I think everyone should look at the example of Winfield/Evans defense comparison with skepticism that it is a statistical anomaly, something isn't being measured or is incorrectly. I mean isn't it obvious? Aren't there many, many cases that can be picked from the database like this. WAR is a very good model. Bottom Line:Models don't work perfectly. And, listening to other opinions, not just the worshipers who agree with everything you said is helpful. Personally, I've been excited about the friendship that has developed between Bill and Tango. I hope the back and forth creates some great work.
1:23 PM Jun 25th
(I hadn't seen Charles' post which also mentions about Win Shares for fielding.)
12:37 PM Jun 25th
Of course I love the willingness to doubt and even overrule what the defensive metric says, and I love the article.

However (BTW this isn't a "however" that will lead to any different take on the article!) .....I always want to check on such a thing to see what Win Shares says. (John talked about Win Shares but didn't show a breakdown for fielding.)

For what it's worth -- and in this case I'd be inclined still to overrule it but it is in line with the "WAR" showings on both of these players.

I looked year-by-year, every year from 1974 to 1988, not so much at the raw Win Share fielding numbers but at how each guy ranked at his position among all players.
(I took the numbers from Baseball Gauge; would have taken them from this site but I don't find fielding breakdowns for that period.)

For something like this, IMO it's the best way to get a fix on what the metric says about the guy as a fielder that year.
I do agree that the raw number says a lot, especially when it's a real real high number. But for something like what we're looking at here, I think the ranking at the position is the most telling.

For what it's worth -- and, as I said, I'm still inclined to largely put aside the defensive metrics (not completely):

Evans shows extremely well in the yearly rankings from 1974 to 1981, then middlingly through 1986 and worse thereafter.

Winfield shows poorly in most years, middlingly in some, good in just a couple of years.

(details available if anyone wants any)
12:36 PM Jun 25th
As an aside, Win Shares also hates Winfield's fielding, and Bill pointed this out in the Win Shares book.

Sticking to putouts in 1990, he had 167 in 919.2 innings, by League RF had 4835 PO in 20182.1 innings. Two areas of adjustment:

The league had 12869 K and 24025 A in those 20182.1 innings; the Angels had 944 K and 1871 A in 1454 innings. Thus, the Angels' PO-A-K was 6.5% of the league totals. Winfield was on the field for 63.2% of the team's PO, so we'd expect him to have 200 PO.

The league had 76800 AB, 19900 H, 12869 K, with 30244 AB, 7883 H, and 4429 K by lefty batters. The Angels' enemy batters had 5548 AB, 1482 H, and 944 K; lefties had 1754 AB, 482 H, and 257 K, so an excess of 256 outs by lefty batters, or 162 prorated to Winfield's innings. Right fielders make 1 fewer PO for every 15 extra outs by left-handed batters, lowering his expected PO to 189.

So, he had 167 when he was expected to have 189, a shortfall of 22 PO. Not taking into account how many hits the Angels gave up (so leaving the team as a whole at 0 runs saved versus average), it's about a run saved for every 5.2 extra PO (this is a number I derived from looking at three PBP methods and comparing them to what traditional stats show), so I get 4.3 runs worse than average, strictly on PO.
11:56 AM Jun 25th
I realize that there is more to the advanced defensive metrics than I'm mentioning, and I don't want to discount any of the work good people have done.

What I'm trying to say is:

1. We know that Dave Winfield was a better offensive player than Dwight Evans. He had a longer career, bigger years, and better career numbers, all while playing in tougher parks. We can all agree on this.

2. Granting that Evans was a better defensive outfielder, are we willing to believe that Dewey's edge as a defense player was SO significant that it more than closes the gap between him and Winfield as offensive players?

If Winfield was viewed by observers as a terrible outfielder, I'd believe it. But he wasn't: he was a decently mobile player who won a bunch of Gold Gloves and threw out a lot of base runners. Did he play too deep? Sure. Was his read of balls lacking? Probably. Did he get his Gold Gloves because he used his height to snag a couple dingers at the fence? Probably. He was likely overrated as a defensive player.

