|
In response to
"TFox: so wait. They weighted the league leaders, or they weighted all home runs?"
by
musubi
|
I'll paste the list and the methodology in the IM. (Warning: its long.)
Posted by
TFox
Jul 6 '10, 08:58
|
The PEET Top 10
The Mag's Player-Era-to-Era Translation leaderboard.
Player Home Runs
Hank Aaron 724
Babe Ruth 663
Barry Bonds 660
Mel Ott 650
Willie Mays 628
Reggie Jackson 595
Frank Robinson 578
Ted Williams 569
Mike Schmidt 557
Harmon Killebrew 552
Total home runs have drifted down from their millennial peak, but what really has decreased is dominance, the ability of top sluggers to blast away from the pack. And when we account for that, we can accurately compare home runs across eras -- and finally put the home run record into proper context.
Let's start with this: Last season, AL hitters smacked 2,560 homers, or 2.92 per 100 plate appearances. That was the sixth-highest frequency in league history, close to the 1999 rate. But Mark Teixeira and Carlos Pena tied for the league lead with just 39 dingers, following a season in which Miguel Cabrera topped the AL with 37 home runs -- the lowest league-leading total since 1989. In the NL, Pujols cracked 47 homers last season, or 6.71 per 100 PAs, to lead the league. That's an impressive rate, but the difference between Pujols' ratio and the NL average (2.49) was just 4.22 jacks per 100 PAs, the smallest disparity since 1992.
You see, every statistic in baseball is compiled within the context of two end points -- the leader and the average. The league average represents the balance between offense and defense in any given year. Replace day games with night games, as baseball did in the 1940s, and league averages for batting and slugging drop. Lower the mound and shrink the strike zone, as the game did in 1969, and offensive numbers jump.
In many categories, the mean has hovered around the same number throughout baseball history, which creates the illusion that it's easy to compare stats across eras. For example, batters have a collective on-base percentage of .329 this season -- the same as in 1977 or 1945. But there's another factor to consider: the league leader, who represents the extent to which any player can exceed his peers. And this gap has changed a lot over time. In 1908, Honus Wagner hit .354, which was 48 percent better than the NL average. In 1925, Rogers Hornsby hit .403, surpassing the league average by 38 percent. In 1991, Terry Pendleton hit .319, just 28 percent better than the league. Yet each player won a batting title in those seasons.
When the talent level improves across baseball, gaps between league leaders and league averages generally shrink. That might seem strange, but evolutionary biologist Stephen Jay Gould used this phenomenon in the mid-'80s to explain why no major leaguer hits .400 anymore. It's also the reason LeBron James can't put up Wilt Chamberlain-like stats in the modern NBA. When everybody gets better, the ability to dominate diminishes. But there have been times when the gap has widened, and league-leading home run hitters have soared away from the average. This happened in the 1920s, when Babe Ruth led baseball out of the dead-ball era and a new crop of power hitters came to dominate the game. And we have just emerged from another such era. In 1998, McGwire hit an eye-popping 7.72 more home runs per 100 plate appearances than the NL average; in 2001, Bonds was 8.02 HRs per 100 PAs better than the league, a measure of dominance that exceeds any posted by Ruth. It is these numbers that have come back to earth over the past couple of seasons. Home run hitting feels more normal now because dominance is returning to historically average levels.
Were steroids behind the widened gaps? Probably. But if we put every player into the context of the league average and leader, we can avoid all the hysteria surrounding performance-enhancing drugs. Let's simply suppose that when Mike Schmidt led the NL with 36 home runs in 1974, his total represented the greatest extent to which any player could have dominated the league that year, and let's say the same thing about Ruth's 60 in 1927 and Bonds' 73 in 2001. For the sake of comparisons, it's the gap between the league and the leaders that really matters, not the possible reasons why the gap changed.
For example, in 1999 Sosa hit 63 home runs in 712 plate appearances, or 8.85 dingers per 100 PAs. The NL average that year was 2.86 per 100 PAs, and the league leader (McGwire) hit 9.83. Thus, Sosa filled 85.8 percent of the gap between the league average and the league leader. For the purposes of this comparison, let's translate Sosa's performance to perfectly average conditions. Since 1920, the start of the live-ball era, league-average hitters have averaged 1.94 homers per 100 plate appearances, while league leaders have averaged 6.60. That gives us a difference of 4.66.
To calculate Sosa's adjusted total, we start with the historical average rate: 1.94 home runs per 100 PAs, which comes out to 13.8 homers in 712 PAs. To that figure, we add 85.8 percent of the difference between the league leader and the league average. In this case, 85.8 percent of 4.66 is 3.99. So, at a rate of 3.99 homers per 100 plate appearances, Sosa would hit 28.5 jacks in 712 PAs. Now add 28.5 to 13.8 and Sosa has an adjusted total of 42.3 round-trippers, an impressive but not historic number.
We don't have to argue about Sosa's training methods, or about expansion or population growth or whether pitching has improved or declined since the good old days. With this method, which we'll call The Mag's Player Era-to-Era Translation system (PEET), we can simply note that about a third of Sosa's home runs were due to the context of his competition -- how often batters homered in 1999 and the degree to which hitters could stand apart from the league.
When we run the leading home run hitters of all time through the PEET system, it becomes clear which sluggers' numbers were inflated or obscured by varying dominance. Translated to an average environment, Ken Griffey Jr., Sosa, Alex Rodriguez and McGwire all drop off the all-time top 10 list, while Mel Ott, Reggie Jackson, Schmidt and Ted Williams move onto it (see the chart at the bottom). These graphs here, which compare how Schmidt and Rafael Palmeiro (one of the biggest losers in our calculations) performed relative to the average and the leader in their careers, should help illustrate how PEET works. Palmeiro, who at times early in his career hit fewer home runs than the average, loses 90 home runs via PEET, and Schmidt, who was consistently close to the league leader, gains nine. And while many modern fans might not know much about Ott, his long ball prowess should not be forgotten. The New York Giants great of the 1930s gains 139 home runs with PEET, the most of any player. Ott finished in the top 10 in his league in home runs 18 times, tied with Aaron and Ruth for the most ever.
It's hard to match The Bambino, who filled a whopping 97 percent of the gap between the league average and the league leaders during the live-ball portion of his career. Aaron wasn't nearly as dominant, filling 70 percent of the gap during his career. But just like in real life, his consistency as a power hitter (not to mention the fact that he had 3,323 more plate appearances than Ruth) puts him at the top of the PEET list. And unlike real life, Aaron holds off Bonds (71 percent of the gap) on our list. According to our calculations, Bonds would have been hard-pressed to hit 700 homers had he played in Aaron's era, while if Aaron had played in the '90s and early 2000s, he probably would have cleared 800. Bonds had all-time great power, but his status as the all-time home run king is a figment of context.
Then again, so is every statistic -- until you look at it the right way.
|
Responses:
|
Replies are disabled on threads older than 7 days.
|
|