Sometimes It's All About Cleanup

Plus Other Observations Around Joe Sewell and the Indian Teams of His Period

Because I became a baseball fan in 1985, I will always think of a triple-digit RBI, single-digit home run season as a Tommy Herr season. Herr hit 8 home runs and had 110 RBI for the 1985 Cardinals. Browsing, I happened to notice that Joe Sewell was part of this family when he had 106 RBI off just 4 HR in 1924.

I am not going to say that this was comparable. Certainly, from a theoretical standpoint, it was a lot more likely. A big reason is that OBP in the 1924 American League was .358; in 1985 in the NL, it was .319. Then, while it was much more common in 1985 than today, there were a much higher percentage of regular players in 1924 with fewer than 10 home runs, so a higher percentage of players who could theoretically qualify if they had 100 or more RBI. Home runs may also occur often enough in a modern environment that they cut down on other RBI opportunities in a non-trivial manner, cannibalizing other RBI, so to speak.

Just taking an empirical approach, we can also see that Sewell’s accomplishment¹ does not rate with Herr’s. Limiting ourselves just to the 1924 American League, the White Sox had two players, Earl Sheely and Bibb Falk, who combined for 9 home runs and 202 RBI.² But on the other hand, Sewell’s total did stack up quite well, as his 106 RBI placed him seventh in the league.

Herr was a one-off, with his top RBI season other than 1985 just 83 RBI, and his third-highest total 61. Sewell, though, looks like a veritable run producer, by the standards of today. He hit 15 home runs at Yankee Stadium while with the Bronx Bombers from 1931-1933,³ but in his approximate 10 years with Cleveland, he hit just 30, with a career high of 7. But despite this lack of home run power, setting aside 1924, he also had RBI totals of 109, 98, 93, and 92.

The 109 came in 1923, and it was an amazing season for Sewell. He hit .353 with a .456 OBP and finished tied for 4th in MVP. His 146 OPS+ was 30 points his best. In the summer months of June, July, and August he hit .416,⁴ and finished August on a 27-for-49 tear.

Not only was this a career year for Sewell, but he had plenty of help. The Indians scored 5.79 runs a game, 0.37 more than the second-place team in runs, the Yankees. In 1924, the were way off that, scoring just 4.94 runs a game. So, while one wants to be careful not to overanalyze or expect perfect alignment in numbers, given these facts, that Sewell had 106 RBI in 1924, almost matching his 1923 total, is surprising.

I’ve gotten wise to the fact that RBI with runners in scoring position and average in runners in scoring position are often dramatically different things (read: opportunity matters), but in cases of RBI that don’t necessarily compute from the basic stats, the textbook maneuver is indeed to consult the player’s runners-in-scoring position average. This more specialized basis does not at all lead to a closing of the gap in the superiority of the 1923 season, though. Here are Sewell’s batting averages by situation for 1923 and 1924.

1923: .358 no one on, .315 1st only, .381 RISP

1924: .335 no one on, .351 1st only, .283 RISP

RISP performance a priori seemed like it would be particularly important to understanding Sewell’s RBI because home runs were hardly in his repertoire. Players like him have to be station-to-station run scorers, as it were. A further look at his “bases occupied” splits confirms the RISP line is where we should indeed be laying down our hat. The Retrosheet numbers have Sewell with 8 RBI in 1923 that didn’t come from RISP, and only 10 that didn’t come from RISP in 1924. So production with “no one on” and “1st only” played very little role in his RBI.

In saying that RISP data is the key, however, I am referring not just to production but to opportunity. As we have seen, the production seems to have been all on the side of 1923. Average-wise Sewell overachieved with RISP in 1923 even versus his impressive baseline, while in 1924 his RISP average was 98 points lower than his 1923 average. Yet the two RISP RBI counts only differ by 7.⁵

Despite the large difference in his RISP averages, it is not true that his number of opportunities was vastly greater in 1924. He had 260 plate appearances with men in RISP in 1924, 250 in 1923.⁶ Upon seeing these numbers, my first thought was that RISP average must just not count for much next to opportunity. I will revisit this assumption later as I try to ultimately come to terms with Sewell’s RBI counts for the two years. But holding off on tackling this issue, the opportunity balance itself is curious, given what I said about how much better the Indians offense was in 1923.

A leading suspect for explaining the difference has to be that Sewell hit cleanup for all of 1924,⁷ while in 1923 he was only there starting July 1, after having batted sixth for his first 63 games. Given the limited scope of this piece, I’ll let you fall back on your own beliefs and knowledge about the relative advantage hitting cleanup might have given him, rather than doing something ambitious like creating a model lineup or obtaining normative batting order data on production and RBI. But the correlation is there to be noted. During his monster August in 1923, all spent in the cleanup slot, Sewell had 33 RBI in 23 games. Of course, with the way that he was hitting at that time, his RBI total figured to be pretty good regardless of his place in the batting order.

