Comparing eras in my system

[Triad]

Are any of you familiar with rating systems which incorporate factors for measuring the strength of competition in a given year? I've been toying with this in the past, and have gotten fairly balanced results based on the philosophy, which is that in its infancy a competition will have fewer participants at a high skill level, and over the course of a few decades (dependent on social and economic trends), the average skill level will continue to improve.

One measurement for this is the variance between the high set and the low set of players in a given year. A higher variance indicates a weaker overall level of competition. This should be demonstrable with most any run value metric. What we see is the gap between the best players and the worst players in the league narrowed over time, until about the early '60s, where it has remained fairly steady since. There are blips for about 2-3 years with any expansion teams coming in and watering it down temporarily.

Another indication of this trend is in the variance between the winning percentages of the best teams and the worst teams in a given year. The same pattern develops over the course of history. (the number of teams in the league must also be factored in)

With the inclusion of blacks into the majors beginning in 1947, and finally culiminating in the early '60s, this seems to have made the biggest impact. It took about 15 years before blacks had been fully integrated. Of course, many other factors are intertwined, and increased player salaries is going to attract more who might not have attempted to play back in the early 1900s.

With the era adjustments in place, we can then plug them in to individual player values. Anyone who seriously believes that Ruth would bat .340 for his career today isn't paying close attention.

Without an era adjustment, we end up with an overabundance of early players as the all-time elite. I can go along with Ruth, Cobb, Wagner, Hornsby, Gehrig, and Johnson being at or near the top of their all-time peers, but unless we make a significant adjustment, any raw metric will tend to give too much credit to the next tier of old-timers. It's antithetical that most of the best players played before 1940. With the increase in population, high-level foreign programs, number of major league teams, economics, etc., we should expect that it would produce more elite players than a hundred years ago. For every Hall of Famer produced in the 1920s, for example, we should expect about two produced in the 1990s.

My formula takes into account various other factors, such as the team environment a player was in. It also adjusts for full seasons missed in between or next to quality seasons, also taking into account a player's age. This credits players who lost time in the war, or blacks who started late (Robinson, Campanella, Minoso), and any of those who had career-ending injuries in the middle of their prime. It's a graduated scale, crediting them for a low estimate of what they could have been expected to produce.

I give a great deal of extra weight to high-level seasons. Thus, five or six high level seasons are worth more than the rest of a player's career combined (for most HOF-level players). Long careers with many average seasons don't have any appreciable value. Seasons under the league average have a zero value.

My preliminary conclusion is that Ruth's value is somewhat overstated, as he was not measured against any of the Negro League players who would be playing in the same league today. And he didn't face many left-handed relief pitchers coming in for the sole purpose of facing him. In addition, the white players themselves were overall of lower quality than in recent decades. I rate Ruth 2nd among right fielders behind Hank Aaron. I do rate Gehrig first among first basemen, Hornsby first at 2B, and Wagner first at SS, with A-Rod coming on strong. He still has a ways to go, though. At 3B is Schmidt, and though I don't have a rating for him yet, I've got Gibson at catcher.

I have not updated the ratings more recently than 2000, so I don't have an adequate rating for Bonds. It's clear to me, though, that he'll be in the top three, with Aaron and Mays. I have a feeling he's probably going to be rated #1 overall. For pitchers, I have the first five as Johnson, Grove, Koufax, Clemens and Feller. In the bullpen, I have Rivera first.

Highest rated fielders at each position:
1B - Keith Hernandez
2B - Bill Mazeroski
SS - Rabbit Maranville
3B - Brooks Robinson
C - Ivan Rodriguez
LF - Rickey Henderson
CF - Richie Ashburn
RF - Roberto Clemente

Other conclusions I've made:
-Center fielders' defensive value is underrated.
-Strikeouts for pitchers are overrated, though they are often a good indicator of a decent pitcher.
-There were not very many elite players in the 1800s or 1910s.
-Designated hitters' overall value is exaggerated.
-The high-level pitchers have quite a bit of control over their W-L record.
-Saves combined with ERA and IP are an excellent gauge of a reliever's ability.
-R and RBI give more context to generalized rate stats.
-Offense is worth about 52% of the game, pitching 35%, and fielding 13%.
-Third basemen can usually easily convert to first base (meaning they can handle either).
-Shortstops can usually easily convert to second base.
-Center fielders can usually easily convert to left or right field.
-No other positions can usually easily convert to another.
-Ron Santo and Dick Allen are criminally underrated.

