Philosophy of batting leadoff
I was skimming through the Bill James Historical Baseball Abstract (1985 edition) the other day and came across an interesting essay on leadoff hitters.
In the player comments sections, James compares Hall of Fame shortshops Luke Appling and Luis Aparicio. James writes at some length about their different styles of offensive play: Appling was a low-power, good contact hitter who walked a lot; Aparicio hit for a lower average and walked less, but he was much faster and stole many more bases. Here are the slash stats for the two:
Appling: .310/.399/.389, normalized: .301/.389/.386
Aparicio: .262/.311/.343, normalized: .274/.324/.358
With Aparicio swiping 506 bags (and leading the league nine times) and Appling stealing only 186. Part of the difference in their raw stats is due to context, the eras and ballparks in which they played. Appling played in the high-average 1930s, while the core of Aparicio’s career falls in the low-scoring ’60s. The normalized numbers above puts them on an even footing.
James on Appling:
Luke Appling represented a type of offensive player that was common at one time and has all but disappeared today, a small man who didn’t often hit the ball hard, but made an offensive contribution by slap-hitting .300 — actually, he once hit .388, but that was kind of flukey — and working the pitcher for 85 to 120 walks a season.
On the disappearance of these hitters:
This type died out abruptly after 1950, with the exception of an occasional Matty Alou. It’s hard to say why. A number of teams pulled their fences in about then, which probably reduced the distance between the infeld and the outfield, and took away many of the dying quail hits that were a staple of the profession. The similar hitters who did exist after that rarely played a key defensive position.
And the emergence of the Aparicio type:
After them came the speed merchants, the little leadoff men who did not hit for quite the same avarage, walked less than half as much, and tried to make their offensive contribution by stealing bases, getting into scoring position. Louis Aparicio was the prootype of such a player.
James then points out that the stolen bases of the Aparacio types (which include Maury Wills and Bert Campaneris, among others) do not nearly make up for the lower OBPs these players produced. He doesn’t go through the details, but does cite the relative low runs-scored totals of these three players as evidence: none of them ever led the league in runs scored and they have only two 100-plus runs scored seasons between them (both owned by Wills).*
*Here’s something that James didn’t include in his essay, though: Appling never led his league in runs scored, either. And he only scored more than 100 runs in a season once. Anyway, moving on…
James then speculates on why the inferior type of leadoff hitter has taken root in the game (sorry for the long direct quotes, but the alternative is that I should come up with enough original content to fill my word quota. Besides, wouldn’t you rather read James than me? Yes, I thought so). Anyway, here’s the quote:
For a decade or more, baseball was held in the grip of the Aparicio/Wills generation of leadoff men. If it seems strange that all of baseball could have fallen victim to the delusion that 40 or 50 stolen bases could compensate for a comparative inability to reach base, I would ask you to consider that all of the institutions of our society — our education systems, our criminal justice system, our entertainment industry, our military — get entangled in strange, illogical, self-defeating practices and habits, and for a decade or more not a single state nor district nor individual will find the courage, wisdom and perspective to escape; indeed, whole countries go quite mad from time to time. Is it stranger that American baseball men became convinced that they had to use a type of player that analysis or simple reasoning can quickly show is not productive?
And finally, how he sees that changing:
Delusions pass, and men do come to their senses. There is no doubt in my mind that by the year 2000, the Jerry Remy-, Damaso Garcia-type of leadoff hitter will be extinct, and of the two types of leadoff men — Appling and Aparicio — will develop a synthesis, a player who combines the speed and base stealing ability of Aparacio with the pesky patience of a Luke Appling. Rickey Henderson and Tim Raines are the harbingers of this new type.
So, I thought I’d have a look and see if James’ prediction about leadoff hitters in 2000 was accurate or not. Here’s what I did: I made a list of all players who got at least 400 plate appearances in the leadoff slot in any given year, going back to 1952 (the Retrosheet era). Then I calculated the on-base percentage for all such players in a given year, to see if there is any historical trend over the last 50-60 years. The plot below shows my findings: The red circles show the average OBP for leadoff hitters in any given year. The red curve shows the same thing, with some of the fluctuations smoothed out.
The blue curve shows the average MLB OBP for the same years. Note that I made the same 400 PA requirement for players going into the MLB average (because players with 400 PA as a group are better hitters than the full complement of MLB players — I want to compare apples to apples). I decided to put the MLB-average curve on the plot after seeing the ups-and-downs of the leadoff hitters’ OBP over the years. I suspected that the large variations were tied to general changes in MLB and not so much a changing philosophy of what a leadoff hitter should be.
