Lopsided batter/pitcher match-upsby Chad Evely
April 27, 2012
Despite being a team game, at its core baseball consists of match-ups between two individuals: the batter and the pitcher. Sadly, as years pass, the results of these individual match-ups fade to splits and are ultimately lost to the ether of season totals. But what about those instances where one player thoroughly dominated another over the course of their careers? This article will establish a method to define a "lopsided" match-up and delve into the most extreme examples.
For the purposes of this study, we’re interested not in those cases where a pitcher dominated a poor batter but where a pitcher dominated a good/great batter. That Hall of Famers feature prominently in the final list seems to indicate that we were successful. I’ll be brief in my description of the methodology, knowing that math-in-prose is not exactly the most popular writing form. If anyone is interested in the nitty-gritty, feel free to ask in the comments section.
After an explanation of the methodology, we'll look at the most lopsided match-ups of the late 1950s and 1960s. I've got three more articles lined up for more recent times, so stay tuned.
At the core of the comparison, I needed a single number to describe offensive performance. I wanted it to be cumulative stat rather than a rate in order to pick out those match-ups with a large sample size rather than limited at-bat flukes. I settled on the 2002 version of Runs Created (hereafter referred to as "RC"), but removed the baserunning components to focus on the batter/pitcher match-up.
After some thought, I also removed the penalty for grounding into a double play; ultimately it didn’t seem fair to penalize the batter for grounding out with a runner on first if I wasn’t going to reward him for getting an RBI with a runner in scoring position. The RC stat gives a good approximation of the value of a plate appearance (hereafter referred to as "PA") on a scale that’s easy to understand and the 2002 version has the added benefit of approximating actual runs created in very small sample sizes, allowing us to view a fairly accurate value for a single PA.
For some context: in 2011, the average team scored 4.28 runs per game, while the total RC tallied was 4.32. Here are a couple of RC values scaled to an average season to give an idea of the range of values we’re talking about:
Now that we’ve decided on a statistic, we can start having some fun. To determine lopsided match-ups, we must first know how we expect the batter to fare. If Jeff Mathis hits like Jeff Mathis against a particular pitcher, it’s to be expected. If new teammate Jose Bautista hits like Jeff Mathis against a pitcher, we’re on to something. To establish this baseline of expected performance, only those seasons where a batter has at least 200 plate appearances are taken into account.
The batter’s season performance is then separated into two groups: all PAs against the pitcher in question and all other PAs. The RC against all other pitchers is divided by total PAs to determine how many runs we’re expecting that batter to create in a single plate appearance. This single PA expectation is then multiplied by the number of PAs against the pitcher in question to determine the total expected RC against that pitcher if the batter were to perform at the same level as his season averages.
The difference between the expected RC and the actual RC against the pitcher tells us whether the batter under- or over-performed over the course of a season and by how much. Once this difference is determined for each season these players faced each other, it is added across their careers to come up with the total difference in expected and actual RC in the match-up. In most cases, a large difference in a single season will even out over the course of a career, but those instances of sustained lopsidedness indicate the truly unbalanced match-ups we’re seeking.
This methodology was applied to all batter/pitcher match-ups between 1956 and 2011. I settled on 1956 as a starting point since it’s the first season where Retrosheet has play-by-play data for at least 95 percent of games played, which seemed like a reasonable threshold. I toyed with the idea of introducing a scaling factor to bring each year into a common scoring environment but ultimately decided the raw numbers better captured the spirit of the study.
Enough with the boring stuff and onto the results. For each match-up, I have given the actual stats against the pitcher as well as the "expected" stats if the batter were to have performed against that pitcher the same way he did against the rest of the competition he faced in the seasons in question. While far from a baseball historian, I have done a little research into each match-up to bring a little life to the numbers.
Over time, the average number of plate appearances between a particular batter and a particular pitcher has decreased, making our method inherently biased toward older match-ups, particularly the '50s and '60s. I would expect that the biggest culprits are expansion, scheduling changes, fifth starters, increased player movement and increased reliever usage. Since more plate appearances mean a higher magnitude of dominance for those older match-ups, rather than presenting a straight Top 10 list the results have been divided to show the top five match-ups in each decade.
For each match-up, I have listed the overall ranking of lopsidedness to give you an idea of where it stands all-time. Match-ups are placed in a decade according to the median year of their career match-ups. So if they first faced off in 1973 and last faced off in 1984, the median year is [(1973+1984) / 2] = 1978.5 and their match-up is placed in the '70s. To give an idea of the aforementioned era bias towards older match-ups, eight of the top nine match-ups come from the 1960s and every single match-up from 1-27 is from the '60s or '70s.
