Using External Factors to Assess Value
by Asher B. Chancey, BaseballEvolution.com
July 30, 2009

One of the greatest challenges when ranking players is trying not to be fooled by their statistics. Looking at a player’s statistics, you want to know whether the statistics purely represent the abilities of the player, or whether they represent factors that have nothing to do with the player that makes that player look better than he should.

This deception manifests itself in many ways. Primarily, there is the home ballpark factor – baseball is the only major American sport in which the dimensions of a playing field are not strictly regulated. As such, what it takes to hit a homerun in one ballpark can be quite different in another ballpark. For the most part, statistics are most deceptive in regard to whether a player would have been able to put up the numbers he put up in a different ballpark (or conversely whether his numbers would have been better in another park).

The second major factor which can distort a player’s statistics is era. A player in the Delgado Era will be more likely to hit a homerun than a player in the Downing Era. A player in the 1920s will be more likely to hit a homerun than in the 1910s.

These first two factors have been grappled with quite capably by baseball statisticians and, according to my understanding, when you say “OPS+” or “adjusted batting runs,” you are referring to numbers which have been adjusted for ballpark and for era.

But these are not the only factors which can distort a player’s statistics. There are a few more factors which, although subtle, remain quite important and can seriously misrepresent a player’s abilities. For example:

Time missed due to outside factors. These factors are essentially limited to wars, segregation, other leagues, and strange fluke diseases. It is important to realize that players have missed playing time due to World War I, World War II, and the Korean War; they have missed time due to the Negro Leagues, the Mexican League, the Federal League, and baseball in Asia. And while injuries are a part of the game, when players succumb to diseases or disasters – meningitis, car wrecks, suicide, plane crashes, earthquakes, etc. – a career cut short by those factors improperly influences our perception of a player and his abilities.

One thing is for sure – the total measure of Roy Campanella’s value cannot be derived from his career statistic totals. Campanella missed, probably, five or six seasons at the beginning of his career due to baseball’s color barrier, and had his career cut short by an automobile accident which left him paralyzed from the shoulders down (we’ll leave aside the question of whether his numbers were benefited by the premature end to his career). For that matter, George Sisler had half his career impacted by a bout with sinusitis, Addie Joss died of meningitis, and Roberto Clemente died in a plane crash. Those factors must be taken into account.

Excellence derived from a supportive lineup. There can be no doubt – zero, zilch, none – that playing in the same lineup with an elite player can bring out a career year in a guy. If a player is a slightly-above average player and then joins a team with one of the top 20 players of all time, or even a roster full of excellent if not elite players, and becomes an All Star and an MVP, then his abilities are probably being misrepresented. Guys like Jeff Kent, Roberto Alomar, Joe Morgan (I don’t mean to be picking on second basemen here), Davey Johnson (no, really), Scott Rolen, Roger Maris, and Jim Edmonds all performed better, or in some cases significantly better, when they were in lineups with future Hall of Famers than they did during other parts of their careers. This is something of which we must be mindful.

Beware the breakout season. If a player is ordinarily quite competent and a contributor to the team, and then busts out for a season like none he’s ever enjoyed before, chances are it had more to do with context than with the individual player’s ability. Whether it is Andre Dawson playing an abundance of day games in the abnormally explosive 1987 season, Brady Anderson blasting 50 homeruns in 1996, or Jim Gentile having one of the great seasons of all time in 1961, the breakout season is usually explained by something other than “the guy just put it all together for a year.”

In particular, maybe a subset of this rule, there is the rule that one should beware any elite play in certain seasons. We know for sure that certain seasons were explosive, and not without justification. If you see a guy blowing up in 1930, 1961, 1969, 1977, 1987, 1994, 1996, 1998, or 2001 in a way that he never did before or since, then you can at least be suspicious (you are at least on notice) that something may be awry.

Having now been through the factors, one may wonder how we deal with them. I have one suggestion: adjusted batting runs is one of the two or three best statistics we have for measuring a player’s overall statistical value. So, what we can do is look at a player’s season-by-season accumulation of adjusted batting runs, and determine whether, for example, Brady Anderson accumulated an adjusted batting runs total in 1996 – the year he hit 50 – that was significantly out of proportion with his career total. Let’s take a look:

Player	Year	Best	Total	%
Brady Anderson	1996	50.1	110.0	46%

What this table tells us is that Brady Anderson derived 46% of his career adjusted batting runs from one single season, which by the way was nicknamed "The Year of the Homerun" at the time. So, chances are pretty good that Brady's 110.0 adjusted batting runs are probably overstating his value, because for all but one season of his career, he was really a 60 ABR guy.

