Remember me

All Fastball-Happy Families Are Alike. Or Aren’t They?

February 3, 2016

I’m drawn to the notion of baseball families, in the Jamesian sense of "players who share strengths and weaknesses," and I’m particularly drawn to the unproven (and very possibly unprovable) idea that if you find two pitchers who are very much alike, then they both should be effective against the same hitters and ineffective against a different group of hitters (else they’re not so much alike as you would think). So, having no book deadlines or any other pressures in my life (and how I wish that that were true!), I decided to spend an afternoon exploring looking for candidates to test this theory out on. That afternoon has grown into a month, as I tried various standards for finding the most suitable group of players, who would yield the largest samples of pitcher/batter matchups in a study where the sample sizes, try as I might, were bound to be very small.

If you’re familiar with the concept of pitching families, you can skip this paragraph and the next one, but for anyone who hasn’t read previous discussions of the concept, families are a refinement of platooning, a complex kind of Stengel/Weaver platooning where we go beyond the simple idea of "lefties bat better against righty pitchers, and vice versa" and try to break players down into smaller units than just lefty/righty in an attempt to find an edge.  For example, if a batter did especially well (or poorly) against a right-handed power pitcher like Bob Gibson, it might stand to reason that that same batter would have similar results against other right-handed power pitchers, such as Don Drysdale.  If this can be demonstrated, that would be very useful information to know about that batter.

The chief problem here is that of sample size. In any given matchup of pitchers, we will be dealing with a sample size no larger than 200-odd Plate Appearances against any one batter, and that’s just too small a sample to be reliable (and very few matchups even approach 200 PAs).  If someone has identical performances against two pitchers in a sample that small, it might be valid information or it might just be a coincidence that in future plate appearances will prove unreliable.  My thinking here is that if we can identify a larger number than a single pitcher/batter matchup, some meaningful patterns might emerge. But how do we identify larger families of pitchers and batters?

I needed to start by identifying a period and league that would give me a group of pitchers who would have sizable matchups with the same batters.  BBREF’s sim scores are astonishingly useless, telling me, for example, that lefthanders are sometimes good matches for right-handers. For this experiment, at least, that’s nonsense.  I’m not interested in outcomes (won-lost records, or ERAs, or any of that). I want pitchers who pitched like each other, similar stuff, similar styles, obviously similar handedness. ( does divide each batter’s performance against "power," "power/finesse," and "finesse" pitchers, but again they don’t distinguish between righty and lefty pitchers.) So I searched instead for one particular trait: I went for high K/IP guys (a stat that BBREF calls "SO/9"), reasoning that almost all of them would have excellent fastballs. By picking the era I was most familiar with, I could then eliminate on sight the lefties and the high K/IP junkballers, if any cropped up.

The era and the league I’m most familiar with is the NL in the 1960s, and I had in mind those two excellent pitchers with very high K/IP ratios whom I knew to have similar styles, Don Drysdale and Bob Gibson, both right handers, both very fast, intimidating, both the very definition of "power pitcher," and for the purposes of this study, both born around the same time. I drew up a complete list of about twenty young (under 26 years old) right-handed pitchers who in the years 1959-1963 put up a K/IP ratio in a season of more than 100 IP that was significantly above the National League average.   I then eliminated those who failed to sustain that K/IP ratio (Don Cardwell had a very consistent 6.5 or so SO/9 ratio in his first 800-odd IP, then fell to a sub-league ratio of about 4.5 in his last 1300 IP); those who were traded to the AL for most of the rest of their careers (Stan Williams and Ray Culp, who both came up with NL teams as real strikeout artists); those whom I knew to employ pitches other than fastballs as their primary pitch (the knuckleballer Ken Johnson turned up one year, 1962, with an 8.1 SO/9 IP ratio); and those who, despite an impressive SO/9 IP ratio early on, couldn’t even stay in MLB for very long past their brilliant beginnings (Art Mahaffey, Ernie Broglio and Glen Hobbie, for example, all washed out of MLB by the age of 30.)

