Finding the Ideal Leadoff Hitter

We know, in 2014, that lineup construction has little effect on winning. And yet, it’s not any less frustrating when managers set their batting orders in ways that seem to defy any semblance of logic. Lineup construction matters to us. We may know it’s not terribly important, but we’re fascinated in spite of ourselves.

The lineup position subject to the most debate is probably leadoff. Multiple writers and analysts have noted that players who would make the best leadoff hitters are normally too valuable to use in the leadoff position. Bill James wrote in his New Historical Abstract, “All of the greatest leadoff men … would be guys who aren’t leadoff men, starting with Ted Williams … if you had two Ted Williamses, and could afford to use one of them as a leadoff man, he would be the greatest leadoff man who ever lived.”

Every method I’ve seen to determine great leadoff batters produces names like Ted Williams, Barry Bonds, Mickey Mantle, Ty Cobb … players who are probably better suited to the second through fourth spots in the batting order. I think I’ve found a simple method that solves the problem. I’ve always been interested in singles hitters who walk. It’s a skill set that matches our image of the prototypical leadoff batter.

Most fans agree that a good leadoff man should get on base and run the bases well. Most fans further agree that a player who both gets on base and hits with power is more valuable a little later in the order, where he can drive in runs. If we accept that we probably can’t have two Ted Williamses, a realistic ideal of the leadoff batter has a high on-base percentage but doesn’t hit with a lot of power.

With this in mind, I’m adapting a stat I’ve talked about elsewhere to identify optimal leadoff men: OBP minus ISO. In my head, I’ve always called this reverse ISO, but that’s sort of a misnomer, and it’s a little unwieldy, so from here on let’s call this stat combination Leadoff Rating, or LOR. We know a good leadoff man gets on base, but most players with high on-base percentage are great all-around hitters. We know power hitters are usually better suited to other spots in the batting order, but many players with low ISO just aren’t that great. By subtracting isolated power from OBP, we can identify players specially suited to hitting leadoff.

This stat does not include baserunning (because I have no idea how to incorporate it with two percentages) but it turns out not to matter very much. A significant majority of players who rank well in LOR were also accomplished baserunners, and base stealers in particular. Among the top 300 hitters of all time (basically everyone with 2,000 career hits), I found a fairly strong positive correlation between LOR and SB (r=.465). The relationship is weaker if you only look at 1947-present (r=.356), but a degree of positive correlation is clear. In both data sets, n=300.

When you calculate LOR for the all-time top 300 hitters, the leader is Billy Hamilton. That’s Sliding Billy Hamilton, the Hall of Fame outfielder for Philadelphia and Boston in the 1890s, not the rookie phenom for the Cincinnati Reds. The original Hamilton retired with 1,782 singles, 1,187 bases on balls, and 376 extra-base hits. He hit .344/.455/.432, with an ISO of just .088, and an OBP higher than his slugging percentage. Hamilton also stole 912 bases. He is a superb example of the hitter we’re looking for, and he leads the new stat by a huge margin. His .367 LOR rates 12% higher than second-place Eddie Collins (.328). Here’s the top 75:

