Measuring greatness (part 1)by Mike Carminati
April 09, 2009
Nobody is glad in the gladness of another, and our system is one of war, of an injurious superiority … It is our system; and a man comes to measure his greatness by the regrets, envies, and hatreds of his competitors.
—Ralph Waldo Emerson
This past December two players secured their place in the Baseball Hall of Fame via the baseball writers’ ballot, but they took quite different routes to that goal.
Rickey Henderson slid through, as expected, in his first time on the ballot. Even his waywardness late in his career—including two separate excursions with the independent Newark Bears—could not tarnish his stellar career. He was a player about whom Bill James once said “split him in two, you'd have two Hall of Famers.”
In fact many were so shocked that Henderson appeared on only 94.81 percent of the ballots that they even "outed" at least one of the negligent voters. Henderson was left off of 28 out of 539 total ballots.
One must keep in mind that Henderson received the 12th highest voting percentage of all time and was just slightly ahead of Willie Mays—not such bad company after all. Also, Henderson 106 more votes than the 405 required. That’s the fifth most votes over the number needed all-time (behind Ripken, Gwynn, Ryan, and Brett). Here are the top 15 Hall of Famers by voting percentage:
Henderson is behind some players, even contemporaries, who were clearly inferior though Hall of Famers all (Gwynn in particular seems unfair), but it is hard to gripe when you reach such (deservedly) rarified air.
Jim Rice, meanwhile, was at the opposite end of the spectrum, finally eking out a plaque in his 15th and final year of eligibility on the writers’ ballot. Rice’s selection was far more controversial than Henderson’s, with his selection on ballots, rather than his omission from them, sending shivers through what Jon Stewart calls the Blogonet. The argument against Rice starts with the idea that he was a sentimental pick due to being in his last year of eligibility (on the Baseball Writers ballot at least). Some have argued that he played in a hitter’s park and did not produce on the road. Some will have said his career was too short, he was a defensive liability, and that he was a slug around the bases. Some (including me) would argue that his teammate and follow outfielder Dwight Evans was a better player and he fell of the ballot years ago.
As Rob Neyer most acerbically put it, "[W]e can simply add him to the list of good players -- Bruce Sutter, Catfish Hunter and Orlando Cepeda come to mind -- who don't really belong in the Hall of Fame but are there anyway…[T]he election of Rice will do little to lower the standards of the institution, as it's unlikely that players like Dave Parker, Albert Belle, Dick Allen and big Frank Howard now will be knocking on the Coop's door (even though, it should be said, all of them were at least Rice's equal)." To quote Tina Fey, “Cat sound!”
Wow, Albert Belle is a hard case to make, but luckily for Neyer, he rarely deigns to make it. And what? No mention of the execrable choices like Tommy McCarthy or the spate of other Veterans Committee choices? (More on Joe Gordon in a minute) Besides, why all the shock and awe when Rice has been at 50% with an extremely slow bullet since 2000 and missed plaque-itude last year by just 16 votes (72.19%)?
So were the pedantic pundits right? Were Rice’s election and Henderson’s omissions so galling? And what of the Veterans Committee selection Joe Gordon, whose selection caused nary a ripple of consternation? Consider that Gordon played just 11 seasons and played 128 or more games in a season just eight times. He was no better than an average major-league hitter (by adjusted OPS) by the age of 34 and was out of the game at 35. He did have an MVP season in 1943 and missed two full seasons to WWII, but was atrocious (park-adjusted OPS 21 points below average) when he returned in 1946. Finally, it took gerrymandering the Veterans Committee worse than a congressional district (I am talking to you, Texas 22nd) to get him into Cooperstown. Getting nine required votes out of 12, as he did on the pre-1943 committee (actually 10), was much easier than getting 62 out of 82 votes as was required on the previous Veterans ballot in 2007. Gordon got just 10 votes (or 12.20 percent) on that, the 2007, ballot. Gordon, in fact, never received more than 28.53 percent of the vote in any Hall election in 17 previous tries, and that—the previous high—was in 1969 when the writers still owned his fate. But the veterans have not voted in a major-league player since Bill Mazeroski in 2001, at least two reorganizations ago when still-capitated Ted Williams ruled the Veterans nest.