We are now going to get to some of those “other observations around the Indians” alluded to in the subtitle, so my apologies that we may go pretty far afield. When Sewell was moved to cleanup, it wasn’t accomplished through a straight swap of positions. First, light-hitting rookie third basemen Rube Lutzke moved from seventh to sixth to take Sewell’s old spot. Then the first base tandem of Lou Guisto and Frank Brower had been hitting cleanup. When Sewell was made the cleanup hitter, Guisto was jettisoned, never to play again after 156 previous games across five seasons, while Brower was dropped to seventh.

That Guisto had been hitting cleanup was odd, to be kind. The only team he’d ever known was the Indians, and the only real shot he had had before came in 1917. Granted, it came under much more difficult conditions; the White Sox had a 2.16 team E.R.A. that year, and the Red Sox were just a tick behind. But Guisto, a first baseman then as well, hit just .185 and slugged just .225 in 200 AB. In 1923, he was almost matching that to a tee at the time his season ended, with a .190 average and a .222 SLG. I would guess that manager Tris Speaker just conceived of first basemen as cleanup hitters, whether they lived up to the standard or not. Or maybe he thought the rest of the players were performing satisfactorily and he didn’t want to jerk people around.⁸

The other half of the first-base duo, Brower, offered more, particularly from our current Moneyball perspective. He made his debut with the Nationals⁹ well after his 27th birthday in 1920. He walked exactly twice for every strikeout for his career, while the league ratio was always about 1:1. The .285 batting average of his 1923 slash line would play much better today than it did then, but the rest of his line (.392/.509) looks like something you’d see with many very good first basemen starting around 1950, although for 1923 it looks out of place.¹⁰ Limited action with the Indians in 1924 wrapped up his major league career, but he seems to have been just the Pacific Coast League type, and he indeed hit .362 and slugged .583 for the 1925 San Francisco Seals, in a busy season of 186 games.¹¹

At the time of the lineup switch, Brower was carrying a .475 SLG, not far off where he eventually ended up, but he was hitting only .237. Put in the lineup every day, like Sewell, he got very hot in July and August, then didn’t quite sustain it in September and the first week of October. But overall, he hit .305 after the lineup switch, and that elevated his on-base percentage as well.

Whether Sewell and Brower offer the full explanation, the Indians offense went as they went, and was at almost dizzying heights in the summer.

Indians’ runs-scored-per-game by month:

April (13 g) 4.8

May 4.5

June 6.0

July 6.4

August 6.9

September 5.6

The potency and vicissitudes of the Indian offense might have served as a point of interest to diehard fans, but in the scheme of things, it was academic and diversionary. The team did not matter. They were felled by run prevention (0.10-a-game-over-league-average, and 6th in the league), bad luck (their record trailed their Pythagorean record by 7 games), and a big season from the Yankees, who roared to a 65-30 start en route to the world championship in the first year of Yankee Stadium. Posting a 82-71 record, the Indians finished 16 1/2 back of the Yankees and a 1/2 back of the Tigers, .

Expectations may well have been high before the season, as the Indians had won the World Series in 1920 and were 2nd in the AL in 1921 with a 94-60 record, which is better than .600 ball. Both of those teams scored over 200 more runs than they allowed, and their actual cumulative record matched their Pythagorean record. Of particular interest to us, their rank in runs scored in both years matched their finish in the standings, so reflected equally good offenses. The Indians led the AL in runs in 1920, and were 23 runs behind the Yankees in 1921. 1922 was a pedestrian year for the Indians, 78-76, but their fans probably had reason to think it was a fluke.

Sewell played in an era of good offense where players without home run power could excel, and he played for a top-hitting team. But hitting cleanup for over half the season for the 1923 Indians, conditions may have been even more ripe for 100 RBI than this alignment of the stars suggests. Sewell himself reached base 300 times on the dot, but two teammates who hit in front of him, leadoff man Charlie Jamieson and Speaker, both reached base 300+ times as well. Logically, the on-base ability of those in front of him was particularly important to his RBI opportunities.

Speaker is the all-time leader in doubles, and he hit 59 in 1923, his most ever in a season by a count of 6. He hit .380 on the year, standard for him as his 1920 (.388), 1922 (.378), and 1925 (.389) averages demonstrate. But he missed quite a bit of time in 1922 and 1925, and he wasn’t available to be driven in if he wasn’t in the lineup.

Three hundred times on base for Speaker (315), Jamieson (308), and Sewell (300). There had never even been teammates¹² who had reached base 300 times in a season before, let alone three guys. The Indians’ feat was matched, however, by the 1930 Cubs, with Woody English reaching base 320 times, Hack Wilson 314 and Kiki Cuyler 310.