In the course of looking into things like this, I always waste a lot of time by getting distracted by things I come across that I know are completely irrelevant to the study at hand—for example, as I was eliminating Glen Hobbie from inclusion in the list of young right-handed power pitchers of the 1960s, I found some statistical oddities in his career. Since they’re irrelevant to this column, let’s just call this my Hobbie paragraph, and I’ll get back to my Job soon. The good part of Hobbie’s career, that sustained him through the bad part that I actually witnessed, was his first two seasons, 1958 and 1959, when he went 26-19 with slightly over 5 strikeouts per 9 IP.  For the rest of his 8-year career, Hobbie went 46-62 with slightly under 5 SO/9.  His decline was almost comically linear. With the exception of 1963, when he had a slight uptick, his W/L percentages dropped literally every year he had a W/L percentage: .625, .552, .444, .350, .263, .167.  That’s like watching a cartoon anvil fall from the Acme building, isn’t it? But it’s astonishing how good Hobbie was in the one really good season he had: at the age of 23 when he went 16-13 for a losing Cub team, he threw 234 innings and had a 106 ERA+ with 5.3 Ks every nine innings, numbers that reminded me of Tom Seaver’s rookie year. That 5.3 SO/9 placed 9th in the NL, and put him on my radar screen for this column, but his good fastball didn’t have a long lifespan. He hurt his back doing track exercises in the first spring training of the Cubs’ ill-fated College of Coaches experiment, and as he put it in his SABR biography "I lost quite a bit off my fastball, and everything went downhill from there."  Since I was looking for pitchers who sustained an excellent fastball throughout the 1960s, I had to drop him from this family of pitchers, but truth to tell, his name wouldn’t have even been considered as a righty power pitcher if I’d been relying only on my memory.

Enough of the Hobbie, and back to the Job:  below is the family of righty pitchers under 26 years of age in the NL from 1959-1963 who had above-league average SO/9 ratios in at least one season of over 100 IP, and who sustained an above-league average SO/9 IP for their NL careers of over 1000 IP. The "height" column is just something that didn’t correlate: I thought they’d all be very tall. They weren’t, though it was freaky when the first two I charted, Drysdale and Marichal, each had identical heights and SO/9 IP ratios. I thought I was onto something for a moment.







Bob Gibson 






Don Drysdale






Juan Marichal






Bob Bolin






Jim Maloney 






T. Cloninger 







My rationale for choosing this family to study is that they faced many of the same batters in the 1960s and early 1970s (I included any batters Drysdale and Gibson also faced in the late 1950s, of course). This was to minimize small sample-size batter-pitcher matchups, while maximizing those that approached statistical significance. If I’d wanted to be a little purer, or make my study simpler, I could rationalize removing Cloninger and even Marichal, as the least dominating SO/9 ratios, or Bolin for throwing the fewest NL innings, but I think a family should consist of more than just 3 or 4 members. Certainly, if I’d wanted to do more work, I could have greatly expanded the family by including all of the pitchers who threw an above-league average SO/9 ratio at least once from 1959-1963, that would include all the rejectees named above (Williams, Culp, Mahaffey, Broglio, Hobbie, Cardwell, plus Joey Jay, Al McBean, Ray Washburn, etc.), plus many of the pitchers who accomplished that in 1959-1963 but at an older age than 26, such as Robin Roberts, which would have brought the family size up over 30 pitchers. And if this smaller study shows anything conclusive, I suppose the next step would be to see if a more loosely defined family also shares these commonalities.

In an earlier draft of this study, I selected pitchers less systematically, choosing purely by eye and memory (I included Culp, and excluded Cloninger and Bolin) and found, as expected, that most hitters didn’t fare well against these pitchers (who were, after all, the best power pitchers I could find near their peaks) but that there was no real pattern in batters who did hit one of these pitchers well. Joe Adcock, for example, really tee’d off against Marichal —he had a 1.063 OPS against Juan, which is almost unimaginably good, and he hit Bob Gibson even better: 1.097 OPS.  So maybe you expect from these numbers that he’d be able to hit Drysdale as well, but no dice: Drysdale held him to a very poor .467 OPS. This lack of a clear "family" pattern was disappointingly typical, if a little extreme, and about what you might expect in small sample sizes, a very unstable set of numbers. But if we took the total number of PAs that Adcock had against the entire family, maybe it would show that his numbers against Marichal and Gibson were unusually high, and Drysdale unusually low, but in the group Adcock’s numbers still registered as overall very high.  In other words, a very low OPS against a particular family isn’t guarantee that he will hit each family member very poorly, but it might be a good reason to give him an off-day when that type of pitcher is scheduled to face him.

