Baseball BeatJanuary 15, 2008
"Listen, Buster" Redux (Part Two)
By Rich Lederer

Moving on to Olney's second article (after reviewing his first piece yesterday), Buster begins by mentioning that he received a lot of email over his Friday piece and then jumps into a discussion on Adjusted OPS (aka OPS+).

Adjusted OPS+ is a useful number. And if this your be-all, end-all statistic, keep in mind that:

Mark McGwire and Frank Thomas rank higher than Willie Mays, Hank Aaron and Joe DiMaggio.

Jim Thome ranks higher than A-Rod and Gary Sheffield.

Lance Berkman ranks higher than Ken Griffey Jr.

Brian Giles ranks higher than George Brett, Al Kaline, Orlando Cepeda, Tony Gwynn, Rod Carew and Roberto Clemente.

Adam Dunn ranks higher than Eddie Murray.

I'm not sure where to start here. But let me say that OPS+ should not be anyone's "be-all, end-all statistic." On this front, Olney and I agree. No single stat, in fact, should be viewed in such a manner.

More than anything, Olney's comparisons demonstrate a lack of understanding and appreciation for walks and OBP. Every player that he questions is well known for his ability to get on base by taking a walk. Maybe it's just me, but I don't find it so outrageous that Frank Thomas (157) has a higher OPS+ than Willie Mays (156), Hank Aaron (155), or Joe DiMaggio (155). That said, I believe Olney's splitting hairs here. These four players all have virtually identical OPS+ totals.

OPS+ measures hitting, not fielding or baserunning. Thomas was a great hitter – one of the best ever. However, the Big Hurt's offensive prowess doesn't mean he was a better player than the Say Hey Kid, Hammerin' Hank, or the Yankee Clipper. Mays and DiMaggio were two of the best defensive center fielders of all time, Aaron was a quality right fielder, and all three ran the bases extremely well. Thomas was a poor fielding first baseman or a DH and was a slow runner for the vast majority of his career. Shake it all up, and I'm quite certain that every reputable baseball historian and analyst would take Mays, Aaron, and DiMaggio for their all-round play over Thomas. But that doesn't detract from Thomas' hitting or from OPS+ as a measurement of offensive production (ex-baserunning).

Jim Thome (.281/.409/.565) over Gary Sheffield (.296/.397/.522)? Sounds plausible to me. Thome has out-homered and out-walked Sheffield in 1726 fewer plate appearances.

Adam Dunn over Eddie Murray? Oh my gosh, who thought of this stat anyway? Comparing a career rate stat for a player through his 27-year-old season to another who played up to the age of 41 tells us more about Olney than it does either Dunn or Murray. But, if the truth be known, Murray had a higher OPS+ (143) than Dunn (130) at a comparable point in their careers.

And if you think that Adjusted OPS+ is a set of numbers that generally creates a level statistical playing field for all of the eras of baseball, then you'd have to ignore the following. Of the top 63 players all time in OPS+, there are: Nineteen players who performed the bulk of their careers in the years leading up to 1920.

Eight players who performed the bulk of their careers in the years from 1920-1939.

Seventeen players who have performed the bulk of their careers from 1990-2007.

And a total of 17 players from the 50-year period of 1940-89.

That all may be true. However, I'm not sure why Olney chose to measure the "top 63", but nine of the top 24 (if you include Hank Greenberg) are from 1940-1989. That works out to 37.5% of the sample size, which just happens to match the sample period (50 years divided by 132 years). Interestingly, only four of the top 24 career leaders (16.7%) are from 1990-2007 (13.6% of the overall time frame).

Getting back to Olney's point, I don't believe OPS+ is necessarily weakened by the fact that the distribution of superior results may not be equal throughout baseball history. I'm not aware of any statistic that "creates a level statistical playing field for all of the eras of baseball." Take league leaders. There is only one per category each year, yet the number of teams and players has increased over time (as Olney acknowledges in the next section), making it increasingly more difficult to lead the league – or rank among the league leaders – in any stat.

Part of the reason, of course, is there are more teams now. But part of the reason is that in years in which there is less offense, generally, it is more difficult to create a plus/minus disparity in this statistic.

I tend to agree with the latter assertion.