1    Billy Hamilton, .367
2    Eddie Collins, .328
3    Richie Ashburn, .322
4    Willie Keeler, .315
5    Luke Appling, .311
6    Jesse Burkett, .307
7    Fred Tenney, .306
t8    Wade Boggs, .300
t8    Dummy Hoy, .300
10    Stan Hack, .298
11    Willie Randolph, .297
12    Charlie Jamieson, .296
13    Patsy Donovan, .293
14    Rod Carew, .292
15    Brett Butler, .291
16    Joe Sewell, .290
17    Ty Cobb, .287
18    Buddy Myer, .286
19    Clyde Milan, .285
20    George Van Haltren, .284
21    Cap Anson, .281
22    Maury Wills, .280
23    Juan Pierre, .277
t24    Dave Bancroft, .276
t24    Jason Kendall, .276
t24    Lloyd Waner, .276
27    Nellie Fox, .274
28    Tris Speaker, .272
t29    Jimmy Sheckard, .271
t29    Ozzie Smith, .271
t31    Stuffy McInnis, .270
t31    Sam Rice, .270
t31    Arky Vaughan, .270
t34    George Burns, .269
t34    Fred Clarke, .269
t34    Pete Rose, .269
t37    Tony Gwynn, .268
t37    Joe Kelley, .268
39    Deacon White, .266
40    Ichiro Suzuki, .265
t41    Billy Herman, .264
t41    Paul Waner, .264
43    Jake Daubert, .263
44    Harry Hooper, .262
t45    Doc Cramer, .261
t45    Rickey Henderson, .261
t47    Max Carey, .260
t47    George Kell, .260
49    Hugh Duffy, .259
t50    Pee Wee Reese, .258
t50    Steve Sax, .258
t52    Joe Judge, .256
t52    Omar Vizquel, .256
t54    Bid McPhee, .254
t54    Tim Raines, .254
t56    Frankie Frisch, .253
t56    Jack Glasscock, .253
t58    Ed Delahanty, .252
t58    Harvey Kuenn, .252
t58    Nap Lajoie, .252
t58    Honus Wagner, .252
t58    Dixie Walker, .252
t63    George Davis, .251
t63    Tony Phillips, .251
t63    George Sisler, .251
66    Ed McKean, .250
t67    Phil Cavarretta, .249
t67    Dick Groat, .249
t69    Bill Dahlen, .248
t69    Kenny Lofton, .248
t69    John Ward, .248
t72    Dick Bartell, .247
t72    Derek Jeter, .247
t72    Pie Traynor, .247
75    4 tied, .246

 

That’s not an altogether surprising list — which means the new stat, simple as it is, works. The top 10, as a group, hit .319/.405/.408. They combined for 26,000 hits, of which 21,000 were singles. They walked 11,000 times and struck out just 3,300, with 4,000 stolen bases. They scored 15,000 runs and drove in just 8,500. But five of those 10 were active in the 1890s, and only two (Richie Ashburn and Wade Boggs) played the majority of their careers after World War II. Players like Billy Hamilton and Fred Tenney certainly fit our leadoff profile, but we’ve isolated an era and a style of play as much as a style of player. The top 20 in LOR, just from the last 70 years, paints a truer picture of the players this stat helps us identify. Min. 2000 hits:

1. Richie Ashburn, .322
2. Wade Boggs, .300
3. Willie Randolph, .297
4. Rod Carew, .292
5. Brett Butler, .291
6. Maury Wills, .280
7. Juan Pierre, .277
8. Jason Kendall, .276
9. Nellie Fox, .274
10. Ozzie Smith, .271
11. Pete Rose, .269
12. Tony Gwynn,.268
13. Ichiro Suzuki, .265
14. Rickey Henderson, .261
15. George Kell, .260
16. Pee Wee Reese, .258
17. Omar Vizquel, .256
18. Tim Raines, .254
19. Harvey Kuenn, .252
20. Tony Phillips, .251

(Kenny Lofton and Derek Jeter rate 22nd and 23rd)

Most of those players are recent enough, and good enough, that you have an immediate image of how they played. All fit our image of leadoff hitters, and most of them did in fact lead off. Henderson and Raines rate lower than we might expect because LOR doesn’t include baserunning, but also because of their relatively strong ISO. I might even argue that this is a strength of the system: it recognizes that Henderson and Raines were such good hitters, they’d be effective in the middle of the lineup — or at least more so than players like Willie Randolph and Juan Pierre. Having said that, LOR identifies Henderson and Raines as possessing top leadoff skills over the last seven decades. I played around with a different formula, (2 * OBP) – ISO, to boost players like Rickey, and it worked, but it also moved the Cobbs and Ted Williamses too high on the list. LOR as I outlined it earlier identifies great leadoff hitters without just finding great hitters regardless of skill set. If you had a team with both Richie Ashburn and Rickey Henderson, you might very well bat Ashburn first and Rickey second.

The potential flaw in all this is that we’re looking at great players. Does the formula still work when we apply it in less exceptional situations? To find out, I ran LOR on a sample of seven 2014 MLB teams. The sample is effectively random: the seven teams named after animals. That covers five of the six divisions (sorry, AL West), with first-place teams (Blue Jays, Tigers), last-place teams (Cubs, D-backs), and those in between (Cardinals, Marlins, Orioles). All stats through June 30, minimum 150 PA.