But there's a long way from 'overrated' to 'terrible.' WAR makes that jump without hesitation, and asks us to believe something that strains credulity. I can't make the same leap.
11:50 AM Jun 25th
rtayatay: excellent presentation of facts. I don't have an answer to your question, but what you are doing is exactly how to move the discussion forward.
11:29 AM Jun 25th
Can we just have a discussion on the facts?

I said this for his career:

How about for his career? Using facts only: 2.01 putouts per game in RF compared to the average of 2.15. Using facts, that 0.14 fewer putouts per game. And 15960 innings is 1773 games. Still only facts. And 1773 x .14 = almost 250 fewer outs. Those are the facts.

The argument is that his opportunity space is going to be so different over a 20 year career that we can't even presume a league average outfielder who made 2.15 putouts per game would have made 2.15 putouts per game given Winfield's conditions.

And that can certainly be true. We don't have all that information. What we have are the facts that Winfield did in fact catch 2.01 balls and the league average RF caught 2.15. And that there's a gap of 250 outs based on that.

That's the starting point. Then, you have to argue that Winfield faced a very disproportionate set of opportunities over 20 years.

This is the important part: no one is making those arguments. The only arguments being made is "nah". That's the crux of the rejection.

Total Zone does more than just look at the putouts as I have. I am looking at the putouts because those are facts.

11:24 AM Jun 25th
So, just looking at the 1990 Angels, a lot of questions pop up for me. The 1990 Angels ranked:

2nd in lowest XBH%
2nd in highest GB/FB%
2nd in highest IF/FB%

Combine those last two and I'd guess they had the lowest FB% to the OF in the league.

Meaning, they were an extreme ground ball team, with what looks like a significantly lower number of balls even making it to the OF.

In addition, you have 38 year old Dave Winfield playing next to 27 year-old Devon White, who was an excellent CF. Who do you want taking a ball hit between them?

On top of that, not surprisingly, Devon White had the lowest range factor during his prime in that 1990 season.

Here are his defensive runs numbers:

1987 +27
1988 +11
1989 +22
1990 +1
1991 +18
1992 +33
1993 +18

All of that makes me question how well the defensive runs number takes into account surrounding factors. Was Devon White really just an average CF in 1990, but one of the best in the AL in the other years? And was Dave Winfield really that bad in 1990, or was at least a chunk of that due to other reasons?

11:20 AM Jun 25th
Just to tango with Tango:

You went from facts straight to speculation without noticing it, Tom.

It is certainly true that Winfield, in 1992, had fewer putouts than an average right-fielder. But you jumped right from THAT to saying Winfield missed 51 flyballs, and you call that a fact.

It's not. It's just not. You don't know how many fly balls Dave Winfield missed in 1992: you're making a guess. Maybe it's a good guess, and maybe it's not, but it's not a fact. You can't call it a fact.

The structure of defensive WAR does the same thing you did in your sample: it starts from facts and then glides straight into making educated guesses, all the while claiming 'fact.'

That's OK...that's the business. But when you get a result that seems to challenge basic clear thinking, you can't just say, 'Well, we're still right' and stick your fingers in your ears.
11:01 AM Jun 25th
Despite being a huge Evans fan, it has always been clear to me that Dave Winfield had the better career. WAR is deeply under valuing Winfield's time in the OF, and to base your argument on the age 40 defensive season seems like an odd jumping off point.

I think another glaring example is the OFF/DEF score weirdness happens with second basemen Roberto Alomar and Bobby Grich. Both players have similar OFF scores (272.5 for Alomar, 254.5 for Grich) but when it comes to defense Grich somehow gets a 134.1 and Alomar gets a 15.8.

There is no way in hell that Bob Grich over the course of his career was that much better playing second base than Roberto Alomar. It just looks way outside the lines of common sense.

Bob Grich career WAR 69.1, Alomar 63.6.
10:50 AM Jun 25th
I think the problem with BWAR is that it either heavily penalizes certain defensive players and heavily rewards other players. Winfield is a case in point where they severely penalizing him for his defense. Ken Singleton is another player who is heavily penalized in BWAR for his defense.