If you don't think that Adjusted OPS+ is a statistic that skews toward the elite players of the Steroid Era, well, then that's your story and you're sticking to it.

OPS+ might be skewed by the higher run-scoring environment and even steroids, but that does not take away from the fact that Jim Rice hit .320/.374/.546 in a highly favorable park environment at home and .277/.330/.459 on the road. His raw counting stats and unadjusted rate stats should not be compared to those from the so-called "Steroid Era," just as they shouldn't be compared to those from the Dead Ball or Live Ball eras.

But Rice's OPS+ should absolutely be compared to players of his own era. I don't think one can quibble with that unless, of course, they don't like what they see. Using Olney's hand-picked years (1975-1986), Rice ranked 11th in OPS+ among players with 4000 or more plate appearances. Jack Clark, Ken Singleton, and Fred Lynn posted higher totals. Greg Luzinski had the same OPS+ as Rice. Keith Hernandez, George Foster, Dave Parker, and Dale Murphy rank directly behind Rice.

I would submit that Foster is a pretty good comp. Same era, same position, same type of hitter/slugger. Both players won an MVP award. Rice led the AL in HR 3x and RBI 2x. Foster led the NL in HR 2x and RBI 3x. Maybe Rice was better. But, if so, the difference between the two was minor (and mostly a function of playing time), yet Foster never received more than 29 votes for the HOF (or 6.9% of the total).

If you place a lot of weight on playing time and counting stats, then perhaps Parker should be viewed more favorably than Rice. Again, both players were from the same era. The Cobra was a Gold Glove-caliber right fielder and a much better baserunner during his peak. Like Rice, he won an MVP. He also led the NL in AVG and SLG 2x, and TB 3x. As it relates to HOF consideration, Parker's vote total peaked at 116 (24.5%) in 1998 and his candidacy has been trending downward ever since.

Joe Morgan, Gary Carter, and Johnny Bench are the only three players with a lower OPS+ during Rice's best years who have been inducted into Cooperstown. None of them played left field indifferently. In fact, all three were outstanding defensively at much more difficult and important positions. And, to be fair to Morgan and Bench, the chosen period didn't capture all of their best seasons.

If we're going to get behind a player from this era, let's focus our time and attention on Bobby Grich. Note that the slick-fielding second baseman ranks 20th in OPS+ during that same period. There are only three non-OF/1B above him on that list. The three? Mike Schmidt, George Brett, and Morgan. Three of the greatest players in the history of the game. But I digress. I'll give Grich, who garnered almost zero support from the voters, his due in a separate article at a later time.

It's not a perfect statistic. There aren't any perfect statistics.

Agree. No single metric is flawless, whether they're counting or rate, unadjusted or adjusted. They all have flaws. None of them are perfect, especially as standalone measures. It's important to put all stats into their proper context.

While I'd generally agree that to focus on building a Hall of Famer's credentials around a single year of MVP voting might be dubious, the numbers cited in Friday's column accounts for hundreds of votes from every AL city over more than a decade. A lot of writers who watched Rice play daily, at the time he was on the field -- rather than through the time-machine prism of Adjusted OPS+ -- thought he was pretty damn good.

I don't doubt or dispute that. But I think it is only fair to point out that writers have always been biased toward RBI (check out the MVP award winners over time for proof) and rarely, if ever, adjusted for park effects – particularly in Rice's day. As a result, I believe it follows that Rice was overrated by writers back then and is benefiting today from the misperception that he was better than he actually was. Importantly, by making this statement, I'm not revising history. The fact that writers "thought he was pretty damn good" doesn't necessarily mean he was pretty damn good.

If you want to quibble with the fact that he won the award in 1978, or with his placement in some particular year, OK, I get that. But to ignore the MVP voting entirely, as if it isn't at least some kind of barometer of his play over the course of his career, is embarrassing. This is like saying, "Hey, forget the Oscar voting of the 1950s. Marlon Brando was clearly overrated."

At the risk of speaking on behalf of serious fans and students of the game, I believe we would all like for the MVP voting to be a "barometer" of Rice's (and everyone's) play. We're not "ignoring the MVP voting entirely." Instead, we're just discounting it. Doing so shouldn't be "embarrassing" to anybody. After all, Roger Maris and Dale Murphy won back-to-back MVPs and never even sniffed the Hall of Fame.