Detroit Tigers (44-34)

LOR Leader: Alex Avila, .202

Avila is hitting .233/.350/.381. He has the third-highest OBP on the team, trailing only Victor Martinez and Miguel Cabrera. Avila has the worst BsR (-4.4) on the team, and the Tigers have a stacked lineup, but it seems silly to bat him seventh or the eighth in the order, which is where he’s been hitting. Austin Jackson ranks second on the Tigers in LOR, .195.

Actual Leadoff Hitters: Ian Kinsler, .161, and Rajai Davis, .187

Kinsler, who is batting .307/.342/.488, might make more sense than Davis (.273/.319/.405 with 21 SB) just because Detroit’s lineup is so loaded. Five Tigers have at least 150 plate appearances and an OBP over .340.

Historical comparison: George Brett retired with a .189 LOR. Dave Winfield was at .161, and Mickey Mantle .162.

Toronto Blue Jays (45-39)

LOR Leader: Adam Lind, .231

Anthony Gose only has 123 plate appearances, but his batting line is .243/.350/.301, giving him an astronomical .292 LOR. Gose has a good OBP and he’s hitting with absolutely no power. He’s only got the fourth-best OBP among the 10 Jays with 100 PA, but I don’t think Toronto fans want to see Jose Bautista and Edwin Encarnacion leading off. They might be okay with Lind, who is batting .331/.406/.506.

Actual Leadoff Hitter: Jose Reyes, .175

Reyes has 16 SB and just 1 CS, but his .321 OBP is among the lowest on the team, and his .146 ISO would work anywhere in the lineup. He’s not a terrible leadoff hitter, but Lind (who’s mostly hit fourth or fifth) or Gose (ninth) would probably make more sense at the top of the order.

Historical comparison: Chipper Jones had a .175 LOR. Like Reyes, he was probably a better fit later in the lineup, though for different reasons.

St. Louis Cardinals (44-39)

LOR Leader: Jon Jay, .271

Jay is batting .296/.356/.381 this season. He has the third-highest OBP on the team (Holliday, Carpenter), with a .085 ISO. The Cardinals are stacked with leadoffable hitters, though. Matt Carpenter (.269), Matt Holliday (.249), Mark Ellis (.231), Kolten Wong (.206), and Yadier Molina (.204) all have strong LORs. Their team ISO of .117 is second-lowest in the majors (Royals, .111).

Actual Leadoff Hitter: Matt Carpenter, .269

There’s effectively no difference between Jay’s .273 and Carpenter’s .269. Carpenter has a .373 OBP and only 4 HR.

Historical comparison: Carpenter’s LOR this season is the same as Pete Rose’s career mark. It’s a great number for a leadoff man, and in Carpenter’s case it’s supported by the best OBP on the team.

Baltimore Orioles (42-39)

LOR Leader: Nick Markakis, .246

Markakis is batting .298/.358/.410. Those are pretty prototypical leadoff numbers. J.J. Hardy (.243) ranks second. Nelson Cruz (.060) ranks last.

Actual Leadoff Hitter: Markakis

The Orioles have a very sensibly constructed lineup. Markakis, with a good OBP and low ISO, leads off. Steve Pearce, who leads the team in wRC+ (160), hits second. Then Adam Jones, with Chris Davis and Cruz alternating in the 4/5 spots. Hardy bats sixth, Manny Machado has been dropped to seventh, then Jonathan Schoop, followed by the non-Wieters catchers.

Historical comparison: Julio Franco had a .246 LOR. Derek Jeter is at .247.

Miami Marlins (39-43)

LOR Leader: Casey McGehee, .300

This is an extraordinary LOR, and it is crazy that McGehee is batting fourth. You’ve probably read the Bill James bit about Mark Grace and Ryne Sandberg, and how teams think in images. McGehee is hitting .309/.376/.385, numbers that scream leadoff, but he’s a third baseman, so he’s hitting cleanup. McGehee ranks third in the majors in WPA, but his .385 SLG seems like a better fit at the top of the order. Adeiny Hechavarria ranks second on the Marlins in LOR (.239), but mostly because of his .060 ISO.

