In many ways Willie Stargell’s twin, Willie McCovey "is probably the only truly great player in history to have been platooned for several years at the start of his career." McCovey’s obstacle, as Bill notes further in the Historical Abstract, was simple: Orlando Cepeda. It took the Giants several years to figure out what to do with two young world-class first-basemen. But Stargell, too, had his career delayed, perhaps not quite as dramatically as McCovey, nor as much by platooning, but significantly.

Coming to the major leagues at age 22, Stargell was relegated to the bench or to shuttling between left field and first base for years.  His first full season, 1963, he basically was the backup left fielder to veteran Jerry Lynch.  Pittsburgh had a logjam at LF and 1B in 1963:  they had Lynch as their primary leftfielder (59 starts) and Bob Skinner (30), Ted Savage (17), and Manny Mota (16), in addition to Stargell (30), who was the youngest and the least tested of the five men starting at least 10% of Pittsburgh’s LF games.  Donn Clendenon, meanwhile, had just established himself at first base, having wrested it away from Dick Stuart in 1962 with a second-place finish in the ‘62 Rookie of the Year voting. (Cepeda had won the ROTY award outright the year before McCovey arrived in SF.) Both LF and 1B seemed amply staffed in 1963, and in 1964, Stargell played the second-most innings on the team at both positions, backing up Lynch and Clendenon once again, starting 53 games in LF and 49 at 1B. Remarkably, however, he made the 1964 All-Star team, thanks to a very hot start.  (He began the season by driving in 28 runs in the team’s first 27 games, batting .358 and slugging .642. For the rest of 1964, he batted .244 and slugged .452.)  The only other positions available to the left-handed-throwing Stargell were CF, for which he was plainly too slow (though they tried), and RF, occupied by a Hall-of-Famer in his prime.

In 1965, Stargell again made the All-Star team in LF, but over the complete season, he started only 118 games in left, sharing it with 34-year-old Lynch (21 starts), 27-year old Manny Mota (who got into 35 games in LF as a defensive substitute)  and 22-year old Bob Bailey (24 starts in LF). The next season, Stargell played even fewer innings in LF, although by now he had established himself as a perennial All-Star, MVP candidate, and slugger. In 1967, Stargell started only 86 games in LF, while serving as Clendenon’s backup at 1B (33 starts). Stargell’s primary position was LF, even after the Pirates let Clendenon go in 1969 (his replacement at 1B, Al Oliver, just as Clendenon had, finished second in the NL ROTY voting as Pittsburgh’s first baseman), but Stargell only played LF on a steady, everyday basis (at least 120 starts and 1100 innings) beginning in 1971, his tenth year in MLB, which seems like an awful lot of being dicked around for a first-ballot Hall-of-Famer. McCovey had established himself as a regular at one position, by way of contrast, at the start of his seventh season in MLB, when McCovey played 156 games at 1B. Cepeda spent most of that year on the DL, and was soon traded away.

All in all, McCovey made 88% of his career starts at a single position, 1B, while Stargell made just 58% of his career starts in LF and another 39% at 1B. Of course, Stargell played most of his career games at 1B late in his career, but the lateness of being settled in one position is just another similarity of their careers, having little to do with their very similar batting records, which is my main focus here.  

I looked into how they fared over a larger group of pitchers than I did a few weeks ago—in addition to their overlapping career spans, and their similar handedness, and their identical career OPSes, with very similar OBPs and SLGs, they faced identical pitchers numerous times over their MLB careers, so I thought that above a certain number of plate appearances, I might find similar outcomes against the same pitchers, which would support in practice the theory of "batter families." More precisely, if I couldn’t find any meaningful similarities in this tiny family of two, then there would be little point to pursuing the family theory any further.

What follows is the list of every pitcher against whom either McCovey or Stargell had 90 or more plate appearances, listed in order of their plate appearances against either batter (i.e., Don Sutton faced Stargell 165 times so he’s first on the list,  Drysdale faced McCovey 151 times so he’s third on the list, etc.)

Since both Stargell and McCovey had lifetime .889 OPSes,  you can see which pitchers lowered and raised that figure against each batter.  Of these fifteen pitchers, we need to eliminate Gaylord Perry and Juan Marichal for comparative purposes, since McCovey was their teammate so didn’t face Marichal at all and faced Perry an insignificant (13) number of times. Of the remaining thirteen pitchers, more than half posted OPSes against both McCovey and Stargell that fall into my description of "very close," i.e., within a .100-point OPS range against each of them. (I’ve highlighted these seven pitchers in yellow.)  Sutton was the toughest for both Willies, Cloninger the easiest, but all in all seven pitchers placed within a hundred OPS-point range against both hitters. (As I’ll explain in a later article, a .100 OPS point range may look large but it actually represents similarity rather than difference.) 

Of the remaining six, Seaver had a .114 OPS point-difference,  Jenkins .167,  Niekro .161,  Bunning .198 and  Jackson .221.  Of the 13 pitchers, only Reuschel had a difference outside the (admittedly arbitrary) range of .300 OPS points that I deemed to be "large," .346—Reuschel pitched effectively against McCovey (.736) and ineffectively against Stargell (1.082). Otherwise, every pitcher on this small list pitched about as well against one batter as he did against the other, suggesting that in this very small family of batters, there is a certain consistency of pitchers’ outcomes.