But why are the hue and cry anyway? Don’t we have a means to determine the Hall-worthiness of these players? Of course, we do: Bill James established a five-step method 15 years ago in his category defining “The Politics of Glory (later renamed “Whatever Happened to the Hall of Fame”). His methodology, which has become so basic to the Hall argument it was added to Baseball-Reference.com’s player page canon, is based on five methods (p. 64):
1. Similarity Scores
2. The Hall of Fame Standards List
3. The Black Ink Test
4. The Hall of Fame Career Monitor
5. Fibonacci Win Scores
Most readers will be familiar with the first four, but Fibonacci Win Scores might need some explanation. They are “a method for dealing with starting pitchers … although [the method] is embroidered with an essentially irrelevant system for dealing with hitters.” (p. 68) James attempts to explain pitchers’ winning in three dimensions (by win total, winning percentage, and games over .500). James seems to have little confidence in the method and later introduced a more effectively way to look at overall player impact, Win Shares, that we will discuss later.
If you compare Rice and Henderson based on James’ methods, they do not appear to be as dissimilar as the controversies above would have you believe. Henderson passes all of the tests. Rice passes all but Hall Of Fame Standards (7.1 points short of the 50 points needed. However, if you throw in the Baseball Reference invention called the Gray Ink Test, based on James’ Black Ink Test but looking at the top 10 per league and not just the league leader, Rice, who passes the test, pulls even with Henderson, who does not—he is one point shy at 143.) Meanwhile, the uncontroversial Gordon fails each and every test for a perfect zero score.
My question is whether James’ methods still apply or whether there are some new ways to look at Hall candidates to determine their worthiness. I would propose that there are further measures that can buttress James’ core set. As James said, “As to how to find potential Hall of Famers, I argue for the democratic, as opposed to the autocratic, interpretation of baseball statistics…I advocate…what I call the democratic use of baseball statistics: Let them all speak.” (p. 74 - 75)
And I must say that the timing is appropriate given that James wrote his book as Phil Rizzuto entered the Hall and James made him “the central figure in the book, the player that [he used] to illustrate the statistical techniques, to examine the logic, to trace the controversy.” (p.63) Now, Rizzuto’s former Yankees keystone mate Joe Gordon has been selected for the Hall.
Consider first that in the last 15 years we have seen a spate of first-year eligibles going straight into the Hall while, exceptions like Rice aside, very few players who languished in the purgatory of the writers’ ballot, got into the Hall. Of the 23 players who were selected by the writers in the last 15 years, 14 were elected in their first year of eligibility. That is a rate of 60.87 percent, almost twice as high as the rate prior to James’ writing book (31.33 percent). A player who has been selected by the writers in the last 15 years has had to wait about 2.52 years on average, whereas a player prior to that had to wait about twice as long (4.82 years).
The first new method I propose is based on another Bill James statistic, Win Shares. The average Hall of Famer amassed 337.32 Win Shares. The stat has remained fairly static through the life of the Hall if one makes allowances for the first half decade, the Thirties, when the backlog of great players was being picked through and the makeup of the Hall was being determined.
Below is a table by decade of the average Win Shares per player elected to the Hall within that decade along with the number of players selected, the average WS at the end of the decade for all Hall of Famers, and difference between the end of decade figure and the average overall (337.32):
|Decade||Avg Win Shares||# Players||Avg as of decade end||Diff||% Diff|
Note that within a little over a decade the makeup of the Hall based on Win Shares was pretty well established (just 12 percent off). The figure has been asymptotically approaching today’s mark ever since and has remained largely unchanged since the end of the 1970s.
I devised a standard by which Win Shares can be used to deliver a letter grade to each candidate, A through F (but no E for Excellent). Below is a table of
Note that there are no eligible candidates with 400 or more Win Shares, Grade A, who is not in the Hall (FYI, Pete Rose is ineligible).