Actual Leadoff Hitter: Christian Yelich, .175

Yelich is batting .259/.340/.424, which is perfectly adequate, but it would make a lot more sense to bat McGehee first and put Yelich in the middle of the order than the other way around. Yelich does have 10 steals with just one CS.

Historical comparison: Chipper Jones again.

Chicago Cubs (34-46)

LOR Leader: Emilio Bonifacio, .228

Bonifacio is Chicago’s actual leadoff hitter! He’s hitting .261/.307/.340, and while that .307 OBP probably does not scream “effective leadoff man” to you, it’s actually 4th (out of nine players with 150 PA) on the Cubs. A weakness of LOR is that it loves low ISO, so a player with minimal power will rate reasonably well if he’s any kind of major league talent. Luis Valbuena (.361 OBP, .185 LOR) might actually be the best leadoff option on the team.

Actual Leadoff Hitter: Bonifacio

Bonifacio leads the team with 13 SB, but he’s also been caught stealing 6 times.

Historical comparison: John Olerud retired with a .228 LOR. Bonifacio is not as good as John Olerud.

Arizona Diamondbacks (35-49)

LOR Leader: Martin Prado, .219

Prado is batting .268/.315/.363, so his good LOR is more about a low ISO than a strong OBP. Miguel Montero (.346 OBP, .181 LOR) might be a more appropriate choice to lead off.

Actual Leadoff Hitter: Gerardo Parra, .201

Parra is hitting .255/.306/.360, but Kirk Gibson has shuffled the lineup a lot this season, and recently Ender Inciarte is the leadoff man. Inciarte actually has the best LOR on the team (.244), but in only 124 plate appearances. Inciarte has a mind-blowing .026 ISO; in 117 AB, his only extra-base hits are three doubles. He’s stolen some bases (6 SB, 1 CS), and it probably wasn’t the .270 OBP that inspired Gibson to move him up from 8th in the order. Inciarte’s .239/.270/.265 batting line suggests that he should be hitting in Reno, not hitting leadoff.

Historical comparison: Roberto Clemente, Kirby Puckett, and Bernie Williams all had .201 LOR.


Writing the section above, I was reminded numerous times of Dusty Baker’s infamous “base-clogging” quotes. The common thread is that most of these teams are putting speed at the top of the lineup. On-base percentage, even batting average, are afterthoughts. It’s all about stolen bases. We don’t need a new stat to tell us that’s bad strategy, but I think LOR passes the test on these teams. It’s drawn astray once or twice by really bad hitters who have low ISO because they’re really bad hitters, but on each of these teams, one of the top two in LOR is probably the optimal leadoff man.

As a final test, I calculated LOR for every qualified batter in 2013 (n=140). The top 10:

1. Marco Scutaro, .285
2. Norichika Aoki, .272
3. Elvis Andrus, .268
4. Dustin Pedroia, .258
5. Jon Jay, .257
6. Gregor Blanco, .256
7. Joe Mauer, .252
8. Billy Butler, .251
9. Joey Votto, .249
10. Shin-Soo Choo, .246

As a group, those players hit .295/.376/.412. All except Andrus had OBPs over .340, with three of the 10 over .400. They combined for 1,198 singles, 716 walks, 401 extra-base hits, 131 stolen bases, 68 HBP, and 54 caught stealing. Thus, the average player in this group got 160 H, 72 BB, 10 HR, 13 SB. That’s a pretty nice season for a leadoff batter, but probably not a good fit batting third or fourth. Most of these players were uniquely valuable at the top of the order.

It’s a strength of the system that it doesn’t suggest great all-around hitters like Troy Tulowitzki and Miguel Cabrera, nor really terrible hitters like Alcides Escobar and Darwin Barney. It’s a simple formula, but it usually locks on to the right skill set. In 2013, Chris Davis was last, by far, in Leadoff Rating (.022). Last season, Davis had a .370 OBP and .348 ISO. Really low LOR is found almost exclusively in high-level sluggers. As a point of reference, Yuniesky Betancourt has a middling career LOR of .158, and his .097 last season in Milwaukee was a career-low. In more than a handful of plate appearances, you need a really high ISO to score much under .100 in LOR.