Next, I decided to test my thesis  by looking in a slightly different way:  what if I lowered my standard to at least 50 PAs against either Willie, and narrowed my search further by limiting it to pitchers who were effective (under .700 OPS) against at least one Willie with the 50 PAs? Would the other Willie also prove ineffective against the same pitcher?

This part of the study worked in a few isolated cases (Dave Giusti and Jon Matlack, as you can see by the chart below) but mostly the results were all over the map. Probably 50 PAs against only one Willie was just too small a sample size to yield stable results (or maybe I have rocks in my head). In any case, five of the matchups ended up with 20 or fewer PAs for one Willie, which is obviously too small a sample to be  reliable, and several other matchups yielded no very strong correlations. The three matchups that featured at least 65 PAs against both batters, Claude Osteen, Jerry Koosman and  Ron Reed, resulted in fairly close OPSes (.185, .173, and .244, respectively) but otherwise these matchups were all over the map, the more so as the number of PAs against one batter approached zero.

Now of course, it might be that the sample size of virtually any pitcher/batter matchup will be too small to yield remotely meaningful results (I haven’t neglected that very real possibility) but it might also be that I’ve simply lowered the PA bar too much to get meaningful results—certainly the 14 plate appearances in which Matlack faced McCovey, holding him to a .220 OPS is not the sample size I’d want to base a huge decision on (like "Matlack’s pitching? Gotta sit McCovey down!"). 

So I decided, next, to look from the opposite end (pitchers whom McCovey and Stargell ate for lunch) and to expand the OPSes and PAs slightly, to a 1.000 OPS (.111 points above the Willies’ lifetime .889 OPS, rather than the .189 OPS below it with .700 OPS) and at least 60 PAs against either McCovey or Stargell. So, alphabetically this time:

This yielded 14 matchups, of which five (Cloninger, Drysdale, Jackson, Maloney and Pappas) had at least 60 PAs against both batters. I don’t want to make too much of Cloninger, Drysdale and Maloney here, because I already discussed those matchups in the previous "Give Me the Willies" article, but Jackson and Pappas now add a .221 and .220 difference, respectively to Cloninger’s, Drysdale’s and Maloney’s even closer differences. If we lower the bar to 50 PAs against both Willies, we can add another four:  Briles (.104 difference), Hands (.516), Rogers (.357) and Stoneman (.293).  In the remaining five matchups that feature at least one Willie without 50 PAs, the differences grow a little wider (Cardwell, .129 difference; Kirby, .188; J. Niekro, .733;  Renko .355;  Reuschel, .346).  The pattern, if pattern there be, is that the lower the number of PAs in a pitcher-batter matchup, the greater disparity between OPSes.

It’s always a bit of a Devil’s bargain in choosing whether to study fewer matchups with higher PAs, or more matchups with fewer PAs, but all indications so far stress that the more PAs, the greater the stability of outcomes, which is encouraging (especially because it seems to mean there’s little point in doing more studies of pitcher/batter matchups of fewer than 60 PAs at least, cutting down on my donkey-work considerably.)

In reviewing both Willies’ almost identical numbers, I decided to look more closely to see what significant differences, if any, there might be, to explain different performances against various pitchers.  They both had fairly sizable platoon splits, McCovey about .150 OPS points higher against lefties than righties (.926/.776), and Stargell .010 points more in each direction (.935/.765) .  Their managers’ strategy of sitting them down, in general, against lefthanded pitchers seems pretty sound: though they have large lifetime splits, for their first half-dozen or so seasons, those large splits were ruinously huge:

Or, in expanded form, vs. lefthanders:

and

And again in expanded form, vs. lefthanders:

In the "Astonishing Facts" Department, more prosaically known as the Department of Utter Coincidences, each Willie drove in exactly 37 runs in his first five years in MLB while facing left-handed pitchers. More significantly, their managers showed excellent judgment in declining to make either Willie a full-time player in this period, as their other batting stats against lefties showed very dubious results. Over their first five years in the majors, where each of them played very nearly a complete 162-game season against lefthanded pitchers, Willie #1 batted .188 with 9 HRs and the aforementioned 37 RBI, while Willie #2 batted .224 against lefties, with 14 HRs and—I forget how many RBI. I’ll combine the chart totals above for you:

Of course, fewer than 350 ABs in a full season of games hardly constitutes an overwhelmingly dispositive number of ABs, but if you add both Willies’ ABs and their RBIs, you get 74 ribbies in 648 ABs, which would make a poor total for a very full season, and their combined BA (.207) is downright abysmal. So you can understand why both of these first-ballot HoFers had some delay becoming everyday players, even aside from the imposing presence of Cepeda and Clendenon. They came by their platoon status the old-fashioned way: they earned it.

I apologize for getting distracted from my original goal here, which was to study a variety of OPS twins, or OPS brothers, or OPS cousins, but I guess I got fascinated by the extreme similarity of the two Willies. Another distraction for me in studying the Willies is the splits that baseball-reference.com runs extensively, including pitcher/batter matchups and platoon splits, but extending far beyond those two. Frankly, some of those splits puzzle me the more I look at them, since they would appear to contain significant data, but the more I think about some of these splits, the more they look like random, almost meaningless garbage data, oddball facts without real significance, though some of them (like platoon splits) bear some obvious value.  So I’m going to venture next into the area of discussing baseball-reference.com’s splits without having any real mathematical or statistical training, which is something like walking a tightrope without prior practice but, in order to steady my nerves, with a shot or two of whiskey first. That is to say, quoting some famous last words: Watch THIS!