It certainly wasn’t the case that everyone was on 300 times a year in 1923; only one non-Indian did it in MLB. If you think about it, you know that had to have been Ruth (379, a MLB record that still stands, surpassing Billy Hamilton’s 362 from 1894).¹³

Ascertaining the batting order positions of Indians in this day does not require a lot of sifting, and conveying them is easy. Jamieson hit leadoff in all 152 of his games in 1923. In the interval between Guisto’s departure and Sewell’s insertion into the cleanup spot, Speaker hit fourth three times. But otherwise, he always hit third. Of course, my narrative does revolve around Speaker’s eventually changing the lineup. But after he did it, he stuck with it for the rest of the season. Maybe that wouldn’t have been the case if Sewell’s performance had left something to be desired, but one gets the feeling that Speaker didn’t make changes to the order unless the performance of his players made inaction seem negligent.

The Indians’ decline in runs from 1923 to 1924¹⁴ looks more stark to me when it is represented as a total, as a 133-run drop-off. Since I am focusing on Sewell’s raw RBI, in some contexts it’s irrelevant, but this was even more notable since the average team in the AL scored 31 more runs than in 1923. So, how did the decline come about?

The downturn is curious because the nucleus of the two teams was very similar, and you might even drop the nucleus qualifier and extend it to their whole cellular makeup. Browsing individual statistics, I couldn’t see right off that the team had really hit much worse in ‘24 than ‘23. Although the team’s moment was in fact over (they would not win 90 games again until their World Series championship in 1948), one gets the sense they weren’t as far away from ‘23 as the runs scored total indicates.

A couple of things went very right for the Tribe in ‘24 that did not in ‘23. The ‘23 catching situation was a split between Steve O’Neill (108 starts, 80 complete games) and Glenn Myatt (45 starts, 41 complete games). Before the season, the Indians engineered a big trade with the Red Sox, which was responsible for what little turnover in the lineup there was, and O’Neill was included in that. Myatt, who turned 27 midway through 1924, therefore saw his workload about double, and Joe Sewell’s brother Luke was made his understudy. Roxy Walters, surely one of the game’s worst hitters among position players,¹⁵ came over in the trade but got only 87 plate appearances.

O’Neill¹⁶ had been a venerable Indian and hit .317 as a regular with the team from ‘20-’22, but in ‘23 he got old, walking a ton but hitting .248 with just 12 extra-base hits.

Sewell brought some of those same walks that O’Neill had. Not hitting a home run until his 468th career game,¹⁷ he would never be confused for a power hitter, but in a race reminiscent of the tail-end of a charity 5K, also bettered the 1923 O’Neill by 73 points in slugging average and 43 points in batting average.

But the main improvement came from Myatt himself. He’d been solid in ‘23, with a .751 OPS that translated to a 97 index, but he improved his mark to .919 in ‘24, largely on the strength of a .342 average.

Reading between the lines, since 1921, the Indians had had a born designated hitter in the name of Riggs Stephenson. Designated hitter not being an option, the Indians experimented with him at third base in 1922, but mostly had him play second or sit on the bench, his bad defense seemingly more tolerable in small doses. But on the hitting side, in Stephenson’s 12 seasons when he had at least 100 plate appearances, his worst OPS+ was 110. The man could fall out of bed and hit.

But while all candy is sweet,¹⁸ not all of it is equally sweet, and the Indians happened to get the very best version of Stephenson in 1924. He increased his 1923 average of .319 to .371 (Ruth, a qualifier unlike Stephenson, hit .378 to lead the league), and he increased his OBP from .357 to .439.

Stephenson was one who got away, and the Indians never reaped the full benefit of his ability. Believe it or not, after hitting like he did in 1924, he spent most of 1925 in the minors, and they traded him to the Cubs in 1926 for cash and two players who would never play for the Indians or in the majors again at all. The Cubs and Joe McCarthy figured out Stephenson should be in the outfield. A regular for most of 1927-1930, he kept doing his thing at the plate, hitting .347 with a .422 OBP. He led the NL in doubles in 1927, and posted a 1.006 OPS in 1929.¹⁹

We may take the liberty of calling the Stephenson situation a second base platoon, even though Riggs hit righties as proficiently as lefties in 1923 and 1924, and even though his platoon mates were also right-handed. The change to the combination in 1924, a product of the Red Sox trade, was most decidedly a downgrade — “unassisted triple play in the World Series” man Bill Wambsganss leaving, and Chick Fewster coming. Offensively, there was generally almost no difference between the two, with Fewster in fact perhaps better (holding a career .672 to .655 edge in OPS). But Wambsganss, who played for a good while if he did nothing else, had one of his best seasons in ‘23, with a .753 OPS. In ‘24 for the Tribe, Fewster was at .641.

Frank Brower was benched in ‘24,²⁰ presumably to make room for the Red Sox acquisition, George Burns²¹. Burns wasn’t bad, but this was a swap of a plus walker and home run hitter for a plus singles and doubles hitter. Unless the acquired player has a big edge in batting average, that is not a trade you want to make. On the strength of 37 doubles in 129 games, Burns had a 106 OPS+. While that doesn’t seem to cut it for a starting baseman, helped a bit by his defense, he did have 2.4 bWAR. He hit 64 doubles two years later, and as that is tied for the second-highest single-season total ever, he is a part of the record book.