[jtur88]

One factor you need to take into account is that before about 1940, virtually every gifted athlete that played professionally was a baseball player. The NFL was marginally viable before the 1940's and the NBA non-esixtent.

Today, about 70% of gifted athletes are not playing baseball at all. It would be impossible to estimate, but it would not be unreasonable to guess that maybe a dozen players who excelled as pro football or basketball players could have been baseball HoFers. We have no idea how many Bob Gibsons or Dave Winfields or Jackie Robinsons elected other sports. We know that there are fewer African Americans playing MLB today than there were a generation ago.

So, even though we can surmise that there are many more elite athletes today than a century ago, we must acklnowledge that many of them were absorbed by other sports, when that became a viable option.

[WilsonC]

BP's WARP2 and WARP3 (Wins Above Replacement Player) metrics adjust for league strength. I don't know much about their methodology, but it's basically:
WARP1 - A player's value relative to his league.
WARP2 - WARP1 adjusted to a historical context, considering league strength.
WARP3 - WARP2 adjusted to a normalized schedule of 162 games.

With their metrics, a player like Cobb loses a sizeable chunk of value (somewhere around 15-20%) due to a league difficulty adjustment, but also gains some of it back due to the shorter schedule in his years.

[Triad]

Originally Posted by jtur88

One factor you need to take into account is that before about 1940, virtually every gifted athlete that played professionally was a baseball player. The NFL was marginally viable before the 1940's and the NBA non-esixtent.

Jim Thorpe was probably the most gifted, but that didn't help him much in baseball. Many other athletes have found this out.

Originally Posted by jtur88

We know that there are fewer African Americans playing MLB today than there were a generation ago.

This is because the number of Latino and Asian players has increased...

_______1996_2001_2006
Whites .... 62% ... 59% .. 60%
Blacks ..... 17% ... 13% ... 8%
Hispanic .. 20% ... 26% .. 29%
Asian/Other 1% .... 1% .... 3%

Originally Posted by jtur88

It would not be unreasonable to guess that maybe a dozen players who excelled as pro football or basketball players could have been baseball HoFers. We have no idea how many Bob Gibsons or Dave Winfields or Jackie Robinsons elected other sports.

That's probably true, but it works both ways. For every one we lose, we could also be gaining others. Some major league players could have played other professional sports but didn't.

Originally Posted by jtur88

So, even though we can surmise that there are many more elite athletes today than a century ago, we must acklnowledge that many of them were absorbed by other sports, when that became a viable option.

The thing about this though is that baseball isn't often about pure athleticism. There aren't a lot of real fast players in baseball, or a lot of real strong or real big players, or real tall players. There are always going to be a group of athletes who could be good at various sports, but that doesn't seem like a significant group. Who in the NFL or NBA do you think could be a regular in the major leagues? I would say not too many. Wouldn't we see more crossing over (even at the college or minor league level) if it were such a phenomenon? Some major leaguers played college football, but most of them had no future with the NFL.

For basketball players, we're talking pretty much the shorter guards (6' 4" and under) who would fit the baseball mold. For football players, nobody on the offensive or defensive lines has any special skills that would translate to baseball. Their contribution to football is strength and size, both of which have little value in baseball unless there's a lot of coordination, flexibility, and discipline. Every play in football is about road rage. In baseball, you can't use that mentality. You have to be calm and collected. Some of the people here at FanHome are probably just as qualified to play in the major leagues as football linemen or linebackers would be. As for backs, some of them could be converted into baseball players, like Brian Jordan, Bo Jackson, Deion Sanders. Have there been very many, though, who've made a big impact? This seems like just a blip.

I would argue that the skills involved in hitting off a major league pitcher are more discriminating than the skills involved in playing in the NFL or NBA. I would liken it to being a PGA tour golfer. You have to be very precise and very disciplined.

And when it comes to pitchers, baseball is about anything but athleticism. Having athletic ability gives you no obvious advantage in being a pitcher. Most pitchers in the majors are rather mediocre athletes.