Look, I've never met Jim Rice, didn't grow up a Red Sox fan, don't think he is one of the very elite players of all time. I understand why someone wouldn't vote for him (but don't agree). But to portray his career as entirely unworthy of Hall of Fame consideration is silly.

I've never met Jim Rice either, didn't grow up a Red Sox fan or hater, and certainly don't think he is one of the very elite players of all time. I understand why someone would vote for him (HR and RBI titles plus an MVP). That said, I disagree. When put in the proper context, Rice was not as good as he appears to be on the surface.


Great article Rich!
I must say I've always been worried with some peculiar numbers relating OPS+.

For example, Pete Rose's OPS+ is 118 which is one of the worst for HOF of Rose's era (asuming Rose has the numbers to enter the HOF). McCovey's 148, Stargell's 147, Allen's 156 (and he's not in the HOF), R Jackson's 139, just a little example. Was Rose that inferior to them? Was Rose extensevely overrated?
And Rose never had a very high OPS+. He had 2 years (when he was 27 and 28) with OPS+ of 152 and 158, but the rest of his career he was below 141.

And take the case of Ripken. His OPS+ is 112. Yount's 115, Larkin's 116 and Trammell's 110 (and the Larkin & Trammell probably won't go to the HOF). I know Ripken was a great player, but how great could he really be if he's just in the same level than Yount, Larkin and Trammell?

Thanks, John. A few things:

* It's extremely important to put offensive production in the context of defensive positions. As I detailed in the case of Frank Thomas vs. Willie Mays, Hank Aaron, and Joe DiMaggio, a higher OPS+ doesn't equate to being a better player.

* Secondly, it's important to recognize length of career when considering rate stats. Rose, for example, had a much longer career than the players you cited. His counting stats were enhanced and his rate stats were harmed by this fact. That's why peak stats and the slope of one's career are necessary factors when evaluating players.

* Thirdly, offense is only part of the equation. Defense must be taken into consideration as well. Baserunning, although to a lesser extent, should come into play, too.

* Lastly, I would like to think that Yount, Larkin, and Trammell were "great" players, too. Yount's excellence has been rewarded by the voters. Trammell's excellence, unfortunately, has not. And Larkin's excellence has not been voted upon as yet. I'm hopeful that Trammell and Larkin will both be inducted into the HOF and join Ripken and Yount in Cooperstown.

It's my view that the best players throughout baseball history are more or less at the same level, and that the quality of the worst players changes drastically from decade to decade. So I go by a rule of thumb that the 5th best hitter of the 20's is about the same as the 5th best hitter from the 50's, 70's, etc. (No matter what the actual number of WS or OPS+ or whatever you're using).

Anyway, one odd thing that I noticed is that a huge number of players that people compare to Rice as being just as good but not in the HOF, always seem to come from the NL. Not sure why, was the NL way better, way worse, or are we missing a step in figuring the numbers?

I can't understand why people are always trying to compare eras -- the game has changed a lot in the last 25-30 years so imagine how much different it was 50-60 years ago. The same thing with the idea that there is a level playing field in sports -- tell that to Barry Bonds and Jay Payton for example. That sort of cuts to the heart of the steroid issue but this isn't the time for that.

I'm also wondering about the sudden support for Jim Rice from the mainstream writers. Do people think that this is backlash against the "bloggers" over control of the information? You wouldn't think so but the writers are so inconsistent on who they support that they might not really care about the players but are really at war with the new media. I mean, why Jim Rice? Was he any better then Dwight Evans, or even Darrell Evans for that matter. It just doesn't make any sense to me.

John: back when Rice was playing, there was the DH, and most of the best hitter's parks in baseball were in the AL while most of the best pitcher's parks were in the NL. Without touching secondary effects like bullpen fatigue, this meant that the top AL sluggers were driving in runs in contexts that either they didn't have to field at all or that they had a guy who could hit like a major leaguer but not field like one helping his contributions. Although the NL still has no DH, Colorado (pre-humidor) and interleague play have done a fair amount to balance things these days.