The other five regulars were the same in both 1923 and 1924: shortstop Sewell; Lutzke; and from left to right in the outfield, Jamieson, Speaker and Homer Summa.

Sewell we have covered.

Lutzke tugs at your heart, because he died at 40, with no great detail in his SABR bio about the cause of his death (although it sounds like it was from natural causes). He certainly wasn’t good in ‘24, a 65 OPS+, but relative to his ‘23 performance, he was fine, and probably thus didn’t alter the fortunes of the team either way.

Over his career, Summa gave fairly consistent performance. He was an average-or-above average hitter (.302 career) and a below-average “secondary average” hitter (in his five years of 100+ games, his average league rank in secondary average was 51st, in an average qualifier field size of 60). But just as the Indians got the best version of Stephenson at the plate in ‘24 and a bit of a subpar version in ‘23, succumbing to the dreaded sophomore slump, Summa was down quite a bit in ‘24. His average slipped from .328 to .290, and his walk percentage from 5.6 to 2.7

By no means was Speaker in the beginning of a sad “Willie Mays with the Mets” phase in 1924. Indeed, in ‘25, he would lead the AL in on-base percentage (.479), and narrowly miss out on the batting title, too (.389, to Harry Heilmann’s .393). But his 74 Baseball Reference Batting Runs in 1923 rates second-highest ever for a 35-year-old²², and in 1924 he fulfilled the near certainly of not being as good. He lost 37 points off his OBP, 100 points off his slugging average, and missed 15 more games. He wasn’t 74 “runs above average” at the plate, but +34 places him in a tie for 20th all-time among 36-year-olds. But, in assessing the difference between the Indians’ 1924 and 1923 offenses, like with Lutzke, he inevitably must be compared to himself.

Jamieson might be said to represent the larger paradox of individual achievement amidst team disappointment in 1924, as he hit .359, was second in the batting race and third in the MVP. He did start 13 fewer games in 1924 than in 1923, but with an OPS of .865 compared to .869 the year before, he can’t honestly be said to have been been intregral in the team’s decline. But he did draw just 47 walks compared to the 80 he had the year before. He reached base 262 times, not 308. Reaching base quite a bit less often for a much less potent offensive team, his runs scored predictably plummeted from 130 to 98. MVP balloting is a result relative to the competition, and the “no previous winners” rule meant Ruth was out of the equation. But if the perception was that Jamieson had had a better season in 1924 than 1923, even from a superficial statistical perspective, that was a case you could only make citing batting average.

I’m not sure either my larger project here or his attainments justify it, but I took a close look at Jamieson’s career. With just a couple of seasons in the American League top 10 and just 131 stolen bases for his career,²³ Jamieson probably had no more than decent speed, but the leadoff hitter framework is a good one for understanding him. He hit for average and he chipped in some walks. You had to live with a lack of power; from my analyst’s perspective, just how acute that was, and just how good his offsetting on-base percentage, determined if it represented too much of a sacrifice.

Both he and Summa were Indians of the 1920s for all intents and purposes²⁴, allowing their Indians numbers to be compared without taking heed of the live ball/dead ball distinction. As much of a singles hitter as Summa was, he still beats Jamieson in Isolated Power for their Cleveland careers, .096 to .090.

Jamieson only hit as many as 4 home runs once, and never finished better in AL doubles than tied for 10th. But he led the AL in singles in 1923 and 1924.

His early major league career gave no signs of stardom. A bench player for the Nationals in 1916 and the beginning of 1917, he moved to the very much rebuilding Athletics midseason.²⁵ Hitting .267, 18 points above league average, he was able to post a 97 OPS+, but this was done in the context of no home runs and just 8 extra-base hits in 345 at-bats.

In 1918, even his average abandoned him; it slipped to .202, and about all he offered was his team-best 54 walks.

Skipping ahead, we find him on the 1920 Indians, and I suppose the rest was history. He hit .319 that season (36 points over the league average) and had a 108 OPS+ for the team that won the championship. This first real success came as a platoon player; he had just 15% of his plate appearances against lefties, and was not in the starting lineup for games 1, 3, and 6 of the World Series.²⁶

Joe Evans was the guy who got the starts in the World Series versus lefties in Jamieson’s place. He came up with the Indians at age 20 in 1915, and had a 52 OPS+ as the team’s starting third baseman in 1917, but first-year manager Speaker carved out a precise role for him starting in the championship season, with the Mississippian getting 189 PA in 1920, 174 in 1921, and 164 in 1922. His career success was short-lived, but he caught lightening in a bottle in 1920, hitting .349 with 9 triples.