What I was looking for was batters who had special trouble hitting power-pitching right-handers across the board (or those rarities who ate up such pitching). My goal behind this search was that maybe, just maybe, we could find a method for identifying such matchups by family, and thus sit down (or move up in the lineup) batters who are overmatched (or who love to hit) such pitching, even if they haven’t faced a particular family member yet.  So this is really the first part of a much longer study, and that will come to pass only if we can actually discover such patterns. For now, I’m just trying to see if any patterns emerge through the miracle of and hindsight.

My plan is, first, to find good and bad anomalies in the largest sample sizes I could find against each family member, and draw up a good-hitting list and a poor-hitting list for each of these six power-pitching young righties, then to expand that list by seeing how each batter did against the other pitchers on the list. I’m using OPS for the batters because that seems to me the most inclusive stat that can be ranked in the pitcher vs. batter matchups on  If there are any batters in this study who hit consistently poorly or well against this type of pitcher, then I’ll look into other pitchers whom this batter did well or poorly against to see if the pattern holds.

For example, in my proto-study (which used Ray Culp instead of Bolin and Cloninger) I found that Bob Aspromonte couldn’t hit any of these guys with a paddle: he "OPSed" (I’m just going to say "batted" or "hit" instead from here on in, but you’ll know I mean OPS) about .550 against the first group of five pitchers:

Aspromonte vs…












After his name popped up on both Drysdale’s "bad" list and Gibson’s (and, oddly, the identical OPS against both pitchers came in a different number of ABs), I looked to see how he did against the other three pitchers, and voila! You have to admit this gives me some reason to think I might be onto something here. Aspromonte, a right-handed batter, had almost no platoon difference, lifetime OPS in the .640s against both righties and lefties, so this suggested perhaps the sort of righty you don’t want him batting against. He underperformed against all of these five pitchers, most of them significantly.                             &nbs​p; 









































































This total is a very poor showing for 400 plate appearances, especially the power figures, 2 HRs and 21 RBI. (The SLG and OPS figures are approximate, but probably not off by very much.) If Aspromonte hit every pitcher like he hit these five, he obviously wouldn’t have had a major league career longer than a week or two.

We will explore how Aspro did against right-handed fastball pitchers other than these five in a future column, but I just wanted to give you some idea of the method I’ll be using, and the pitchers I’ll be applying that method to. I also wanted to share some of the funnier quirks I came across while studying these matchups in very small, and not-so-small, samples.  One weird thing happened when I decided to study the batters one-by-one.  Once all my data at the time was in, I thought I might as well organize them alphabetically:  Hank Aaron, Richie Allen, Matty Alou, Bob Aspromonte and a lightbulb went off—Hey, all these guys have brothers who played MLB! Then I realized that I’d forgotten about Joe Adcock, and the lightbulb dimmed—Hey, don’t draw conclusions too fast from tiny samples!                                 ​     

I might as well explain right here how I happened to select those particular "A"-surnamed batters, and not their brothers. I ranked each of the five pitchers by opposing batters’ OPS against each of them, and then scrolled down that list, looking for batters who had at least 70 PAs against them. I then drew up two lists for each pitcher, one of batters who hit them particularly well, and the other of batters who hit them particularly poorly, in 70 or more PAs.  Finally, I would then cross-check each batter for how he did against each of the other pitchers—sometimes, as with Adcock, he would show up on multiple good and bad lists, or as with Aspro, he’d show up on multiple bad lists,  but most times he would show up on only one pitcher’s good or bad list. My original five pitchers had 31 different batters between them on their bad lists.

The 70 PA standard was arbitrary, and so was my standard for deciding what constituted good and bad OPS performances. The average league OPS during the 1960s was in the high .600s, so anything in the .600-.800 range generally was considered normal. Below .600 or above .800 generally got flagged as an outlier.  Managers, not being notably stupid, tend to figure out when someone can’t hit a particular pitcher with a shotgun, so naturally the "bad" lists tended to include more batters who were in the lineup for reasons other than their bats— unsurprisingly, a lot of middle infielders and catchers. BUT (I figured) it would be interesting to see if these pitchers made their livings by totally dominating that sort of batter, while some batting stars did quite well when facing them, which turned out to be the case.