It's true that basketball and football are on the stage as well, but aren't there more guys involved in sports in general? And think of all the other countries which are contributing to the pool of major leaguers in baseball.

Originally Posted by WilsonC

BP's WARP2 and WARP3 (Wins Above Replacement Player) metrics adjust for league strength. I don't know much about their methodology, but it's basically:
WARP1 - A player's value relative to his league.
WARP2 - WARP1 adjusted to a historical context, considering league strength.
WARP3 - WARP2 adjusted to a normalized schedule of 162 games.

With their metrics, a player like Cobb loses a sizeable chunk of value (somewhere around 15-20%) due to a league difficulty adjustment, but also gains some of it back due to the shorter schedule in his years.

Thanks, Wilson. I'll be interested in checking it out. My system actually deducts for leagues that played fewer games per season, because the laws of disparity say that as the frequencies increase, this causes the extremes to move more toward the norm. So in a 154-game schedule or 140-game schedule, the leaders will have an advantage of not having the extremes come toward the center. We see this in the strike-shortened seasons of 1981 and 1994. There are more extremes in those seasons because they were shorter. If we went to a 180-game schedule, we'd have to give those players more credit, not less. I'm writing my Congressman. Does Baseball Prospectus have a Congressman?

[WilsonC]

If you're talking ability, then a shorter schedule does allow for more variance in rate stats per season - so I could see giving less weight to outlier seasons. On the flip side, when we're talking value, scaling up for shorter seasons is the correct approach. If a player creates the same number of runs over a shorter season, or the same number of wins, then he's creating more runs-per-game and contributing a higher percentage of what's needed to win a pennant. That's the nice thing about showing the 3 different metrics, is it allows you to look at different methods of evaluating players using the same system, from a few perspectives.

[jtur88]

Re: drain to other sports. A couple of more factors that ought to be considered is that prior to about the 1950s, virtually every athletic boy in the US played some baseball, and a high number of high schools had a baseball program. Nowadays, schools in northern climates seldom have a baseball program, with many athletes getting relatively little exposure to the game. Given a choice of playing football or basketball in front of a packed house, or playing baseball where their mom is there because she has the SUV, shift a huge incentive away from basesball.

Baseball is very much a white boy's game in America. I lived in a town in north Florida that was more than 50% black, but the community college (with a majority black enrollment) did not have a single black player on its baseball team, and there were only a few blacks in the entire league they played in. I've done head-counts at minor league ballparks in cities with substantial black populations (like Greenville SC and Battle Creek MI) and could count the black spectators on my fingers. Only 12% of American-born playes in MLB are black, which is less than the population at large. This leads to the inescapable conclusion that athletic blacks (even those with physiques more suited to baseball) are indeed opting (usually unsuccessfully) for football or basketball, perhaps never bothering or never having the opportunity to try to acquire the skills specific to baseball.

[Triad]

Originally Posted by WilsonC

If you're talking ability, then a shorter schedule does allow for more variance in rate stats per season - so I could see giving less weight to outlier seasons. On the flip side, when we're talking value, scaling up for shorter seasons is the correct approach.

I'm not clear on why you're separating ability and value. Pardon my ignorance on that subject. If we're looking at 154 and 162 game seasons, having 8 fewer games makes it somewhat easier to maintain a high rate relative to the league.

Originally Posted by WilsonC

If a player creates the same number of runs over a shorter season, or the same number of wins, then he's creating more runs-per-game and contributing a higher percentage of what's needed to win a pennant.

But it's easier to do it the shorter the season is. And we're not really comparing two players as having the same number of created runs, with one in a shorter period. What we're comparing is two players with the same rates, but one done in a shorter period. All else being equal, which is what should then apply more credit to longer seasons. Your example assumes that the player in the shorter period already had better rates, which is an unbalanced comparison.

So by your rationale, players in 1981 and 1994 should get heightened credit for having each game supposedly meaning more. Is that what you believe?

Originally Posted by WilsonC

That's the nice thing about showing the 3 different metrics, is it allows you to look at different methods of evaluating players using the same system, from a few perspectives.

I do agree with having the three perspectives. I've used the same ones going back to the early 1990s, though the application of them might have differences.