What about James’ prediction that by 2000 leadoff hitters would be a synthesis of the Appling and Aparacio types, i.e. fast runners and patient hitters? Actually, for about 10 years following the publication of James’ book (1985-1995), leadoff OBP rose sharply, from about .350 to almost .365. That increase was fueled not only by Henderson (.409) and Raines (.382), as James’ predicted, but also by Wade Boggs (.428), Tony Phillips (.391), Kenny Lofton (.383) and Brett Butler (.382). However, looking past the mid-’90s, I get the feeling that this was more a question of happenstance rather than a philosophical change in the way managers view the leadoff spot.
In 2000, when James thought that the leadoff position would have been transformed to speedy, patient guys like Rickey Henderson, we had the following leadoff men (min. 400 PA):
+-------------------+------+-------+-------+-------+------+ | Name | PA | avg | obp | slg | SB | +-------------------+------+-------+-------+-------+------+ | Castillo, Luis | 625 | 0.334 | 0.412 | 0.388 | 62 | | Erstad, Darin | 746 | 0.355 | 0.408 | 0.541 | 28 | | Furcal, Rafael | 402 | 0.295 | 0.387 | 0.382 | 40 | | Vina, Fernando | 549 | 0.300 | 0.379 | 0.398 | 10 | | Damon, Johnny | 740 | 0.327 | 0.378 | 0.495 | 46 | | Anderson, Brady | 613 | 0.257 | 0.373 | 0.422 | 16 | | Henderson, Rickey | 519 | 0.233 | 0.366 | 0.305 | 36 | | Lofton, Kenny | 585 | 0.278 | 0.366 | 0.422 | 30 | | Knoblauch, Chuck | 456 | 0.283 | 0.365 | 0.385 | 15 | | Young, Eric | 686 | 0.297 | 0.364 | 0.399 | 54 | | Stewart, Shannon | 631 | 0.319 | 0.363 | 0.518 | 20 | | Alicea, Luis | 506 | 0.294 | 0.360 | 0.404 | 1 | | Durham, Ray | 705 | 0.280 | 0.358 | 0.450 | 25 | | Offerman, Jose | 410 | 0.255 | 0.353 | 0.359 | 0 | | Belliard, Ron | 429 | 0.263 | 0.352 | 0.389 | 7 | | Owens, Eric | 428 | 0.293 | 0.346 | 0.381 | 29 | | Goodwin, Tom | 582 | 0.263 | 0.343 | 0.352 | 55 | | Benard, Marvin | 600 | 0.263 | 0.341 | 0.396 | 22 | | Long, Terrence | 561 | 0.288 | 0.336 | 0.452 | 5 | | Reese, Pokey | 426 | 0.255 | 0.317 | 0.386 | 29 | | Bergeron, Peter | 489 | 0.245 | 0.313 | 0.349 | 11 | | Williams, Gerald | 616 | 0.274 | 0.308 | 0.427 | 12 | | Womack, Tony | 658 | 0.271 | 0.307 | 0.384 | 45 | | Glanville, Doug | 534 | 0.275 | 0.302 | 0.374 | 31 | +-------------------+------+-------+-------+-------+------+
Well, at least Rickey Henderson is like Rickey Henderson.
Seriously, I’m not sure quite what to make of this list. It doesn’t look dominated by this Appling/Aparacio hybrid, does it? The MLB average of OBP (taking only players with 400-plus PA), was around .360 in 2000, and I count exactly half of our leadoff hitters above average. And a number of them are way down at around .300 in OBP. For each Rafael Furcal, we have a Gerald Williams. Fernando Vina is offset by Tony Womack. Darin Erstad (in his career year) is bookended by Doug Glanville. No, I don’t see Bill James’ utopian leadoff hitter taking over by the year 2000.