When I had re-arranged each pitcher’s stats by batter in descending order of OPS, of course, the numbers at the top (in teeny-tiny samples, of course) would belong to batters with very high percentages in very few plate appearances. I got some very obscure names.  Juan Marichal’s batters arranged in OPS order, for example, features the name of our old friend Don Leppert at the very top of the list. Leppert faced Marichal twice in his career, with a HR and a walk, for an OPS of 5.000. The complete list of batters who have OPSes 2.000 and higher against Marichal is a list of batters you’ve either never heard of, or terrible hitters, or pitchers-- very few regulars, let alone stars, show up on this list (Dave Winfield at the start of his career and Harvey Kuenn at the end of his are about it).  Of course this is perfectly understandable--if they were regulars, they’d have faced Marichal more often, duh-- but it’s kind of funny to me to think about such marginal MLB players as Leppert and Howie Goss and Joe Lis, all of whom rank among the top-seven OPSes vs. Marichal, knowing that their faces would light up for the remainder of their lives if you ever asked them how they batted against Juan Marichal. (Goss had four at-bats, three singles and a HR—now, there’s one big stupid grin.)

And before I get started, a note about how long ago this project began, way back on the afternoon of July 20, 1966, when I was a teenage guitarist. I have a startlingly clear memory of my kid brother interrupting a taping session with my garage band, screaming "Luplow hit a homer!", ruining the song we were trying to record.  ("You Really Got Me," I think. The music industry survived this setback somehow.)  So when BBREF came out with their boxscores of old games, I realized that I could figure out exactly what day this recording session took place, and why the kid was so jazzed about the HR—turned out it was hit off Marichal, and it was very dramatic, and so on (I wrote a whole essay about memory-triggers from this incident), but I also learned that Luplow, in only 26 at-bats, hit Marichal like a gong, batting .308 off him with 3 HRs. Again, this just pleased me to think that if you ever mentioned either Luplow’s name or Marichal’s to the other, you might get a strong emotional response. And if you ever run into either of them, try it and tell me what they say. Looking further into Luplow’s stats (he was a very marginal outfielder who spent about a season with the Mets), I also learned that he hit more than one HR off only five pitchers, and all of the five are members of this extended family of pitchers: three Hall of Famers, Marichal, Gibson and Robin Roberts, who was too old to qualify but otherwise a pure power-pitching NL righty, plus Milt Pappas, a 200-game winner who would have made the cutoff in 1962 but he was pitching in the AL at the time, and Ray Culp. I don’t know if this is just another small sample-size quirk, but it suggests at least that Luplow could get around on a good fastball, doesn’t it?

Then at the end of December, 2015, Sandy Koufax’s 80th birthday got a lot of play—on Facebook, a woman I vaguely know claimed to be a cousin of Koufax, so I asked her if she was an actual cousin or if she just batted particularly well against him.  (She was Sandy’s mother’s sister’s daughter.)  Then, on a whim, I decided to look up who Koufax’s baseball cousins actually were, and learned that Henry Aaron hit Sandy Koufax better, not worse, than he hit an average pitcher (.362, with 7 HRs in 116 ABs), which I found very impressive. Noodling around, I learned that Clemente and Frank Robinson also hit Koufax really well, while Joe Torre and Bill White couldn’t hit Koufax to save their lives (all five in over 100 Plate Appearances).

Of course, 100 Plate Appearances is nothing, but OTOH very few pitcher/batter matchups reach triple figures, so while it may be nothing statistically speaking, in terms of baseball it’s as close to significant as you can get.  Do White and Torre have some commonality that might explain their miserable histories batting against Koufax, other than bad luck in small samples? If I studied them against pitchers similar to Koufax, how would Aaron, Clemente, and Robinson do? Who else batted especially well or poorly against Koufax, and other left-handed flamethrowers? Is there anything to any of this other than small samples and luck?  Anyway, that’s what got me started on this small sample-size matchup stuff. If this righty power-pitcher study proves fruitful, I’ll repeat it for lefties.