As another example, let's look at the following players from the 1930 season:

Player	Year	1930	Total	%
Carl Reynolds	1930	34.8	30.2	115%
Fred Lindstrom	1930	40.8	66.7	61%
Ed Morgan	1930	43.9	85.9	51%
Lefty O'Doul	1930	63.4	210.8	30%
Hack Wilson	1930	75.9	317.3	24%
Babe Herman	1930	71.1	324.3	22%
Chuck Klein	1930	62	337.8	18%
Bill Terry	1930	60	333.4	18%
Al Simmons	1930	61.1	377.1	16%
Earle Combs	1930	36.2	229	16%
Gabby Hartnett	1930	35.3	242.9	15%
Joe Cronin	1930	34.9	240.7	14%
Kiki Cuyler	1930	36.8	264.4	14%
Frankie Frisch	1930	18.5	142.5	13%
Sam Rice	1930	18.6	160.4	12%
Wally Berger	1930	29.8	260.3	11%
Mickey Cochrane	1930	28.7	256.2	11%
Goose Goslin	1930	35.9	350.3	10%
Lou Gehrig	1930	93.8	982.9	10%
Charlie Gehringer	1930	32.4	354.6	9%
Earl Averill	1930	26.6	310.6	9%
Jimmie Foxx	1930	54.1	782.9	7%
Babe Ruth	1930	95.2	1387.5	7%
Harry Heilmann	1930	34	539.2	6%
Mel Ott	1930	49.2	793.8	6%
Tony Lazzeri	1930	12.5	204.8	6%
Paul Waner	1930	28.8	488.8	6%
Jim Bottomley	1930	1.6	256.7	1%

I am not sure how clear or obvious this is, but anything over 20% of a career value in a single season is very high. On this list, you'll notice that Babe Ruth had 95.2 adjusted batting runs in 1930, and that was only 7% of his career value. But Babe Herman (22%) and Hack Wilson (24%) basically made their careers in 1930 alone. Given that we know that 1930 was the most offensively explosive season of all time (particularly in the National League), these numbers confirm the fact that Herman and Wilson's careers are distorted by 1930.

In my opinion, if Hack Wilson accumulated 317.1 adjusted batting runs in his career, but almost a quarter of them came in one season, then his abilities as a player have been inflated by that one season, and he likely shouldn't be compared to players with a similar number of adjusted batting runs.

Here are some other notable individual seasons:

Player	Year	1930	Total	%
Brooks Robinson	1964	37.5	57.3	65%
Davey Johnson	1973	33.5	74.6	45%
Jim Gentile	1961	63.8	158.6	40%
Alan Trammell	1987	48.1	123.9	39%
Roger Maris	1961	54.2	187.2	29%
Sammy Sosa	2001	92.6	330.6	28%
Luis Gonzalez	2001	70.6	270.8	26%
Nomar Garciaparra	2000	48.0	188.7	25%
Ryne Sandberg	1992	36.8	150.5	24%
Dale Murphy	1987	51.4	224.8	23%
Norm Cash	1961	85.1	377.1	23%
Toby Harrah	1982	38.1	181.9	21%
Jeff Kent	2000	60.4	289.4	21%
Derek Jeter	1999	55.4	277.6	20%
John Olerud	1993	77.0	386.6	20%
Arky Vaughan	1935	71.6	364.3	20%
Scott Rolen	2004	44.8	246.9	18%
Ernie Banks	1958	46.8	259.8	18%
Darrell Evans	1973	49.6	278.6	18%
Ernie Banks	1959	45.5	259.8	18%
Jim Edmonds	2004	56.1	328.7	17%
Moises Alou	1998	49.8	295.1	17%
Charlie Keller	1942	49.3	295.2	17%
Mark McGwire	1998	95.2	589.2	16%
Andre Dawson	1983	33.1	216.4	15%
Ken Singleton	1977	55.4	365.6	15%
Todd Helton	2000	66.5	450.9	15%
Reggie Smith	1977	51.8	355.7	15%
Orlando Cepeda	1967	49.9	343.9	15%
Bob Johnson	1944	55.6	392.4	14%
Wade Boggs	1987	66.1	475.0	14%
Wade Boggs	1988	64.6	475.0	14%
Carl Yastrzemski	1967	72.3	535.7	13%
Edgar Martinez	1995	74.2	565.9	13%
Tim Raines	1987	42.1	332.8	13%
Joe Morgan	1976	62.2	496.4	13%
Jim Thome	2002	73.3	587.1	12%
Alex Rodriguez	2007	69.2	575.7	12%
Ken Griffey, Jr.	1993	60.8	513.7	12%
Joe Morgan	1975	57.7	496.4	12%
Rickey Henderson	1990	62.4	544.4	11%
Harold Baines	1989	32.4	287.2	11%
Eddie Murray	1984	49.8	462.9	11%
Hank Aaron	1959	71.6	919.7	8%
Willie Mays	1965	62.6	844.4	7%
Ozzie Smith	1987	9.7	-132.3	-7%

You'll note that in their best seasons, Ruth, Hank Aaron, and Willie Mays were all in the same general class, seven to eight percent of their total ABR. These guys put up the numbers every year.

By the way, Ozzie Smith's number is what happens when you divide a positive single season total by a negative career total. If I were a math major, I'd know how to make that represent a number that comports with what we are doing. Basically, I suspect that -7% is the equivelant of well over 100% of total value.

So there you have it. We'll do some more guys later, as they occur to me.



Questions? Concerns? Comments? Asher lives in Philadelphia, PA, and can be reached at asher@baseballevolution.com.