Rich: In thinking of Bobby Grich, and Olney's points in general, I came up with a theory of loud versus quiet stats. Loud stats get the ink. Everybody remembers Kirk Gibson's homer to beat Eckersley in 1988. Few remember Mike Davis's walk to bring Gibson up as the potential winning run, or for that matter Davis's stolen base to put the tying run in scoring position. I'm sure in Eckersley's mind walking a badly hurt Gibson to bring up Steve Sax (in a high hit season) with the tying run on second seemed like a bad risk, so he made sure to throw a strike.

The loud stats get the headlines. The triple crown: homers, RBI, and batting average. The quiet stats help teams win. Walks, HBP, advancing runners into scoring position at key moments, high percentage stolen bases even with low theft rates, all are valuable. But they're hard to quantify, especially because walking with a runner on first to put a man in scoring position doesn't do anything unless the next man drives him in. But that RBI single speaks loud and clear.

The thing is, it's hard for the casual observer to grasp the notion. Listen to a ball game (with most announcers). The batter comes up and his batting average usually gets mentions, and often all his triple crown stats. Vin Scully, who I am fortunate to listen to, has long mentioned the quiet stats. After all, the only run that scored in Sandy Koufax's perfect game came on a walk, a sacrifice, a stolen base, and a throwing error, with no RBI. And Scully marvelled about Maury Wills' ability to manufacture a run without a hit back in his heyday. But even as generally enlightened and informed observer as Scully hardly ever mentions walks these days except when Barry Bonds comes up.

For Joe Sixpack, who gets fed a constant diet of triple crown stats, it's hard to even recognize the positive value of Frank Thomas's ability to work a walk, or the negative value that should accrue to players like Jim Rice and Steve Garvey and Juan Pierre. I don't know how we can expand the public consciousness: maybe (just as Rolaid's has the Relief award) we could get Dr. Scholl's to do an ad campaign and award for the player who scores the most runs after reaching base via a base on balls, for players who appreciate the value of walking. I dunno, I guess there are reasons I don't work in advertising.

But barring a sudden nationwide craze in Strat-O-Matic, or changing the Roto rules, I think pusing the quiet stats is going to fall on those of us who want to know why teams really win.

The main problem with OPS+ and using it to compare players is pretty simple. First basemen should outhit guys that play up the middle. Sure, Frank Thomas had a higher OPS+ than a guy like Willie Mays, but how much better way Mays versus other centerfielders of the time than Thomas was versus other first basemen of the time? I don't want to run the numbers, but my gut feeling tells me that at that point its not even close. Frank Thomas may be one of the best hitters ever, but we have to look at certain things in context. Players need to be judged among their peers, not just against the era that they played in. A centerfielder with an OPS equal to that of a first basemen is almost always the better player unless their defense is absolutely atrocious while the first basemen is pulling an Albert Pujols at first base.

If power is one of Rice's two main strengths along with run production, explain how a righty pull hitter with less than 400 (382) career HRs in Fenway belongs in the HOF.

What about another number - 315? That is Rice's career total for hitting into double plays. He had a four-year stretch with 131 total DPs.

While Rice was a fine player, his overall value doesn't equal more deserving Cooperstown candidates such as Trammell, Raines and Blyleven.

For the record, I deleted two comments from someone who told us in another thread just one week ago that we had "lost a reader." I wasn't the least bit bothered by this revelation and am simply not interested in playing games here.

I was a Red Sox fan in 1975, and thought the world of Jim Rice. That does mean that I cannot look back on his career with the knowledge gained in 32 years since.

OPS+ is, if anything, a kind statistic to Jim Rice. OPS weights slugging percentage on a 1-1 basis with on-base percentage. The actual value of each in run creation is about 1-1.7. Rice's OPS was, of course, heavy on the S.

George Foster and Dave Parker are good comparison points for Rice. His teammate Dwight Evans is another. Evans was a far better player, put up far more runs and prevented far more.

The problem isn't Rice, as such, although he's certainly typical of the problem. The problem is that people value baseball players differently.

Growing up in Los Angeles and depending on the Los Angeles Times for my sports coverage, it was a revelation to me in the early 1980s on a trip to Michigan to find out that the Detroit newspapers carried all the little statistics. Walks! HBP! All the players in the league, not just the guys with the most at bats and decisions until the page was full! I only suppose the Times, which IIRC was owned by the same folks who owned The Sporting News, published lousy stats to encourage people to subscribe to TSN.