More small-sample oddities I ran across:

  • Harvey Kuenn, who made the final out in Koufax’s no-hitters against the Giants and the Cubs, was 8-for-70 (6 singles, 2 doubles) against Koufax, lifetime. By the time of the Cubs’ no-hitter,  Kuenn was at the end of his career, so these numbers were in the books. Maybe the Cubs’ manager should have sent a different hitter –any different hitter, including a batboy-- to the plate?
  • Clemente appears on Bob Gibson’s list of batters he could get out (.563 OPS), in a largish number of at-bats (166), but I’d bet the most common response if you said "Gibson vs. Clemente" to a fan of that era would be the (almost) killer line-drive Clemente hit literally off Gibson, breaking his leg in 1967.  Clemente’s teammate Manny Mota, incidentally, claimed that Clemente deliberately broke Gibson’s leg as retaliation for Gibson’s pitching the Pirates high-and-tight, which would be phenomenal bat-control if true. That sort of bat control would also result, you’d think, in Clemente having an unusually good OPS against Gibson, not an unusually poor one.
  • Sometimes light-hitting middle-infielders do pretty well against HoF pitchers: Julio Gotay, a .260 BA hitter in over a thousand MLB PA, hit exactly a hundred points higher (18-for-50) against the five pitchers in my first attempt at a study. He also went 3-for-6 against Koufax.  That includes 8-for-16 against Gibson and 7-for-16 against Marichal.  Elio Chacon batted .353 against Koufax with one of his lifetime 4 HRs, and went 2-for-4 vs Gibson and Maloney.
  • As noted above, Drysdale had Joe Adcock’s number: in 70 at bats, he struck Adcock out 28 times, and gave up only 9 hits. And Richie Allen was practically helpless against Drysdale: in 49 PAs, Allen managed only 3 singles but 17 Ks, for an OPS of .150.
  • Luplow’s small-samples suggest further he could really hit a fastball: among his 1.000 or better OPSes (in small samples, of course) are Ray Culp (2.548 in 7 PA) and Robin Roberts (1.748 in 15 PA), plus Marichal  (.962 in 26) and Gibson (.925 in 18), just among those named here already. In the very high .800s, still an impressive OPS, Luplow’s list includes Ray Washburn (.883 in 15) and Don Drysdale (.871 in 12).

PART TWO: How Consistently (or not) This Family of Pitchers Performed Against a Common Group of Batters.


COMMENTS (9 Comments, most recent shown first)

Brock Hanke
I'm not completely sure about your method for identifying pitcher families, because it doesn't take into account which pitches the people actually threw. Marichal, for example, was famous for his curve more than his fastball, while Gibson featured the slider as his breaking pitch. I don't know how you can account for this, since I don't think we have type-of-pitch date for many pitchers. I wonder if Aspromonte's problem, for example, was that he couldn't hit a slider, but could pick up a curve and get around on it. If Maloney featured a curve, that would be real good evidence for this.
11:10 AM Feb 10th
I think the concept of pitching "families" is more interesting in the sense of possibly predicting the career arc of similar pitchers. I think using the concept to attempt predicting how batting "families" might do at the plate against the pitching "families" is less useful than just studying how specific batters do against specific pitchers...yeah, I get the small sample problem.

1:30 AM Feb 6th
Steven Goldleaf
From that paraphrase, you might suppose that Koufax was able to get Mays out with some dispatch, but actually Mays killed him, too, a .962 OPS with the most doubles and the most walks (by far) off Koufax than anyone, including Aaron.
8:39 AM Feb 5th
Two unrelated points ...

1. I feel pretty sure that Aaron hit Koufax well throughout their time together in the NL, because Koufax mentioned that at some length in his 1966 autobiography. In a very loose paraphrase, he said something like, "With Willie Mays I have some options that might work, but with Hank Aaron I just throw it up there and hope it stays in the ballpark."

2. As for using Kuenn against Koufax in that 1965 game, Kuenn must've been a pinch-hitter, because it was a perfect game, so the last batting spot would have been the pitcher's. So there wasn't any question of keeping him out of the lineup for a lesser player; it was just a matter of picking whoever might have the best chance of getting a hit in that one at-bat. Which might've been Kuenn anyway, I suppose.