Leadoff Rating is not a perfect stat, but it’s the best formula I’ve seen to identify leadoff men without repeatedly tabbing great middle-of-the-order hitters. It’s a simple formula and it works as a general guide to select effective leadoff batters without resorting to the Tulos and Trouts and Ted Williamses who can do more damage a little deeper in the lineup.





Brad Oremland is a columnist for Sports Central, where he writes mostly about the NFL. He roots for baseball teams named after birds, except the Blue Jays.

16 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
jose
9 years ago

For a moment there, I thought it said “Forging the ideal Adolf Hitler”.

LHPSU
9 years ago

To me, you need speed at leadoff not because of stolen bases or some base-clogging theory, but simply to reduce the chance of getting runners wiped out, especially on double plays, and to move up efficiently on balls put in play (more high percentage extra bases/tag ups). OBA is absolutely the more important skill, but the absolute disregard for speed from some quarters seems to have reached the point of ridiculousness.

Lanidrac
9 years ago
Reply to  LHPSU

You don’t need that much speed to be an effective baserunner. Yadier Molina is one of the slowest guys in the National League, yet he’s a great baserunner and manages to swipe about 10 bases a year just by picking the right spots.

vslykemember
9 years ago

Doing this calculation for the Braves suggests La Stella should be the leadoff hitter, which sounds about right.

Lanidrac
9 years ago

Very impressive! The stat could use some more work to clog down guys with mediocre OBP, but I really like where you’re going with this. Just one question: what do you think of guys with a high OBP and low homer totals yet hit lots of doubles?

Strangely, Ozzie Smith always batted 2nd in the order, or at least he did for Joe Torre’s Cardinals, as the Whiteyball era slightly predates my following of baseball.

filihok
9 years ago

Interesting.

I follow the Dodgers most closely and LOR suggests AJ Ellis (.343) or Dee Gordon (.230) as leadoff hitters.

I’m still not convinced that an overall better hitter getting the most at bats per game isn’t the best idea.

Also, instead of just looking at season-to-date stats we can look at projected ROS stats. Using ZiPS ROS:
1) Ben Revere .259
2) Craig Gentry .254
3) Elvis Andrus .253
4) Joe Mauer .253
5) Nori Aoki .252
6) Munenori Kawakami .252
7) Ruben Tejada .250
8) Daric Barton .250
9) Juan Centeno .247
10) Nick Punto .246

522) Evan Gattis .070
523) Alfonso Soriano .069
524) JP Arencibia .068
525) Miguel Olivo .059
526) Juan Francisco .059

And for the Dodgers:
Dee Gordon .236

foxinsoxmember
9 years ago

Cool article. Baserunning is an area for improvement. That’s probably the reason why someone like Avila is not a ‘leadoff guy’.

How about adding something simple for BsR like: OBP – ISO + Constant * BsR, maybe Constant = .01?

Alternatively, simply eliminate anyone with negative BsR?

filihok
9 years ago
Reply to  foxinsox

Is there a reason to believe that base running is more important at leadoff than anywhere else?

It seems that leadoff would be the least important place for baserunning since the hitters coming up should be able to drive runners in at a higher rate than hitters lower in the order.

Leo Martinic
9 years ago
Reply to  Brad Oremland

This is a really good idea. Why are you incorporating PA/600? I think it would be a mistake to discard all players with a negative BsR since that specific metric is not perfect. Also think a good modification to the formula would be to replace OBP with wOBA. Using wOBA would be more indicative of true ability to get on base.

Also, do you think there is a correlation between with players that have a high LOR and pitches per plate appearance?

tz
9 years ago

One issue with subtracting ISO is that there’s always a little piece of it tied to a guy stretching a single to a double or a double to a triple. If you replace ISO with (XBH+4*HR)/PA, and reflect baserunning by (SB – CS)/PA so you can reflect baserunning across eras, you might get a better reflection of speed into the equation.

Tim A
9 years ago

We already know the best leadoff hitter, his name is Ricky Henderson, and your welcome to try and find another as good, but you won’t…

Go Nats
9 years ago

I think your measure eliminates to many doubles. I would prefer a doubles hitter who walks to any other type of leadoff hitter. Wade Boggs was a great leadoff hitter. He was not fast, but he hit tons of doubles and had an all time great on base percentage for his career.