We are in a golden age of baseball, and it's not just because of all the great sabremetricians and analysts that have broadened our understanding to where we cringe every time Brady Little orders his number two hitter to sacrifice. It's also because the information is retrievable. Barry Bonds has had some of the best years in the history of baseball lately, but while the Olneys of the world will list the 73 homer 2001 season as Bonds' best, the rest of us might choose instead 2002 and 2004, when despite fewer than 50 homers his OPS+ was even better. We know it's from Bonds doing so many good things while not making outs. And yes, it was in my lifetime when Bonds would not have qualified for a batting title in 2002 because his 232 walks didn't count: it was at bats, not plate appearances.

Given that walks were so devalued that they didn't help hitters with good eyes simply become eligible for batting titles, how can we reasonably expect everybody to change over night? We can't. A sixty year old sportswriter grew up in an atmosphere where the MVP was the guy on a pennant winning team who had the best triple crown stats, and maybe in extraordinary years it would go to a player on a weaker team or even a starting pitcher. He may try to keep abreast of modern statistics; I know I do. But the older you get, the harder it is to learn new things. And one of the ingrained elements is that triple crown stats lead to All Star selections and MVPs, and All Star selections and MVP awards determine who gets into the Hall of Fame.

So it is important to keep the discourse civil. We're not going to persuade many people that if Jim Rice walked as often as Dwight Evans then we'd vote for Rice in the HOF. I mean, the rumor flies that Rice would get intentionally walked with the bases loaded, a rumor no doubt started by a Red Sox beat writer who'd gotten loaded himself. The Olneys of the world remember that Rice was more feared than Eddie Murray (which he probably was). Rice got an MVP and Murray didn't, and so never mind that Murray had a much longer career with twice as many walks as Rice, adequate enough defensive play to get a Gold Glove, and the two magic numbers of 500 HRs and 3,000 hits.
Murray got into the HOF, Rice was better at his best, so Rice deserves to be there.

And, lets be fair: there is something to that argument. Murray was not noticeably a better hitter over his career; their *OPS+ are only one point apart despite Murray's huge edge in walks. Murray had a longer career, the last seven years of which were not helping his OPS+ numbers, but did push him over the magic numbers of 500 and 3000. But Murray was good enough to make a roster his last seven years, all of which had double digit homers (three of them over 20) in seasons with generally reduced plate appearances. Rice left baseball when he was no longer going to be put on the field, and he didn't have enough gas left in him, even at a lower level, to stick around to get 500.

My late father, who was a physician, often lamented that if he wanted to stay good at his job, he had to spend a couple of hundred hours a year studying. Reading about new treatments, new drugs, new procedures. He did it, because it was literally life and death for him. But there is no life and death in baseball. Oldtimers *don't* have to stay abreast of new diagnostic techniques for baseball talent. If a HOF voter wants to remember Jim Rice, a marginal HOF candidate, as the most feared slugger in his day, we're probably not going to change his mind.

But I bet that darned few sports writers of tomorrow who are growing up today are going to become HOF voters without knowing and appreciating OPS+ and Range Factor and VORP.

One last thing: while looking up stats for writing this "comment", I noticed just how good Hall of Fame Barry Bonds was. Not Steroid Barry: HOF Barry. From 1990-1998, when Barry was still thin and quick and a perennial 30/30 threat, he posted 9 straight seasons with an OPS+ of 160 or higher. Eddie Murray, a clear Hall of Famer, never bettered 158. In 1999 Bonds dropped from 28 to 15 SB, admittedly in an injury riddled season, and he was never the same again. And I have to wonder if Bonds started using steroids not to keep up with McGwire and Sosa, but simply to rehabilitate an injury that limited him to 102 games.

If I were a HOF voter, Bonds would get my vote. I wouldn't like it, I might not vote for him the first year just because it seems that he's gone his whole life never denied anything, but he won three MVPs and finished in the top five 7 out of 8 seasons. If Bonds had retired in 1999, he'd be the offensive HOF answer to Sandy Koufax; great enough to make up for lack of longevity.

These two articles are persuasive and well presented.