Another decade has gone by since 2000, though, has the Bill James Leadoff Era taken hold in that time? Referring to the graphic above, we see that over the last 10 years, leadoff OBP has seen a downward swing, followed by an upward trend more recently, while average OBP seems to be dropping over the same period. It doesn’t look, though, that managers/GMs have yet seen the light on how optimum deployment of leadoff hitters. In case you’re curious, here is the above table repeated for 2009:
+--------------------+------+-------+-------+-------+------+ | Name | PA | avg | obp | slg | SB | +--------------------+------+-------+-------+-------+------+ | Jeter, Derek | 693 | 0.334 | 0.404 | 0.465 | 30 | | Figgins, Chone | 729 | 0.298 | 0.391 | 0.393 | 42 | | Suzuki, Ichiro | 678 | 0.352 | 0.385 | 0.465 | 26 | | Span, Denard | 676 | 0.311 | 0.385 | 0.415 | 23 | | Lopez, Felipe | 661 | 0.310 | 0.382 | 0.427 | 6 | | Scutaro, Marco | 680 | 0.282 | 0.376 | 0.409 | 14 | | Schumaker, Skip | 530 | 0.303 | 0.363 | 0.393 | 2 | | Morgan, Nyjer | 429 | 0.307 | 0.362 | 0.388 | 42 | | Roberts, Brian | 692 | 0.283 | 0.356 | 0.451 | 30 | | Fowler, Dexter | 440 | 0.266 | 0.353 | 0.406 | 27 | | Bourn, Michael | 552 | 0.285 | 0.351 | 0.384 | 61 | | Ellsbury, Jacoby | 553 | 0.301 | 0.351 | 0.415 | 70 | | Podsednik, Scott | 553 | 0.304 | 0.349 | 0.412 | 30 | | Kennedy, Adam | 441 | 0.289 | 0.344 | 0.410 | 20 | | DeJesus, David | 426 | 0.281 | 0.344 | 0.434 | 4 | | Furcal, Rafael | 503 | 0.269 | 0.334 | 0.375 | 12 | | Granderson, Curtis | 601 | 0.249 | 0.325 | 0.453 | 20 | | Kinsler, Ian | 505 | 0.253 | 0.325 | 0.488 | 31 | | Upton, B.J. | 457 | 0.241 | 0.312 | 0.373 | 42 | | Rollins, Jimmy | 691 | 0.250 | 0.295 | 0.423 | 31 | +--------------------+------+-------+-------+-------+------+
It’s an interesting piece of work. I think the one point to add is that there’s a difference between what a manager is trying to accomplish, and what the player on the field is doing. For example, if the manager knew that Furcal and BJ Upton would have terrible OBP’s for the year, he wouldn’t have batted them first. Their unexpectedly poor performance doesn’t undermine Bill James’ point. On the other hand, Podsednik and Jimmy Rollins are very close to the Aparicio model (though Rollins peaked as a much, much better hitter), so it’s not like every manager focused on OBP in the leadoff spot.
Kevin’s point is a good one. Also, the Red Sox last year and the Yankees in most previous years hint at a different stance on this question evolving. Boston has a near hybrid type in Pedoria (more to the Appling end) and yet they bat him second, going with lower OBP, higher SB Ellsbury in the leadoff. Previous to last season, Jeter usually batted second for the Yankees with some good hybrids taking the top spot (Damon) and some bad ones (Womack). Taking the chance on lower OBP, higher steal guy up top gives your best contact/OBP combo a chance to drive in a run if your Ellsbury type does take second, and your #3 hitter still has a quality OBP ahead of him in the lineup.
I think this might be the more accurate understanding of the logical evolution going on. “Traditional” #2 hitters were very often slap hitters with special bunting/h&r skills who really shouldn’t have been batting so high up in the order.
Looking at the 2000 list, there are a few guys toward the bottom of the list who are more like Appling or a hybrid of the 2. Belliard, Offerman, Bernard, Bergeron and Goodwin all had fairly high OBP when you look at how low their BA were, and only Goodwin stole bases at an above avergare clip. i’m not sure if just looking at the difference between BA and OBP will show any type of trend, but might shed some light.
Maybe someday the Phillies will figure out leadoff men…
On the graph… it almost looks like the swells & valleys follow the careers & retirements of particular leadoff men. If you were to do this again, but remove the top 5 OBP & lowest 5 OBP (the extremes every year) from the equation, how would the curve move then? Doing that could show even moreso whether there’s been any philosophical changes/trends or whether there’s simply been better leadoff talent available during those periods (not every generation can have their Henderon, Rainers, or Boggs).
With leadoff hitters you have to consider the team’s overall talent. Bad hitting teams lack good hitters in general so they typically do not have the luxury of a good leadoff hitter. If a really bad hitting team had someone like Tim Raines, they may bat him third. So a parallel analysis would be to look at the leadoff hitter compared to the rest of his team and see if the manager, given his situation, is picking the best available leadoff hitter or if he is just drooling over stolen bases.
And if you really wanted to get deep, you look at the player development and see if teams are favoring certain skills in the minors. See how they treat guys in AA with high OBP but no speed compared to low OBP/high speed options. That could indicate the type of player they want. However, I sense that analysis would be very complex at best.