Much of that may have come about because he was platooned in an uncompromising manner. His averages were .368 versus lefties, .289 against righties. From those, you can see that his “against left-handers” mark more tracked with his overall average, and he had only 53 PA versus righties. This was a guy (yes, back in 1920) who absolutely could not hit the curveball. He had a .218 average and a .270 SLG for his career versus righties, a .308 average and a .398 SLG for his career versus lefties. The 39.5% of plate appearances versus lefties he had in his career would presumably rate as very high for the era. Against righties, his BB/SO ratio was 99/118; against lefties, 97/29.²⁷

This discovery of Evans’ role and of the size of his platoon split establishes just how far back the platoon focus goes. Additionally, the data we now have on Evans shows that, although he was working without statistics, Speaker assessed Evans correctly.²⁸ Even if he was hidebound in his deployment of the batting order, here he distinguished himself as a manager.

Speaker apparently kept the platoon in place while Evans remained on the team in 1921 and 1922. Hitting leadoff in all but one of his starts, Jamieson’s plate appearances rose to over 600 in those years, but he began games on the bench some 30 times in 1921 and some 20 times in 1922. Meanwhile, Evans started over 30 games in left field both years.²⁹ Not until Evans was traded to Washington for Brower before the next year was Jamieson presumably able to come to the ballpark without having to make sure his name was in the lineup.

Filling in the remaining blanks for his career, he hit .310 in 1921 and .323 in 1922, so produced at about the same level he had in 1920. He fell back to .296 in 1925. Never able to quite recapture even his early Indians form, he didn’t regress either, and kept his average in the .291 to .309 range over the next five seasons.³⁰ With the arguable exception of 1926,³¹ the Indians never contended, but there is something to be said for a player who can stay in the lineup for that length of time, presuming a club is trying to be competitive. He did see his plate appearances decrease in each of his last seven seasons with the club (in each of his last nine seasons, I suppose, if one counts 67 plate appearances over 1931 and 1932).

Instead of taking individual players in trying to understand the decline from 1923, we could take individual categories for the team as a whole. The most glaring place the Indians were off was in walks, where they went from 626 in 1923 (leading the league) to 491 in 1924 (a bit below average).

With categories the framework, we again find that the ‘24 Indians did some things more or less on the level that they did them in 1923, at least if one doesn’t adjust for the somewhat better hitting in the league overall. First, their batting average was only down 3 points in 1924, from .301 to .298. Then while Speaker’s doubles tumbled from 59 to 36, the team increased its doubles from 301 to 309.

Home runs wouldn’t have seemed to count for too much, since most games were played without them, but somehow the Indians slugging percentage was down 19 points, considerably more than their batting average. The team was off 18 home runs (from 59 to 41) and 15 triples (from 75 to 60).

Adjusting for 52 more at-bats in 1924, we can make those differences 18.6 and 15.7. Converting the counts to total bases, they become 74.4 and 47.1. 121.5 total bases (74.4 + 47.1) over the 5329 at-bats of 1924 comes to 23 slugging average points. For comparison, the ‘24 Indians were behind the ‘23 Indians 19 points in on-base percentage.

After going this, I wasn’t sure if seemingly small differences in extra-base hits are more important than I would have guessed, or if maybe 20 OBP and SA points (and hence 40 OPS points) doesn’t represent as significant a difference as I would have thought. For context, taking the 16 American League teams over the two years, the 1923 Indians ranked 1st in OBP, while the 1924 Indians were tied for 5th. The 1923 Indians were 3rd in SA, and the 1924 Indians 6th. So the losses in triples and home runs did drop them down quite a bit.

It doesn’t rate with the 2024 Padres versus the 2023 Padres,³² but that decline of 135 walks for the Indians in the corresponding seasons 100 years before had me searching for the right strong adjective to describe it. You could already piece the causes together pretty well just from my previous narrative, but perhaps three positions best explain the change.

First, both Sewell and Jamieson had what would prove to be “career walk years”³³ in 1923. With his seasons differing by just one plate appearance, Sewell had 31 fewer walks in 1924; Jamieson fell off by 33 walks, which we can cut to 27 if we prorate to his 1923 plate appearances.

Sewell’s 1924 9.8% walk rate wasn’t low compared to what he normally did, particularly in terms of his future seasons; once he assumed those historic low strikeout rates of around half-a-dozen a year, he also walked a little less, perhaps taking fewer strikes. From 1925 on, he had a 9.3% walk rate.

On the other hand, although Jamieson had a career high number of walks in 1923, that was really fueled by his 746 plate appearances. Jamieson’s walk leakage did reflect a low anomaly for him in 1924: of his 13 seasons with 100 or more games, the 7.1% walks he had in ‘24 was the worst of his 13 seasons, and his walk percentage actually inched up beyond his 10.7% 1923 percentage for much of the second half of the 1920s.

The third position I’ll highlight, and the one involving a change in personnel, was Brower essentially being replaced at first by Burns. In 1924, the Indians got 27 fewer walks from first base.