I have included a graph of each match-up, an explanation of which can be found in the comments for the No. 1 match-up (Eddie Mathews vs. Don Drysdale). I have also listed which additional pitchers a batter "struggled" against and conversely which additional batters a pitcher "dominated," showing any additional match-ups involving these players that appear in the top 500 all-time.
Let me know if you like this idea. I had originally planned to include the match-ups from the pitchers' perspective—those cases where a good pitcher struggled against a particular batter—but decided the article was long enough as is. I may still do this in the future if this idea seems as interesting to others as it has been to me.
Following are the top five lopsided match-ups between 1956 and 1960:
No. 1. Eddie Mathews vs. Don Drysdale (18.4 RC, 229 PAs, 1956-1967) View Matchup
Actual: 20.0 RC | 40-206 | .194/.275/.335 | 0.610 OPS | 32 SO, 21 BB, 2 HBP
Expected: 38.4 RC | 53-194 | .275/.378/.506 | 0.884 OPS | 33 SO, 32 BB, 1 HBP
On June 13, 1957 in the top of the 2nd, a brawl erupted after Drysdale hit Mathews’ teammate Johnny Logan with a pitch that led to both players’ ejections. Clicking the second link will take you to a Life article of the altercation, complete with a photograph of Mathews clocking Drysdale with a leaping right hand. The next season, on June 26, 1958, Drysdale beaned Mathews in the first inning. Up until this point, Mathews was dominating the match-up to the tune of 7-22 (.318/.400/.591) with two home runs. Whether the beanball allowed Drysdale to gain the upper hand or not, the rest of their career match-up belonged to Drysdale (33-184, .179/.256/.304).
The following graph gives a detailed account of the career match-up between Mathews and Drysdale. It contains a whole lot of information so it might look somewhat intimidating at first, but here is a brief explanation:
- The red line indicates the cumulative Runs Created (plotted on left-hand scale) that we would have expected Mathews to have produced as their career match-up progresses.
- The black line indicates the cumulative Runs Created that Mathews actually produced, plotted on the left-hand scale.
- The bars indicate the result of every plate appearance between these two players. The height corresponds to the Runs Created by that PA, plotted on the right-hand scale. A key at the top shows what each color means. For example: all green bars are hits, with the shortest being singles and the highest HRs.
Analyzing their match-up, we can see the aforementioned HBP indicated by the dark blue bar early in 1958. At that point, the black line is slightly above the red, showing us that Mathews had been producing better than expected up until that point in their match-up. Mathews’ production then leveled off and he couldn’t keep up with the slow, steady train of expectation.
Eddie Mathews also struggled against Mike McCormick (9.8 RC), Turk Farrell (7.9 RC), Johnny Klippstein (7.5 RC), Johnny Antonelli (7.1 RC) and Harvey Haddix (6.6 RC).
Don Drysdale also dominated Orlando Cepeda (15.1 RC), Dick Allen (11.5 RC), Joe Adcock (9.4 RC), Mike Shannon (7.6 RC), Felipe Alou (7.9 RC), Frank Robinson (7.6 RC) and Tom Haller (7.2 RC).
No. 2. Willie Davis vs. Juan Marichal (16.2 RC, 196 PAs, 1961-1973) View Matchup
Actual: 9.1 RC | 33-189 | .175/.185/.296 | 0.481 OPS | 23 SO, 2 BB, 1 HBP
Expected: 25.3 RC | 52-182 | .283/.318/.417 | 0.735 OPS | 19 SO, 9 BB, 1 HBP
Marichal hit Willie Davis with only one pitch during their long, intertwined careers but it was a big one, breaking Davis’ jaw in the third inning of a game on July 19, 1969. They had already matched up over 100 times at that point, so it’s hard to see a correlation to Davis’ struggles against Marichal but it is certainly interesting that two famous HBPs come up in our first two match-ups. Davis showed good power in this match-up but hit for a terrible average with very few walks.
Willie Davis also struggled against NONE.
Marichal also dominated Willie Stargell (12.5 RC), Johnny Roseboro (10.7 RC), Tony Taylor (10.6 RC), Ron Fairly (9.0 RC), Ernie Banks (7.5 RC), Gordy Coleman (7.1 RC), Johnny Edwards (7.0 RC), Tony Perez (7.0 RC), Rusty Staub (6.7 RC) and Sonny Jackson (6.5 RC).