What really annoys me are the writers who defend their votes for marginal candidates and won't vote for Raines and Blyleven (and can't make a reasonable argument for that).

If you vote responsibly for the top two candidates, I can see giving you a pass on your next choice for "most feared" or whoever.

I guess that it's an obvious point, but there's no real comparison between Rice and Dwight Evans; Evans had much the greater career. But I also notice that Evans batted second much of his career. I wonder how many walks he drew that turned into RBI for Jim Rice, in effect enabling Rice to make his Hall of Fame case while hurting Evans' case. Also, the eight Gold Gloves would indicate that someone thought Evans could play right field.

The simple question is this...who would you rather have on your team for the best 12 years of his career, Rice or Raines--neutral park factors of course.
Me, I'm taking Raines

Forgot to tack the "1962" onto my name on previous post, and wanted to raise another point: Jim Rice is certainly worthy of HOF consideration. He was a fine, fine player.

But the HOF is all about drawing a line and putting players on one side or the other. I think length of career is a huge factor, and I look at Rice and Dale Murphy, and they don't quite have it.

well, best i can tell it is not so much a vote for rice as a vote against anyone who played after 1994.

and best i can tell, rice was never even the best player on his own team, but he will be lauded because of the RBIs.

me, i think RBIs are to baseball players as giant boobs are to female humans - they might could be nice for males to look at, but they shouldn't be used as the be all end all to define someone...

Just wanted to draw folks' attention to an interesting post related to Rice by Joe Posnanski:

Admittedly flawed analysis, but interesting nonetheless. As a Yankee fan from back in Rice's day, it makes me think about why it kind of "feels right" to have Rice in the HOF.

In regards to the last article mentioning Sandy Koufax's hm/rd ERA splits: while his road ERA was higher, his road winning pct. was almost exactly the same. Pitchers from his time were resposnible for a lot more decisions, so their W/L record is a pretty valid indicator of quality. So his effectiveness, relative to his environment, was just as good on the road as it was at home.

Excellent articles, Rich, as usual.

I'm surprised that there hasn't been more "meta-discussion" -- particularly in the sabermetric/blogger (ie, non BBWAA community) about the size and inclusiveness of the HOF. This is the "big" vs. "small" argument. As a baseball fan, while I completely agree that Rice would be a marginal Hall of Famer (and probably would not vote for him, at least not before I voted for Raines, Trammell, Blyleven, Dawson, and possibly Parker) -- I really don't get that upset with the notion of Rice actually being part of the Hall of Fame. And I readily concede that putting Rice in might allow people to start seriously talking about Dale Murphy, Don Mattingly, Steve Garvey and others who are a notch further below still -- but I'm not really sure why we've drawn the line where we have.

Ultimately, wherever the line is drawn, there will always be a few players on the bubble, particularly when compared to other Hall of Famers (like the onerous VC selections of previous generations). But more and more, I'm coming around the the idea that lowering the bar a little bit for players like Rice isn't hell freezing over.

And let's put it this way -- I would GLADLY welcome Rice into the HOF if everyone who voted for him also agreed to vote for Blyleven and Raines.

Olney's response today indicates that he still doesn't understand the point. As the number of players has increased over time due to expansion, the tails of the distribution have pushed out. Of course the league leaders in OPS+, Ks, and BBs will likely post higher totals today than they did in in the 1970's. It is harder to lead the league in these categories than ever before.

Why area disproportionate number of today's players ranked in the top 63 in Adjusted OPS+? Simple - they haven't yet entered the decline phase of their careers. Rich implied this when he compared Dunn and Murray at the same ages, but it should be pretty clear that a Berkman or A-Rod will probably see drops in their "career average" stats before they finish out their careers.

"nine of the top 24 (if you include Hank Greenberg) are from 1940-1989. That works out to 37.5% of the sample size, which just happens to match the sample period (50 years divided by 132 years)."

I wonder if the fact that many of the best players went to war during this period skews the numbers. (1) The replacement players might all be interchangable (thus compressing the bell curve and making an ops+ of 100 more common); (2) the good players who did not go to war would look correspondingly better (since the replacement players drag down the actual mean OPS).

It would probably only take one or two very, very good players leaving for a span of years to make a difference (Ted Williams).