This is not a frivolous piece, but it should not be mistaken for a quantitative endeavor. That kind of demonstration will have to wait. I am working on one of the longest things I have ever done for my Substack blog, and it is a quantitative analysis of the meaning of RBI. Instead, my mindset has been to show how you could make an exact account of Sewell’s RBI in the two years if you wanted to, and where it looks like that might lead. But in the spirit of being as conceptually accurate as I can be, I must acknowledge that in the process of writing this piece up, I came to realize that my initial analysis contained some erroneous thinking, or at least some overstatement.

My argument began with the observation that almost all of Sewell’s RBI came with runners in scoring position. Since basically any hit with a runner in scoring position will score the runner, I then concluded that batting average with RISP could be used as a proxy for RBI production and was really all that we needed. So with Sewell hitting .381 with RISP in 1923 and .283 in 1924, the close counts in his actual RBI indicated that his personal performance mattered less than one would think, and what then mattered was RISP opportunities. Although the difference in his RISP plate appearances in the two seasons was just 10, directionally it did favor 1924, consistent with a big role for RBI opportunity, an even bigger role than my intuition suggested.

But in finalizing the piece, it dawned on me that the fact that almost all of Sewell’s RBI came in his at-bats with RISP did not mean that almost all of the runners he drove in were actually in scoring position; no, he didn’t hit home runs, but what about his success getting in runners from first base via doubles and triples, in a) first and second b) first and third and c) bases loaded situations?

Retrieving this information, here you do find a real advantage for Sewell 1924 over 1923. In terms of RISP batting average, he was less clutch in 1924. But in 1923, over all situations with a man on first base, he hit a double or triple in 12 of 256 opportunities (4.7%). By contrast, when there wasn’t a runner on first, he hit a double or triple in 39 of 420 opportunities (9.3%), almost twice the rate of when those doubles or triples had incremental RBI value.

1924 was the opposite case. Sewell had 259 chances with a runner on 1st, and hit a double or triple 25 times (9.7%). By comparison, in what we might call low-leverage double/triple scenarios, he mismatched only 23 times out of 413 plate appearances (5.7%).³⁴

Pending a quantitative analysis, I would still consider this a relatively small factor in his RBI counts. For one thing, approximately half of the doubles with a runner on first only advance that runner to third and do not score him. But understanding that hitting with runners in scoring position does not always mean hitting only with runners in scoring position is a real reframing, and it does help close the gap between Sewell’s RBI performance in the two seasons.

My other initial misrepresentation was that I referred to plate appearances with RISP and not at-bats. Plate appearances are absolutely the right way to look at things if you are trying to document the raw opportunity Sewell had in the two years, and that was a question that interested us in light of the unlikely finding that Sewell batted more often with RISP in the year that he was on the much worse offensive team. But in trying to figure out how Sewell’s RBI advantage in 1923 didn’t amount to more, the right framework isn’t plate appearances, it’s at-bats. Unless the bases are loaded, from an RBI perspective, a walk is a wasted plate appearance. Joe walked a good deal more often with RISP in 1923 (41 times) than he did in 1924 (30), consistent with the overall difference in his walks. So his AB in the two years came out to 189 and 212, respectively.

If you insist on using plate appearances and not at-bats, then not counting walks in calculating the RISP average is really an error. If one figures Sewell’s RISP averages as a function of plate appearances and not at-bats, the difference in them decreases from 98 points to 57 points (.288 and .231, respectively). So trying to square his ultimate RBI with his batting averages and plate appearances is applying data that isn’t aligned.

Appreciating that walks are an albatross on RBI means that we can amplify our complaints about their unfairness to incorporate their de facto punishment of a good deed, and take a break from our usual fury at the uneven playing field we inevitably find when we look at opportunity.

1. I fully own that a great RBI total without commensurate home runs could be seen as a curiosity as much as an accomplishment, although the factors behind the player’s RBI total are paramount here.

2. 3 and 103 for Sheely; 6 and 99 for Falk.

3. This included all 11 of his HR in 1932 coming at home. His slugging average that year was .516 at home, .296 on the road. He did have just 2 strikeouts in 321 road plate appearances, though. He had 3 strikeouts on the year, and it was one of five seasons he had as a batting average qualifier when he had just 3 or 4 strikeouts.

4. Bested by Babe Ruth here, who hit .431. This was a strong year for average hitting, with Harry Heilmann breaking through for a .400 season, and Ruth hitting .393.

5. Whether because of a missing game or two or mistakes in the official record, Retrosheet begins with Sewell 5 RBI ahead in 1923 (109 to 104), not 3. That is why the numbers don’t add up (although the discrepancy is trifling, not ominous, as is usually the case when someone uses the latter expression).

6. On the high end, maybe 5 RISP plate appearances from each season are missing. Footnote 34 will be helpful here. I am of course working with and quoting from the data that is available.

7. And hence the title of the piece!

8. This said, he did eventually make the much-needed change.

9. Or Senators.

10. I put this to the test. My conjecture was off from the standpoint that the Brower slash line has never been terribly common, but I was right on a few counts:

(1) that the nature of the three slash numbers together was apparently unusual in his day.