No. 3. Charlie Neal vs. Bob Friend (15.5 RC, 103 PAs, 1957-1963) View Matchup
Actual: -1.9 RC | 9-95 | .095/.139/.116 | 0.254 OPS | 19 SO, 4 BB, 1 HBP
Expected: 13.6 RC | 24-90 | .265/.337/.413 | 0.750 OPS | 15 SO, 10 BB, 1 HBP
This match-up has less than half the career PAs of Mathews/Drysdale and just over half the career PAs of Davis/Marichal. It shows up this high in our rankings purely on the magnitude of awfulness that was Charlie Neal’s performance against Bob Friend. Take a look at those numbers; they’re wondrous.
Charlie Neal also struggled against Don Elston (6.9 RC).
Bob Friend also dominated NONE.
No. 4. Orlando Cepeda vs. Don Drysdale (15.1 RC, 164 PAs, 1958-1968) View Matchup
Actual: 12.8 RC | 35-154 | .227/.262/.305 | 0.567 OPS | 23 SO, 4 BB, 4 HBP
Expected: 28.0 RC | 47-150 | .310/.358/.528 | 0.886 OPS | 23 SO, 10 BB, 2 HBP
In his very first game as a big leaguer, Orlando Cepeda went 0-2 against Drysdale. However, once Drysdale was replaced in the fourth by Don Bessent, Cepeda homered in what was only his third major league at-bat. The Giants were on their way to an 8-0 victory and Cepeda was on his way to a Hall of Fame career. He did, however, never fare much better against Drysdale, who appears once again as the primary nemesis for a batter.
Drysdale appears four times in the all-time top 75 of this list (Dick Allen and Joe Adcock being the other tortured batters); surely leading the league in HBP in five different seasons must have played some part in getting into the heads of specific batters. Orlando Cepeda seems to have agreed, once saying "The trick against Drysdale is to hit him before he hits you."
Orlando Cepeda also struggled against Don Cardwell (8.4 RC)
Don Drysdale also dominated Eddie Mathews (18.4 RC), Dick Allen (11.5 RC), Joe Adcock (9.4 RC), Mike Shannon (7.6 RC), Felipe Alou (7.9 RC), Frank Robinson (7.6 RC) and Tom Haller (7.2 RC)
No. 6. Hank Aaron vs. Bob Gibson (13.9 RC, 180 PAs, 1959-1974) View Matchup
Actual: 19.5 RC | 35-163 | .215/.278/.423 | 0.701 OPS | 32 SO, 15 BB, 0 HBP
Expected: 33.4 RC | 48-159 | .305/.378/.569 | 0.946 OPS | 19 SO, 19 BB, 0 HBP
Hank Aaron wasn’t particularly terrible in this match-up, but he suffers from years of establishing high expectations. Although he was never hit by a Bob Gibson pitch Aaron still seems to have thought about it, once giving the following advice to a young Dusty Baker:
"Don't dig in against Bob Gibson; he'll knock you down. He'd knock down his own grandmother if she dared to challenge him. Don't stare at him, don't smile at him, don't talk to him. He doesn't like it. If you happen to hit a home run, don't run too slow, don't run too fast. If you happen to want to celebrate, get in the tunnel first. And if he hits you, don't charge the mound, because he's a Gold Glove boxer."
Hank Aaron also struggled against Jack Sanford (9.9 RC), Jim Brosnan (9.2), Glen Hobbie (8.8 RC), Turk Farrell (7.5 RC), Don Sutton (7.4 RC), Tom Seaver (6.9 RC) and Bob Bruce (6.8 RC).
Bob Gibson also dominated Roberto Clemente (12.9 RC), Willie Mays (9.8 RC), Tony Perez (9.6 RC), Ron Santo (8.7 RC), Bill Mazeroski (8.6 RC), Rusty Staub (7.9 RC), Jim Hickman (7.6 RC), Tony Taylor (7.4 RC), Maury Wills (7.1 RC), Glenn Beckert (7.0 RC), Jerry Grote (6.9 RC) and Frank Robinson (6.3 RC).
References and Resources
The information used here was obtained free of charge from and is copyrighted by Retrosheet. Interested parties may contact Retrosheet at "www.retrosheet.org"
Chad is a geospatial application developer/mediocre slow-pitch softball pitcher living in Lexington, KY with his wife and 1-year-old daughter. You can find his semi-regular thoughts on baseball at Statistical Meanderings or email him (chad AT statisticalmeanderings DOT com).