The B range (300-399 WS) drops off dramatically to 65 percent of eligible candidates in the Hall, so I split it into B+ and B- to get a clearer picture. Of the B+ candidates, just seven (16 percent) are not in the Hall. This consists of 19th Century/turn of century players (Tony Mullane, Bill Dahlen, and Sherry Magee) and fairly recent players (Darrell Evans, Rusty Staub, Lou Whitaker, and Tim Raines, who just became eligible in 2008).
Just 57 percent of the B- Players are in the Hall, along with 25 percent of the Cs, 5 percent of the Ds and none of the Fs. If you graph that, it looks like a nice bit of a sine wave from 100 percent down to 0 percent.
I would say that this is a fairly good predictor of Hall-worthiness. Incidentally, this method puts this year’s candidates into clearer focus. Henderson’s 535 Win Shares (Grade A) did indeed make him a lock when compared to Jim Rice’s 282, which put him in the range (Grade C) in which very few candidates, just one in four, acquire a plaque at Cooperstown. Joe Gordon (242 Win Shares) was also a C-grade candidate and a low one at that. Aside from Kirby Puckett in 2001, the last Grade C player selected by the writers for the Hall was Hoyt Wilhelm in 1985.
However, one could argue that Win Shares alone should not tell the whole story. Using Win Shares alone is too facile. Consider that having an egregiously average 20-year career tallying 10 Win Shares a year—though some would argue that 20 years at any level would be an accomplishment—would put you in the 200-Win Share range, a place at which players can start to get serious consideration for the Hall. Koufax had just 194 Win Shares. Ross Youngs, Catfish Hunter, and Rock Ferrell all had 206.
The point is we want to develop a system that selects the Koufaxes and ignores the Ross Youngses (or is it Youngsi?). The way to do this is by looking at how well the player performed above the established baseline level of his fellow players.
We need to look at the Win Shares Above Bench (WSAB). Using this method, players, like Koufax, who performed well above his contemporaries but for a short period will perform well and players who performed slightly above his contemporaries over a long career will perform well. The players who will not stand out will be those who played at or below average no matter the length of their careers.
I held these players to a higher standard than just the average ballplayer of his day. I used only starting position players (one per position per team and year, playing a min. of half his team’s games) and starting/relief pitchers who played a significant role in his team’s games (starting or appearing in at least one tenth of his team’s games).
Using this standard (WSAB), the top Hall of Fame- eligible players are as follows, and they are all in the Hall:
The truth is that the highest ranked player in this method not in the Hall is Bert Blyleven at 216 WSAB. The only other player at 200 or more WSAB is Wilbur Cooper at 205.
Below is a breakdown of eligible players—in total and in the Hall of Fame—by WSAB:
|WSAB Range||Count Players||Count Hall||% in Hall|
The dividing line is also the average WSAB across all Hall of Fame players: 149. If you had 150 WSAB or more, then you have at least a 50-50 chance to get in the Hall. In actuality there are 108 players with 150 or more WSAB and 78 (72.22%) are in the Hall.
If you failed to register 150 Win Shares Above Bench over the course of your career, you had about a one-in-20chance of being elected to the Hall (124 out of 2414 eligible or 5.14 percent).
I propose using a method that measures by WSAB. If a player amasses 200 or more WSAB over his career, he is a virtual lock for the Hall (43 out of 45 or 95.56 percent). If a player hits 150 WSAB, then he is at least a 50 percent chance.
I propose one test for 200 WSAB and another for 150. If a player collected at least 200 WSAB, he passes both. If he hit just 150, he passes just one. If he failed to garner at least 150 WSAB, he fails both.
I see these two methods to evaluate player replacing the half-hearted Fibonacci Win Scores in the James canon.
Of the other four methods, I would say that the Hall of Fame Standards and the Similarity Scores should remain unchanged. The Hall Standards looks at career standards that have remained largely unchanged over time. Yes, there are standards that have become somewhat tarnish, such as 500 home runs in the steroid era. However, for the most part the bulk of these career standards still retain their cachet. Besides if we adjust, say, for era, what would relative standards really mean anyway? The same basic argument can be made for Hall of Fame Monitor, which is based on career and single-season milestones. Next week, I'll look at the Black Ink Test and Similarity Scores, and what can be done to improve them.
My site with my opinions, but I hope that, like Irish Spring, you like it, too.