I've seen Rob Neyer make reference to era adjustments, but it seemed his measurements (possibly based on a Jamesean model) were quite conservative.

Originally Posted by jtur88

Re: drain to other sports. A couple of more factors that ought to be considered is that prior to about the 1950s, virtually every athletic boy in the US played some baseball, and a high number of high schools had a baseball program.

But prior to 1950 or so, a man often couldn't make a good living as a baseball player unless he became a major league regular. So, that made it very difficult to hold down a job and play at the same time, if your minor league team was travelling out of town a lot. It stands to reason that a great deal of worthy players back then never had a good chance to be even a minor leaguer for very long.

Originally Posted by jtur88

I lived in a town in north Florida that was more than 50% black, but the community college (with a majority black enrollment) did not have a single black player on its baseball team, and there were only a few blacks in the entire league they played in.

Keep in mind also that there are quite a bit fewer blacks in college basketball (percentage-wise) than there are in the NBA. They seem to have a better ratio of retention into the higher levels.

Originally Posted by jtur88

I've done head-counts at minor league ballparks in cities with substantial black populations (like Greenville SC and Battle Creek MI) and could count the black spectators on my fingers.

I can definitely concur with that. The black fan base in baseball does seem quite low. And while related to player participation, I don't think one hinges on the other.

Originally Posted by jtur88

Only 12% of American-born playes in MLB are black, which is less than the population at large.

That's the same as the population... 12%.

Originally Posted by jtur88

This leads to the inescapable conclusion that athletic blacks (even those with physiques more suited to baseball) are indeed opting (usually unsuccessfully) for football or basketball, perhaps never bothering or never having the opportunity to try to acquire the skills specific to baseball.

I agree with this, though we have to keep this in the context of the early 1900s, when there were no blacks in the majors at all.

[jtur88]

"But prior to 1950 or so, a man often couldn't make a good living as a baseball player unless he became a major league regular."

I don't think that argument has any relevance, because even pre-1960, where were probably only about a dozen MLB players who did not need to have an off-season job in order to support their families. There are HoFers from the '50's who had winter jobs during the years when they were all-stars.

Today, in the lower minors, a player get paid less than his little sister who has a counter job at Burger King. Class A players fall in one of two groups---family sufficiently well off that they don't need a real job in the summer, or Latin Americans who have nowhere else to go but San Pedro de Macoris.

There are very few blacks playing in the independent leagues. Most teams I've seen, just one or two. Or none. They can't afford the luxury of the hobby, which is largely what independent league baseball is. The typical indy league player is a kid who was drafted by a MLB team, got enough signing bonus to pay for a Ford Bronco, got released after a year or two in Bluefield WV, still has the Bronco, and doesn't want to go home a failure and work in his uncle's insurance agency..

[WilsonC]

Originally Posted by Triad

But it's easier to do it the shorter the season is. And we're not really comparing two players as having the same number of created runs, with one in a shorter period. What we're comparing is two players with the same rates, but one done in a shorter period. All else being equal, which is what should then apply more credit to longer seasons. Your example assumes that the player in the shorter period already had better rates, which is an unbalanced comparison.

The particular metrics mentioned are a seasonal total value metric, not a rate metric. Using a strike-year analogy, a player who created 10 more wins over the replacment during a 110 game strike shortened season is clearly more valuable than one who creates the same number of wins over a 162 game season, whereas a player who creates about 1 win per 16 games played would not be more valuable in the shorter season. That's the adjustment made here - it's not suggesting that a .800 OPS hitter has more value in the shorter season, but rather that his total value needs to be extrapolated over a full season in order to balance the differences in schedule. If any adjustment is to be made for rate differencials in a shorter schedule, it would make sense for that to be studied and included in the league-difficulty portion, which I'm not sure if BP does or not.

But it's easier to do it the shorter the season is. And we're not really comparing two players as having the same number of created runs, with one in a shorter period. What we're comparing is two players with the same rates, but one done in a shorter period. All else being equal, which is what should then apply more credit to longer seasons. Your example assumes that the player in the shorter period already had better rates, which is an unbalanced comparison.

Originally Posted by Triad

So by your rationale, players in 1981 and 1994 should get heightened credit for having each game supposedly meaning more. Is that what you believe?

With a rate metric, no. For a cumulative metric, yes.