After much debate recently on this subject, I’m gonna throw this topic out for discussion: when comparing players of different eras, does adjusting players’ stats to account for changes in the game, the ball, the stadiums, the steroid use make any useful difference? I mean, in some cases we’re almost trying to compare apples and oranges. Can we adjust Honus Wagner’s numbers accurately to compare him to Alex Rodriguez when they play almost two entirely different games?
In Wagner’s time they played on big fields with soft, scuffed balls, with no carpet, crappy gloves, and pitchers throwing off higher mounds. Hitting for average was the name of the game, not hitting for power. A-Rod plays in smaller parks with pristine balls, amazing mitts, astroturf and domes. The talent pool is watered down, especially the pitching. Home runs are favored over hitting .350 (although these days you can do both).
I don’t think you can accurately compare players from two different eras. This is especially true for pitchers, who have gone from pitching every other day to pitching one in five. You can’t say that Cy Young is the best pitcher on the basis of his 511 wins, simply becuase the nature of the game has changed to the point where (barring another fundamental shift in the game) no one is going to ever get close to that number again. So, can you realistically compare Cy Young to Roger Clemens to determine which is better? IMHO, the answer is no.
However, the fact that Cy Young pitched more often also shows that he has the stamina that many pitchers these days do not. When a pitcher pitches off 3 or sometimes even 4 days rest, it is considered the act of a true workhorse. Back then, if you waited 4 days between starts, it was usually because you needed an extra day to shake off a broken arm or something.
Comparing players like Wagner and Rodriguez is simply not possible because of a myriad of factors. New training techniques, dietary supplements, gloves, bats, balls…the list goes on and on. The sport has changed so much that it is impossible to logically compare the talent of players in such different times.
Bushwah.
It was common for pitchers in those days to pitch 50+ games a season for several reasons; one of which was that it was harder to hit the ball hard and far. Today, nearly every hitter can go yard at any time; pitchers have to put maximum effort into every pitch. In Young’s time, pitchers made less effort on each pitch, only cranking up their heaters when there were men on base or the game was close. (Hence Christy Mathewson’s book “Pitching in a Pinch,” a terrific source.) See the strikout rates of that era.
Another reality is that most guys had shorter careers. Given the ethos and the economics of the time, a young pitcher would essentially go out there twice a week all season … and 80% were done by the time they were 27 or 28, their arms wiped out. There were a few freaks, such as Young, but they were the exception.
Nowadays, teams have a care for the pitcher’s long-term career (and so does the pitcher), in part because of guaranteed contracts. We still have freaks – I have no doubt that Greg Maddux could pitch 300-350 innings a year if he wanted to. But teams prefer to play it safe with those multi-million-dollar arms.
Finally, and related to #1, realize that there are way more pitches per inning now than then. Because walks and strikeouts were less common, it was quite normal for pitchers in the early decades of the century to average 10-12 pitches per inning or less-- i.e. maybe 100 pitches in a game. Nowadays, most average 16-18 per inning, meaning they hit 100 pitches around the sixth inning.
As far as the OP, I think you can certainly make some comparisons across eras by looking at how they did relative to their peers.
Ted Williams was a better player than, say Shane Spencer, because Williams was better compared to his peers than Spencer is to his. Williams won multiple awards from observers, was widely praised, and dominated statistical categories. Spencer does not.
The fact that Spencer is likely a better athlete – faster, stronger, etc. and plays in a different era is a moot point.
Obviously, theres’s a big disparity there. Comparing Wagner to A-Rod is much harder, and would no doubt depend on how much you are willing to accept performance-based (i.e. statistical) analysis.
FWIW, Rodriguez is clearly the best ever at his age: Wagner was not a fulltime shortstop until he was 30…
I’m not sure if one could make a direct, accurate comparison between players like Wagner and Rodriguez, as they played in such incredibly disparate eras with different rules, park dimensions, and equipment. However, one could measure the relative value each player had by first finding out what type of statistics the run-of-the-mill major leaguer had from each year. Then, by comparing the stats of players like Wagner and Rodriguez to what Joe Average accomplished, you’d have a crude measurement for displaying value over-and-above the league average, irrespective of what era the player actually played in.
There’s a statistic that statheads use, Runs Above Replacement Level (I think), that is used to indicate this type of value. I have no idea how it’s produced, however.
Oh, to address one common misconception…
You could make the argument that, when compared to Wagner’s era, the talent pool nowadays is much, much deeper than what it once was. The breaking of the colour barrier, an increased focus on international scouting, a general increase in North American population, and the sudden trend of Japanese players wanting to “prove” themselves in North America combined make a huge contribution to the amount of MLB-caliber players available to current markets.
RAR is one measure, but it’s not the end-all be-all. There are a number of things you can use, including the traditional batting average, etc.