5:29 PM Feb 4th
Steven Goldleaf
Bruce, good point about Koufax, although he threw fewer than 700 of his 2324 IP before 1961 (and he wasn't a terrible pitcher then, just not Koufax), so I don't imagine that would skew the numbers very much.
4:23 PM Feb 4th
About Kuenn and the sample size: Unless the difference between Kuenn's record against Koufax and the average record against Koufax were larger, and especially since (as I said) the guy you'd be putting in there instead of him would be a below-average player, I don't think I'd ever consider replacing him in the lineup, no matter now large the sample size were. I mentioned "sample size" because it happens to be an additional factor in such an instance, an extra 'argument,' but I don't think Kuenn's record vs. Koufax was poor enough (compared to average) to make one ever think of holding him back.​
9:41 AM Feb 4th
Hey you really got me goin'
You got me so I can't sleep at night


Steven: This is a fascinating approach to a very interesting subject. I eagerly await the next installment.

I especically found it interesting that Aaron and Robinson, who I think belong to the same family of hitters, both did well against Koufax. Part of it could be that they were in the league in 1956-1960, I suppose.

I thought Chris Krug made the last out of Koufax's perfect game against the Cubs. Why did I think that? Bob Hendley pitched a one hitter in that game.
9:38 AM Feb 4th
Steven Goldleaf
The idea may already be in practice--see my early reference to Stengel and Weaver, whose ideas have been copied by several generations of managers--but should it be? One of the things I'm finding not clear at all is the wisdom of saying "This guy shouldn't be able to hit this pitcher, so I'm going to sit him down." Drysdale and Gibson, to take the two pitchers who seem extremely similar to me, often DON'T have similar performances against individual batters. But I can't really think of a reason a batter would hit Gibson well and Drysdale poorly, or why a manager would think so in advance, so the manager who decides that someone can't hit a power-pitching righty might just be missing a chance to hit Gibson if he sits that someone down. Julio Gotay should be overwhelmed by Gibson, except he wasn't.

And again it might just be small sample size that accounts for Kuenn's performance against Koufax, but at what point would you decide that maybe it's NOT small sample size? 170 ABs? 370? 700? Or do you just keep saying, "He's Harvey Kuenn, he's got to start hitting Koufax sometime"? If the Cubs' manager (or Head Coach, I guess) had had these numbers in front of him, do you think he might have chosen otherwise?
4:38 AM Feb 4th
A couple of things (unrelated or at least not very related):

I think the basic thing you're getting at -- "maybe, just maybe, we could find a method for identifying such matchups by family, and thus sit down (or move up in the lineup) batters who are overmatched (or who love to hit) such pitching, even if they haven’t faced a particular family member yet" -- is a thing that is already very much done, only I think it's done mostly the old-fashioned way: by looking. Do you not think so? What makes me say that is simply that we hear about it all the time. I hear it so commonly -- it's so familiar ("he's sitting because he doesn't hit this type of pitcher very well" sometimes with an addition like "and he can use the day off anyway") -- that I couldn't even tell you if I've ever heard it from managers or coaches or just from announcers. I realize that what you're doing is looking for a systematic way of doing it, rather than just through observation -- but I think the idea is already in common practice.

I realize also that part of what you're doing is looking for instances where it might be worth sitting a guy who wouldn't ordinarily be thought of as a guy to sit against anyone, for example Harvey Kuenn vs. Koufax. BUT, the overall averages against Koufax during his great stretch weren't much better than Kuenn's -- yes, better, but not a whole lot, and of course the sample size of Kuenn's plate appearances against Koufax was pretty small. Even in this example which seems like an extreme one, I don't think the difference between Kuenn's record and the average record would have been enough to overcome the relatively small sample size, and the fact that this was Harvey Kuenn -- not to mention that the guy you'd be putting in there instead of Kuenn would be a bench guy, a below-average player.
12:25 AM Feb 4th
©2023 Be Jolly, Inc. All Rights Reserved.|Powered by Sports Info Solutions|Terms & Conditions|Privacy Policy