Is There More to the Marlins’ Dominance in Close Games Than Mere Chance?

The Marlins began this season on an historic pace in one-run games. They won every single one of their first 12 such contests, besting a record that had stood for 51 years. Since then, they’ve played to a more modest 15-11 mark in those bouts, but their .711 winning percentage in such games on the whole would still tie them for the fifth-best in a single season since the Live Ball Era began in 1920 (min. 20 one-run games).

Last Thursday, my colleague Michael Baumann wrote a piece that got me thinking. Specifically, he found that of the three teams outperforming their Pythagorean (run differential-based) record by at least five games at the time, two — the Orioles and Brewers — had outstanding bullpens in one form or another. This idea isn’t new — it’s been hashed, and re-hashed, and re-hashed again. The Tigers, which have since joined that group of five-game overperformers, have also had a remarkably clutch relief corps. But the Marlins, outperforming their expected wins by six games, have a middling ‘pen by any measure. Marlins position players have come through in big moments more than expected, but they also haven’t wowed in those situations to the same extent that the Orioles’ crop of hitters have.

Instead, the Fish stand out the most in terms of their prowess in one-run games. Historically, that ability hasn’t been enough to consistently overperform Pythagorean records. I recreated Michael’s graph to that effect with an additional 82 seasons, including this one up through Saturday’s games (data via Baseball Reference):

In the entire dataset, the r-squared was about .125; in other words, over/underperformance in one-run games explained only 12.5% of the variance in Pythagorean over/underperformance. Given the breadth of the dataset and the prominence of relievers in one-run games, I split the sample in two beginning with the 1970 season, which is around when relievers began to play an outsize role in the late innings (think Goose Gossage, Rollie Fingers, and Bruce Sutter). My reasoning was, maybe performance in one-run games would be a more meaningful indicator if it was dependent on something like relief prowess, which could carry over into other types of close games — leading to more overperformance elsewhere. This wasn’t borne out here.

Still, despite the randomness inherent in one-run games, relief ERA- has been a slightly better predictor of one-run winning percentage than starting ERA- since 1970, while the impact of starters has dwarfed that of relievers in terms of overall winning percentage:

So, relievers do seem to shine in the late innings of close games, but that still doesn’t explain why the Marlins have been so good in them. I mentioned that their hitters have been clutch; one reason for that is that they’ve been particularly successful at bunting in high-leverage situations. It’s a small sample, but they’ve laid down sacrifices on nine of 10 attempted bunts here. In medium-leverage scenarios, they’ve failed five times out of 20 but have notched seven singles and eight sacrifices.

In these two scenarios combined, they’ve only failed to move the runner over or reach base six times; their failure rate is tied for the eighth-lowest among the 28 teams with at least 10 bunts attempted in these situations. Bill James found that teams adept at small-ball indeed had an advantage in one-run games, but that was also before bunts became so scarce; in this day and age, this skill isn’t enough to explain the Marlins’ propensity to win one-run games. So let’s turn back to pitching.

Overall, the pitching Fish (ERA- of 100) have performed slightly better than their hitting counterparts (wRC+ of 96). At least when it comes to these two measures, according to the table above, pitching wins more games than hitting. Plus, proportionally, this is even more true when it comes to one-run games. The way I see it is, if your bats are superior to your arms, your best wins will be blowouts, but if things are flipped, your most satisfying victories will include more 3-0, 2-0, and 1-0 scores. Russell Carleton over at Baseball Prospectus found that one-run games were significantly lower-scoring on average, even compared to other close — two- or three-run — contests, indicating the potential for good pitching teams to weather — and ultimately prevail in — more of them. Simulation-based methods have demonstrated the same result.

The Marlins’ slightly better arms have seemingly helped them a bit in close situations, but their pitchers aren’t better than their hitters by enough to make a major impact. And during Miami’s record-setting streak, up through their 12th consecutive one-run win on May 10, their pitching was actually a lot worse while their hitting hasn’t really changed: their ERA- was 107 and their wRC+ a similar 97.

This isn’t how the Marlins’ were supposed to look in their first non-pandemic playoff run since their 2003 World Series campaign; they were supposed to be pitching heavy. Instead, their hurlers have been merely average, barely better than their below-average bats. Aside from Eury Pérez, who’s been as good as advertised, the staff shares in the disappointment: Sandy Alcantara has regressed after a Cy Young 2022, Jesús Luzardo had a string of four poor outings before last night’s return to form, and Braxton Garrett hasn’t displayed the same dominance that he showed from mid-May to mid-June. Towards the back end of the rotation, Trevor Rogers has hardly been able to take the field after struggling last season to replicate his 2021 success, Edward Cabrera has walked his way back to the minors, and veteran Johnny Cueto is showing his age and is currently on the IL.

And while it’s hard to argue against a trade that netted them the decade’s first legitimate threat to hit .400, acquiring Luis Arraez cost the Marlins a current Cy Young candidate in Pablo López. Plus, Arraez is beginning to flounder: the slap-hitting second baseman has posted just a 64 wRC+ this month to go along with a putrid 1.2% walk rate, demonstrating the downsides inherent in his batted-ball dependent profile.

This all goes to show the risks that come with trading away pitching depth, even when it seems you have a wealth of it. Before you know it, you’ll have to replace that depth with more, and in the absence of López, you’ll have to turn to the likes of Cueto and Ryan Weathers. In a year or two, if his command returns post-Tommy John, we might be saying the same thing about Jake Eder, whom Miami dealt for the lower-ceilinged corner infielder Jake Burger.

To bring things full circle, there’s nothing about the Marlins’ current setup that screams “close game winner.” One-run game overperformers typically benefit from a lot of luck, and the Marlins don’t possess any of the qualities — like a strong bullpen or a proclivity for small-ball — that would indicate they’re an exception. While it’s commendable that they went all-in at the trade deadline, their anticipated regression has outweighed the impact of their new additions to the extent that they’ve gone 8-12 in August.

Perhaps it would have been better if they hadn’t won all of those one-run games to start the season; that way, they wouldn’t have spent as much prospect capital at the deadline in what increasingly looks to be a premature and futile attempt at a postseason push.