They all work out generally the same, since research has established pretty clearly what sort of things create the most runs. But with a really close matchup, guy A can win by one measure, guy B by another.
Well, there’s Keith Woolner’s Value Over Replacement Player (VORP) and it’s defensively-adjusted sibling, Value Over Replacement Player + Defense (VORPD). Assuming you’re comfortable with run-denominated measures of offensive performance (Runs Created, Runs Produced, Batting Runs, etc.), Woolner’s explanation will probably make sense. If you don’t have any idea what those are, you can start with Woolner’s Stats Glossary, which gives a brief description of a wide variety of sabermetric statistics and formulas for their calculation, where feasible.
In a nutshell, VORP starts with a Linear Weights-style evaluation of the number of runs generated by a player for a particular season and then compares that against the number of runs one would expect to be generated during that season by any warm body that could be called up from Triple-A and plugged into a lineup spot (this is far different from a comparison to the average player in the league, who would have much better numbers than our hypothetical replacement). The difference between a particular player’s actual runs generated and the hypothetical replacment is VORP.
Certainly the overall talent pool available to the MLB has grown drastically over the years. However, the number of Major League teams, and their corresponding farm teams has also grown drastically. I think the state of pitching overall in The Show demonstrates that the pool has not been able to keep up with the demand. The elite pitchers continue to dominate the game, but the collection of chaff at the other end of the scale is massive – among pitchers with 100+ innings last year, 11 had an ERA under 3.00 while 33 were over 5.00 (only 3 of whom pitched in Colorado). In fact, MLB.com lists 105 pitchers who threw in the majors last year who stayed under 3.00 and 245 over the 5.00 mark – that’s 17% of all Major League pitchers getting top marks and 39% qualifying as batting practice.
Good point there. I’d like to see some of the big names from the middle of the 20th century hit against Satchel Paige in his prime. He took Josh Gibson down on three high heaters, you know. I’d like to see him face off against a repeat .400 hitter like Hornsby. Or to see Pedro against those guys.
You’re misunderstanding the nature of statistics. They are affected by playing conditions as much or more as by performance.
In 1968, the National League ERA was 3.43; in 1970 it was 4.52. Am I to conclude that in two years, suddenly all the pitchers were no good anymore? Or is it more likely that lowering the mound had something to do with it?
In 1930, the NL as a whole hit .303. Were they all great hitters then, who forgot how to hit by 1933, when the league average was .266?
These are extreme examples of a basic truth: what constitutes “good” changes over time. Hence the imprtance of comparing people to their peers.
Hence the fact that Mickey Mantle never stole more than 20 bases doesn’t mean he wasn’t fast: just that he played in an era that didn’t favor stolen bases much.
Vitrtually all of the changes in the game in the last 20 years have favored hitters: Weight training/steroids, smaller strike zones, smaller ballparks, the shift to lighter bats, pitchers discouraged from going inside … almost every change has helped hitters.
Make guys use the thicker-handled bats and call the strike zone the way the rulebook says and watch the runs per game drop.
And as for the talent pool … the US population has almost doubled from 1950 to 2002. That alone would account for expansion from 16 to 30 teams.
Not to mention that 25% of major leaguers are now foreign-born, and the vast majority of them would not have been in the majors in the pre-expansion era, and easily another 20% or more are african-american, and the majors were not fully integrated until the 60s.
The result is that the talent pool is much, much deeper today than in the past.
Comparison of the growth in population of the US with the growth in total number of MLB roster slots would seem to refute your argument, even without considering the effect of the widening of the talent pool to include far more Latin American and black players. Consider: in 1920, near the end of the “dead-ball” era when pitching and defense and “small ball” strategies ruled, the population of the US was approximately 105 million. By 1960, the US population was 178 million, or about 1.7 times the 1920 figure. In 1920, there were 16 major league teams. In 1960, there were still 16 major league teams. Assuming that the percentage of the population capable of playing major league baseball remained constant, there should have been sufficient talent in the US to field 27 major league teams at the same talent level as in 1920. By 2000, the population had grown to 280 million. Again assuming that the major-league-level talent rate in the population remained constant, there should now be enough Americans to support over 42 MLB teams at the 1920 level of play. Add in all of the black players who weren’t allowed to play in 1920 and the Latin American players who’ve been added to the talent pool in the last 40 years, and it seems absurd to contend that there’s been some precipitous increase in the number of minor-league quality players occupying places on major league rosters.
Moreover, if the talent pool were indeed diluted, why would this affect only pitching? Shouldn’t there be just as many mediocre-to-bad hitters in the game, neutralizing the effect?