I think the problem with Dw. Evans in terms of the HOF is that there are 5 Right Fielders from his generation (1970’s-80’s) that have similar Win Share numbers but are not in the HOF: D. Parker, R. Smith, K. Singleton, and Bo. Bonds.

There’s the big 3 RF from that generation in the HOF: R. Jackson, D. Winfield & T. Gwynn.

Then you have these five:
Career Win Shares:
D. Evans: 339
D. Smith: 327
R. Smith: 326
Bo. Bonds: 302
K. Singleton: 297

Best 7 seasons Peak WS:
K. Singleton: 204
Bo. Bonds: 199
D. Parker: 197
R. Smith: 184
D. Evans: 175

I think you can make a strong HOF case for all 5 of these players.

Strawberry had a 184 peak but only had about 250 win shares.

Parker had a period from 1977-79 where he was considered the best player in baseball. Overall he has the most win shares. Bo Bonds has the most win shares in baseball from 1969-71. I think he’s 5th overall from ‘69-73.

You also have Staub from 60’s and J. Clark from the 70’s-80’s in that mix.
10:08 AM Jun 25th
Prorated to Evans's games, Evans had about 20 more assists and 20 fewer errors than Winfield, which is about 15 runs saved between the two events, off the top of my head. Errors by outfielders are mostly allowing advancements on hits, so aren't that costly (I use 0.375 runs). I assume the value of an assist is taking away a walk (0.33 runs) and preventing an advance (0.2 runs); I'm undercounting a bit, as you can also include the value of not letting another batter come to bat (0.11 runs). Regardless, these two cannot possibly account for the defensive differences between them.

As a note, when I was a kid in the 1980s, Winfield was a local legend, and we heard a lot about him constantly. My mother spent three years at Winfield's alma mater, Central High in Saint Paul, albeit a few years before Winfield was there (she graduated in 1965, Winfield in 1969); I had a friend whose father was there with Winfield. There are team pictures of Winfield (often with his brother Steve) in Little League in a shrine near the gym there. I like to contrast that with Paul Molitor, who grew up in the same area (though he went to Cretin, a prep school), and Jack Morris, who also grew up in the same area (who went to Highland, from which my mother actually graduated; the school didn't open until 1964), who were stars at the same time, but were almost unmentioned, even in my house (Molitor's father was an accountant with Burlington Northern Railroad, where my father worked as a lawyer in the early 1980s). So, to answer a rhetorical remark, I'd say that Dave Winfield was a damn good player in his childhood.
10:00 AM Jun 25th
You don't mention Johnny Callison even though he is one of the assists leaders. A fine arm, a decent hitter for a few years, and one of my favorite players in my early teens.
9:39 AM Jun 25th
Just as a for instance, in 1990, Dave Winfield is shown as -19 runs relative to the average RF, which is a terrible terrible number. Is it possible?

Well, Winfield had 1.70 putouts per 9 innings. The other Angels RF, Dante Bichette, had 2.22 po/9. The league average RF was 2.23. That's our starting point, based on the facts. And the facts are that Winfield caught 0.5 fewer balls per 9 innings than his peers.

Winfield played for 920 innings, which is the equivalent of 102 9-inning games. And 102 games x 0.5 balls per game is 51 balls. That's based on the facts. So far, all I've done is state facts.

Catching 51 fewer balls than your peers is a terrible terrible number. Very very terrible. How many runs should that be? Knowing nothing about anything, that should be about 40 runs. Total Zone at Baseball Reference suggests 19 runs.

How about for his career? Using facts only: 2.01 putouts per game in RF compared to the average of 2.15. Using facts, that 0.14 fewer putouts per game. And 15960 innings is 1773 games. Still only facts. And 1773 x .14 = almost 250 fewer outs. Those are the facts.

How many runs should that be? Knowing nothing about nothing, I'd say 200 runs.
9:18 AM Jun 25th
©2024 Be Jolly, Inc. All Rights Reserved.|Powered by Sports Info Solutions|Terms & Conditions|Privacy Policy