Thanks for the article!
Baseball-Reference.com lists the AL split for the first batter as having a .355 OBP and that number as .340 in the NL for 2009. That might make this a league-dependent issue, which would be interesting to look at.
Also, you might want to clarify which columns in your table are leadoff stats only and which are full-year stats. Jacoby Ellsbury, for example, had 551 PA and a .347 OBP in the leadoff role and only stole 60 bases while in that capacity. Players could run differently based on where they hit in the lineup.
Wow, lot’s of good comments. A few of my own:
kevin/ Yes, agreed. Of course, some guys do better than expected, so to some degree it cancels outs. The graphic is supposed to show the general tendencies.
cards.devonyoung.com/ An interesting suggestion, I may look into that. However, the general trends of leadoff OBP do follow pretty closely with league average (and I bet SB’s do, too, though I didn’t check).
Paul/ Very good points. (In fact, Raines did get more than 1500 PA’s batting third.) A deeper analysis should take team context into account.
Red Sox Talk/ You’re perfectly correct: the slash stats in the tables are only while the player was leading off, while the SB’s are the player’s total for the season. Good catch.
I disagree with John’s conclusion here. It seems to me the data actually show James’ prediction of better use of the leadoff slot to have been largely correct. If you look at OBP from the leadoff slot (not just those with over 400 PA hitting leadoff), they had a .347 OBP last year vs. league average of .333. That seems about right to me. If you’re basically picking the player on your team with the highest OBP but who doesn’t hit for power, you will often end up with someone just a bit above average on OBP. Even looking at 2003, the apparent recent low point on John’s chart, leadoff hitters overall had a league-average OBP which isn’t that bad. Even with a smart allocation of talent, there will probably be occasional seasons in which leadoff guys are only average. If you look at the last table, there are at most a handful of fast players hitting leadoff today despite a poor OBP —that’s huge progress from the 1960s/70s.
If there’s a slot that is still misused by managers, it’s the #2 spot much more than leadoff. Teams still regularly bat weak hitters second, even though it’s where your best hitter should be (according to The Book analysis). #2 hitters are actually weaker than #6 hitters on average, which is clearly a mistake.
[cross-posted at The Book Blog]
Damn, Paul made the point I wanted to make. I’d just like to add that even if there are more managers today who recognize the value of OBP leading off (probably true in my opinion), they are likely being counterbalanced by two things: one, a lack of pure “OBP and nothing else” hitters to BAT leadoff, and two, a much more powerful tradition of batting your best hitters in the middle of the lineup. Albert Pujols may have the best OBP on your team, but no manager today (and few fans) would bat him at leadoff. And I think once you get down into the realm of unremarkable (<.340) OBPs, managers then tend to favor speed, even if it costs them 20 or 25 points of OBP.
I very much agree with Guy’s idea that the second batter should be stronger than the sixth batter. Here in Washington, the manager plans on batting Cristian Guzman second while recently acquired Adam Kennedy might bat as low as eight. However last year Guzman’s BA was .284, his OBP was .306, he hit 6 home runs, knocked in 52 and was 4-9 in steals. A caveat is that Guzman had the good fortune of batting in front of Zimmerman and Dunn. On the other hand, Kennedy hit .289, his OBP was .348, he hit 11 home runs, knocked in 63 and was 20-26 in steals. Kennedy should bat second but Guzman sulks when he doesn’t bat up in the line-up. Fortunately, his contract is up at the end of the year.
Appling never scored all that many runs because he wasn’t used as a top of the lineup hitter. Retrosheet is missing a few seasons, but here are Appling’s lineup slots:
5th 876 games
6th 331 games
2nd 245 games
4th 207 games
3rd 144 games
7th 82 games
9th 39 games (entered game as pinch-hitter)
8th 5 games (probably all as pinch-hitter)
1st 4 games
Players who could both hit .300 and draw around 100 walks without slugging power were always extremely rare. The players who drew the most walks and batted leadoff rarely hit .300, and the guys who could hit .300 regularly were often put in the middle of the lineup. Johnny Pesky was a #2 hitter, Ferris Fain hit 4th or 2nd, Arky Vaughan usually batted 4th. Of the Appling types I could only find Stan Hack used as a leadoff hitter. The Stanky/Bishop/Yost leadoff hitters batted around .250-.270 with 120 walks a year.