In fact, the opposite of your contention is much more likely to be true: the overall level of talent in the game is significantly higher than in the early days, for a whole lot of reasons. Systematic and ongoing scouting, together with improved communications, mean that the players who’re signed and developed are more likely to be quality players, and not just guys who happened to have a single good day during the game that a manager’s buddy happened to see. The discrepancy between the rate of expansion of the talent pool and the rate of expansion of the number of roster slots available has forced the worst players out of the game at every level (there are far fewer minor league teams now than there were eighty or even forty years ago). Conditioning, systematic instruction, and other nominally rational player development approaches have increased the likelihood that players with natural talent develop that talent more fully.
The superstar players of today probably aren’t dramatically better than the superstars of eighty years ago, but the benchwarmers and utility players almost certainly are substantially better than their ancient equivalents.
So why the apparent difference in pitching performance? Because of real differences in the nature and level of hitting performance. The game goes through phases (the offensive explosions of the 20s and 30s and the 90s, the dominance of pitching in the dead-ball era, the mid/late 1960s, and the 1980s), but on the whole the changes in the game over the last century have favored the hitters. Better quality balls (changed more frequently to maintain visibility), changes in the approach batters take at the plate, greater strength, “scientific” teaching of technique, and a host of other factors have made even the worst hitters far more dangerous than even average players of the dead-ball era.
Eighty years ago, a team’s pitching staff typically recorded 500-600 strikeouts per season; these days, the figure for most staffs is well over a thousand. If the pitchers were so much better then, why were they striking out half as many batters as now? furt got it right: Mathewson, Johnson, and their contemporaries could count on facing several guys in any opposing lineup who, even if they hit the ball, were not going to hit it out, or even very far. So they threw strikes and let the defense take care of recording most outs. The downside risk if they left a fat one in the middle of the plate was far lower than now, when any player on a major league roster is capable of yanking a hanging curve or slider out of the park. And because there were fewer guys capable of hitting with any real power, if a guy did manage to get a hit, his chances of being driven in by another hit were much lower; consider this: in 1920, the NL league BA was .270, and the average number of runs per team per game was under 3.97, while in 2002 the NL league BA was .259, but teams scored an average of 4.45 runs per game. Any way you slice it, a higher league batting average and half as many strikeouts means a whole lot more balls in play, but the result was 10% fewer runs than now. Once guys started swinging for the fences and reaching them regularly, a pitcher who made the same number of bad (i.e., easily hittable) pitches as before would find the effect on his ERA to be much bigger. Another factor in today’s higher ERAs is the substantial decline in errors, and hence in unearned runs; the difference between runs allowed/game and ERA in the NL was about .86 in 1920, and was just over half that, at .44 in 2002.
Lessee. Let’s take something that’s very largely subjective, namely a player’s worth to a team – or their potential worth – and reduce it to a statistic. Create statistics on every player. Let’s assume they are robots whose value can be reviewed in an issue of PC World. Oh, I’m sorry, Sports World.
Now, let’s argue that our worthless, bogus statistics need to be calibrated according to era. No, wait. According to weather. WAIT, WAIT, according to race, since some players were obviously disadvantaged.
According to ballpark, since some stadiums were using astroturf?
Get. A. Life.
Stop trying to pretend sports – an emotional outlet – is connected to science.
SABR can come in handy in instances where you try to compare era to era. To understand the value a player had to his team (and to baseball), it would probably be a good idea to look at the measure of a player versus various statistical measurings and means of his era on the field.
I would suggest to you that the increase in population does not, by definition, mean that there has been an increase in quality players. There are far more activities available to athletes these days which has sent the top athletes in many different directions. The explosion of football, basketball, soccer and hockey, not to mention skateboarding and other “non-pro” sports, has pulled people away from baseball. As an example, I coach high school age baseball here in Calgary. Our league produced Chris Reitsma, now with Cinncinati, aa well as the Rangers Jeff Zimmerman. Two other local boys have seen action in the Majors recently. So we have a pretty good system. Since the mid-80s, Calgary’s population has increased by more than one third, however, we have actually been forced to reduce the number of teams in our league in the last few years due to a lack of registrations–the pool of kids playing baseball has shrunk even as the number of kids in the city has grown.
Further, I would contend that the overall calibre of hitting talent has eroded as well. The increase in strikeouts that you noted can also be accounted for by batters who are simply not able to put the bat on the ball consistently. I suggest to you that a guy like Jose Hernandez, striking out once per game while only batting in the .250s, would not have had an everyday job 20-30 years ago. I do agree with you that the development of the game has helped hitters and I would further suggest that the hitter-friendly nature of the modern game has helped to hide many of the weaker players.
Certainly, the level of training and athleticism in the modern player is well above the level of the turn-of-the-century player. We know that as a fact. But how do we compare the relative performance of these guys? VORP, as rackensack noted, is interesting. But I am still left wondering if there is another way, short of time travel, to effectively analyse these comparisons?
Sigh…another post swallowed by an errant keystroke.