(2) that it was in concert with Moneyball; the 2000s, not the 1990s (or any other decade), are really the decade when this fits.

(3) starting in 2010, with the exception of Carlos Santana, who did it in the aberrant 2019 season, it stopped happening. I say this vindicated me in part because I identified the Brower batting average as generally “too high” for today, and I would take it that is why the combination now isn’t found

(4) I didn’t count ‘em up, but I was right that this is disproportionately a first base thing.

Specifically, here’s what I did. Brower had 479 PA, clearing the 477 that would go with 3.1 per 154 games. Again, he had a .285/.392/.509 line. I looked for players with (a) a minimum of 477 PA (b) a BA from .280-.289 (cc) an OBP from .380-.399 (d) a SA from .480-.519.

Four players have done this twice. Sid Gordon and Brad Hawpe did it in consecutive years, and Jack Clark repeated five years and one team after he did it the first time. Tim Salmon rounds out the quartet.

With a minimum of an .860 OPS, it is a fine slash line; even Brower’s season, achieved under favorable conditions, gave him a 136 OPS+. But Will Clark’s 1988 (.386/.508) translates to a 160 OPS+.

The complete list:

Gavy Cravath, 1915; Frank Brower, 1923; Gus Suhr, 1930; Joe Cronin, 1940; Bobby Doerr, 1948; Pat Mullin, 1948; Roy Campanella, 1949; Sid Gordon, 1951; Vic Wertz, 1951; Sid Gordon, 1952; Mickey Mantle, 1959; Ron Santo, 1969; Jim Wynn, 1970; Frank Robinson, 1971; Reggie Jackson, 1974; Ron Cey, 1979; Jack Clark, 1980; Jason Thompson, 1982; Alvin Davis, 1984; Jack Clark, 1985; Will Clark, 1988; Tim Salmon, 1996; J.T. Snow, 1997; Ryan Klesko, 2000; Tim Salmon, 2002; Pat Burrell, 2005; J.D. Drew, 2006; Aramis Ramirez, 2008; Brad Hawpe, 2008; Brad Hawpe, 2009; Chase Utley, 2009; Carlos Santana, 2019.

It’s interesting that the hit rate for making the Hall of Fame if you did this was about 50% through Reggie Jackson, but since then we’re 0 for 16. I can believe that it didn’t represent the same level of achievement in the 2000s, but why it didn’t translate to a 1980s Hall of Famer, for instance, I couldn’t tell you.

11. Although I don’t want to make too much of this. A speedster named Merlin Kopp played 202 games in the PCL that year (hey, you play the schedule in front of you!)

12. In what might be considered somewhat recent time, I believe the only two teammate pairs to be on base 300 times apiece were Derek Jeter and Bernie Williams on the ‘99 Yankees and Jeff Bagwell and Craig Biggio on the ‘97 Astros.

13. Ruth also reached base 353 times in 1921, surpassing Ty Cob (336, 1915) for the most of the century.

With 376 times on base in 2004, Barry Bonds came up just short of Ruth’s 1923 record. With the Giants in the thick of a pennant race (they ended up 2 back in the West, and 1 shy of the wild card), Bonds started 27 of the Giants 28 September/October games and had a .619 OBP, his second best of any month that year. So it can’t be said he didn’t give it everything he had.

In looking at the highest “Time On Base” totals, these were maybe fueled more by walks and less by hits than I would expected. There have been 43 seasons where a player has reached base 320 times, but the only .400 seasons were produced by Williams, 1941, and Hamilton, 1894, again. There is also Lefty O’Doul, 1929/.398, but the only other season of .380 or better is Ruth’s 1923. However, Mark McGwire, 1998, is the only player to reach base 320 times and not hit .300 (he reached 320 times exactly and hit .299). The mean batting average of seasons where a player reached 320 times was .356.

Forty of the 43 players walked 100 or more times. Falling short of the century mark in walks but reaching base 320 times were Wade Boggs, 1985, who walked 96 times; and Derek Jeter, 1999, who walked 91. O’Doul, 1929, is an outlier — 76 walks.

The hit leaderboard for the 320+ times-on-base club then makes sense. It is headed by O’Doul with 254, followed by Boggs with 240. Hamilton is a distant third with 225.

Mean walks for players with 320+ times on base — 134.

Mean HBP — 5. (0 HBP and on the list: Ruth, 1927; Gehrig, 1931; Foxx, 1932. And, yes, HBP were kept in these years.)

Mean plate appearances were up there, 694. They ranged from Williams, 1941 (606) to Steroids Lenny Dykstra, 1993 (773).

14. This was mirrored by a 15-game slide in the standings; at 67-86, the ‘24 Indians were bad. They finished sixth, but only bettered the cellar-dwelling White Sox by a game.