John Walsh-
And Dillinger never walked more than 65 times, so he doesn’t belong in the group. I forgot about Ashburn and The Little Professor, but they fit. The Appling family of hitters is pretty small, and probably owes its origin to the deadball era when teams scratched out runs without benefit of the longball. 100 walks was rare in the deadball era, and back then 80 might lead the league. But Appling had a lot in common with hitters such as George Burns and Harry Hooper.
Another factor to consider is how the unequal distribution of high-OBP players between teams relates to the league average. For instance, there are six(!) Yankees who would have had above-average OBPs as leadoff hitters last year: Johnson (.426), Jeter (.406), ARod (.402), Texeira (.383), Swisher (.371) and Posada (.363). Obviously, only one of them can actually bat leadoff. On the other hand, on a bad offensive team a low-OBP guy may bat leadoff, not because his manager thinks he’s well suited to the role, but simply for a lack of better options.
It seems like this inequality would depress the gap between the league’s OBP and leadoff hitters’ OBP, making you underestimate how much managers appreciate its importance. A comparison of each leadoff hitter’s OBP to that of his team might shed som light on this.
In 1986 Vincee Coleman led the NL’s lead-off men in runs scored, despite a .301 obp. Speed tends to be undervalued by the saber-set.
stevebogus/ D’oh! I can’t believe Bill James wrote a 1,000 word essay on Appling as a leadoff hitter, when in fact he didn’t bat leadoff! And then I did the same thing!
I wonder if James confused Appling with some other Appling-type hitter who did bat leadoff? He mentions several others in his essay: Ashburn, Pesky, D. DiMaggio, Fain, Dillinger, Stan Hack, Harry Walker, Elmer Valo. Only Ashburn, Dom, Dillinger, and Hack were primarily leadoff men.
zubin/ Vince Coleman in 1986 stole 107 bases and was caught only 14 times. It’s one of the best SB performances in history. I don’t know who is in your “saber-set”, but if you took a poll of the saber-types that I know, I think 100% would welcome that kind of SB performance.
Nevertheless, Coleman, despite coming to the plate 664 times that year, all in the leadoff position , he did not score 100 runs, nor did he lead the league overall. He did lead NL leadoff hitters, that’s true. But, only 4 NL players had at least 400 PAs in the leadoff spot and Coleman had the most PAs, by far.
The next year, 1987, Coleman stole a similar number of bases (109 SB, 22 CS), but upped his OBP to .363. He scored 121 runs (27 more than in 1986) and the Cardinals won the pennant. He still did not lead the league in runs, by the way. Von Hayes and Tony Gwynn, with higher OBPs and lower SB totals, scored more often than Coleman.
Coleman was a fascinating player: in the 6-year period from 1985 to 1990, Coleman by far led the majors in stolen bases, he had over a 100 more than the #2 guy, some fella named Rickey Henderson. He trailed Henderson in runs scored by almost 150, though, in that period. He also scored 85 runs fewer than Wade Boggs, despite stealing, let’s see here, 542 more bases than Boggs.
One of the greatest leadoff batters was Eddie “The Walking Man” Yost, who played most of his career with the hapless Washington Senators. Yost played 18 seasons, two of which were brief call-ups from the minors and his last two years were spent largely on the bench. In his 14 years as a starter, Yost had 1528 walks, an average of 109 per season. He led the American League in walks six times (remember that he played during the Ted Williams era) and had over 100 walks in a season eight times. His career On Base Percentage was .394. However, Yost only stole 72 bases in 18 years.
A few more words about “Ol Aches and Pain” Luke Appling: He played through injuries. That fact, in and of itself, would have added an intangible and benefit most teams.
A second notable trait was that Appling could hit foul ball after foul ball, thereby tiring the opposing pitcher. that characteristic would have also been beneficial to any baseball team.
Of course neither of the above two characteristics can be easily melded into a statistical study and this is where Sabremetrics always falls short.
John, the point is that when a high obp guy isn’t available, an extreme sb threat isn’t a bad substitute. Baseball managers aren’t stupid—they know what obp is and how valuable it is in the lead-off spot.
Actually, speed is more valuable at the bottom of the lineup than the top. Consider the following:
What value does a stolen base have at the top of the Cardinals lineup? If Albert Pujols gets a hit, it will likely be for extra bases so the runner would likely score from either 1st or 2nd. However, the loss of a baserunner on a caught stealing would be very deleterious.
Now, consider the case of the guy batting 6th in the lineup. If he gets on base, he’ll likely only get a single. The 7th or 8th place hitters would similarly only get singles. Now a SB means that you only need one extra hit to score a run instead of two.
If a team has someone with a low OBP at the top of the lineup, you’re providing fewer opportunities for your high SLG guys.