Anyways, as I was saying: thanks, rackensack, for providing a link to Woolner’s explanation of VORP. I’d always wondered how statistics like VORP were measured, but I could never find (or more accurately, was too lazy to look up) a precise explanation.
Dread Pirate Jimbo, you do raise a good point when you claim that, despite the increasing “globalization” (for lack of a better word) of MLB, North American children in particular have an increasing number of alternate activities available to them, and that baseball may no longer be the top priority of most kids. And while I’m just beginning to wrap my brain around the concept that OPS (on-base percentage plus slugging average, for the uninitiated) is a more accurate indicator of a player’s value than batting average, and consequently am not really familiar with the more technical stats that sabermetricians use, you might want to check out some of the discussions that they have over at Baseball Primer. Amazing, incredibly detailed site.
Oh, and a few other comments…
Well, it’s not that large. A player who gets on base and hits homeruns is necessarily more valuable to a team than one who never walks and can’t hit the ball out of the infield. And we’re not “reducing,” but “evaluating” or “measuring.” Would you consider measurements like standard of living or intelligence quotient meaningless reductions?
Sports are definitely, without a doubt, connected to science. I can’t imagine how you might think they aren’t – do real-life effects like altitude, wind speed, etc. not have a demonstrable and measurable effect upon what happens on the playing field? I am definitely neither a scientist nor an expert in statistics, but this assertion just sounds ludicrous to me.
Whether it’s as worthwhile an area of study as, say, meteorology or quantum physics, is a different question altogether, as that seems to be what you’re suggesting with your fairly dismissive comment about baseball analysts not having lives.
[sub]On a complete side note, I think every single one of my GD posts have dealt with baseball. Maybe I do need a life, after all.[/sub]
True enough, but it’s a fair assumption to make. I’ll give you 100 guys and you can pick your best nine, I’ll take 300 and pick my best nine. I might lose the game, but I like my odds.
I would also point out that today, virtually anyone who can play professionally does, because of the money. It was not uncommon 50-70 years ago for guys who were successful at college or the sandlot to forgo a professional career (or quit after a couple of years) because they could make way more money as doctors or lawyers.
Moreover, the progression of players to the big leagues is more merit-based today. In the free minors era, players would often be in the minors for a couple of years even when they were clearly good enough be be in the majors. Minor league teams were legally capable of holding onto them longer, and many were able to provide competitive salaries. Today, the minors exist almost solely for the improvement of the major-league game. (A damn shame, IMO)
Asserting that the players in the league as a whole were on average better then than players now relys on the assumption that athletes were somehow inherently better then; and this contradicts all common sense.
Take out Speier, and Hernandez hit more homers than the other 13 shortstops of thirty years ago COMBINED. The goal of the game is not to put the bat on the ball or to make contact, the goal is to score more runs and win games. Players and coaches today believe (and research bears them out) that given the current conditions, swinging for fences every time up is a more efficient strategy.
The point is not that Jose Hernandez was better than Davey Concepcion. (Although Hernandez last year was probably better than Concepcion as a rookie in 72). The point is that it’s comparing apples and oranges. Bill Russell, were he playing today, would probably hit 15-20 homers and struck out more. Jose Hernandez in 1972 would have hit less, choked up and gone for singles. Players adjust to the styles of their era and do what it takes to win, and coaches pick players who match the style as well. In 2002, teams feel they can’t afford offensive zeroes at short. in 1972, teams felt they could play no-hit good-glove guys like Dal Maxvill.
**
Some things can be done with statistics, although as I said, no two measurements are exactly the same. Defense especially is hard to evaluate. But it makes for some great arguments…
By my own admission, in terms of pure athleticism the players now are better than the players of the past. My argument is that there is a wider gap between the top players and the bottom players now, as compared to then. Players who could only bring one tool to the game out of five didn’t make the cut. There really weren’t any specialty players, no closers, no late inning defensive replacements, no caddies for prima donna pitchers.
But you make another interesting point regarding defence. It is particularly hard to gauge and compare, partly because the methods for tracking defensive prowess are far more detailed now, although not much more revealing. Also, the gloves used now are so much better than those used even 30 years ago that fielding percentages now are inconceivably high compared to years gone by. This, I think is a very interesting point of discussion because, unlike pitching and hitting stats, defensive stats rarely tell the whole tale–great defensive players make errors on plays that no other fielder would have gotten to; catchers who can’t block don’t always get called for passed balls; great defensive first basemen save countless errors for the rest of their team by fielding bad throws; most significantly, good fielders are always in position, hit cutoff men, and make good relays, all things for which there are no stats. The raw stats can’t explain why Cal Ripken and Ozzie Smith wound up with virtually identical FPCTs (.979 and .978 respectively) even though most would agree that Ozzie was the better defensive player.