15. Included in his impressively bad record, he came to the plate 1606 times in his career and never hit a home run. He was a .222 career hitter but posted consecutive averages of .199, .193, .198, .201 and .194.

The Mendoza line, my ass! It should be the Walters line.

Mendoza does have Walters, though, in terms of the worse OPS+ — 41 to 51.

16. Would go on to a successful career as a manager. Fourteen seasons, and he got a World Series trophy with the 1945 Tigers. Managed the Indians, although not for very long.

That was his first job. He was Bob Feller’s first manager.

17. He hit it to left field against Waite Hoyt in 1928 at Yankee Stadium.

18. You know I am ever careful, so I asked ChatGPT. ChatGPT confirms: “If we’re being careful with definitions, the honest answer is almost no [regarding the existence of non-sweet candy]—but there are a few edge cases that make the question interesting….”

19. Of course, holding a regular spot in the outfield doesn’t prove he played it well. But from his statistics, he does not appear to have been a liability. He committed 23 outfield errors over 1927-1930, and had the best fielding percentage among National League left fielders in 1929. His composite Rfield over the four years was -4.

20. Brower saw action in every month and had a .434 OBP on the year. He was given 41 pinch-hitting tries, but didn’t do particularly well in the role, with just five hits and no home runs. In 19 games as a first baseman, he had a 1.069 OPS.

21. Remarkably, the National League had a George Burns worth taking account of, too, and he was only three years older. The National League George Burns was mostly a New York Giant, an outfielder, who led the league in walks and runs five times. To my surprise, the bWAR of the George Burns is close (favoring the National Leaguer 39.9 to 34.7). One had 2018 career hits, the other 2077.

22. Ruth is first with 96 in his age-35 season, Nap Lajoie third with 65.

23. That ranked him 12th in stolen bases among active MLBers at the time of his retirement, behind two guys I wouldn’t have thought of as base stealers, Rogers Hornsby and Goose Goslin.

24. Jamieson was with Cleveland from 1919 to 1932 and took 92% of his plate appearances with them in the ‘20s; Summa’s Cleveland career was 1922-1928.

25. Baseball Reference (courtesy of Retrosheet) says the Athletics got him by waiver, confirming the “down and out” start he had.

26. A left-handed, technically platoon player in that era might generally have had little to worry about and gotten the lion’s share of the action, but the Dodgers started lefties Sherry Smith twice and Rube Marquard once. I suppose this may have been strategic, as Smith, Marquard and lefty Clarence Mitchell only accounted for 30% of the Dodgers starts during the regular season. And with Sewell in for the deceased Ray Chapman, six of the Indians’ “Baseball Reference box” regulars were left-handed.

Another addendum is that Smith would be Jamieson’s teammate from 1922-1927. Smith started 101 games for the Indians and had 13.0 bWAR.

27. It is true that intentional walks tend to only be issued to players who have the platoon advantage. I’d be very surprised if they played any real role in this split of Evans’ (he wasn’t a prominent enough player), but they were only kept for two of his 11 seasons (and after he left Cleveland), so we can’t know.

28. But Speaker platooned Evans so strictly that the improvement in his overall hitting statistics had to at least establish the correlation. Speaker was presumably using statistics to evaluate the platoon question with Evans, but he was looking at his 1920-1922 seasons versus the earlier ones, rather than his sample versus right- and left-handed pitchers.

29. Jamieson had 36 appearances in center field as well, so the numbers don’t quite add up.

30. I went with average, because as Jamieson’s average went, so did his OPS.

31. Finishing second at 88-66, they made a late run at the Yankees, but never really threatened. They were ten games back after losing to the Bombers on August 23, just two back after game 150 on 9/22, and three back at the wire.

32. Through my Substack Notes, you could say this is something I “covered” in real time, and the Padres fell from 653 walks to 458. So they started quite a bit ahead of where the Indians were in 1923 and finished quite a bit behind their 1924 total. I suppose these things are always idiosyncratic, to a degree reflecting players who inexplicably walk less in the second season, but they lost Juan Soto and his 132 walks. Then they added Luis Arraez (16 BB in 524 PA in 2024) and Jackson Merrill (29 BB in 593 PA), a case of subtraction by addition. The funny thing is the Padres in 2024 had a better year, winning 93 games when they’d won only 82 the year before, and scoring 8 more runs than they had in 2023, too.

33. Today this certainly has a different meaning!

34. As implied in earlier footnotes, Retrosheet data from this era typically has minor discrepancies from the official record, as well as the very occasional missing game. Without a very close look, any discrepancies are shrouded in the larger number of missing games. Obviously, the breakdown of the situations I give here must necessarily reflect the Retrosheet totals.

Anyway, Retrosheet has 676 PA for Sewell in 1923, versus his official 685. The two records both have him with 51 2B+3B.

For 1924, Retrosheet has 672 PA for Sewell versus his official 686. Retrosheet has 48 2B+3B for him, versus his official count of 50.