MLB's luckiest and unluckiest first-half teams: Time to buy low on the Yankees and Red Sox?

New York Times

6 days ago

With MLB's All-Star break in the rearview mirror, we're heading into a roughly 10-week, 60-game sprint to the finish, culminating in the start of October baseball. A lot can and will change during that stretch, from the looming trade deadline on July 31 to the usual churn of injuries and hot/cold streaks. And as we've seen time and again, even bad teams can pose as good ones (or vice-versa) over any given 60-game span.
Advertisement
But one of the biggest factors that can still shape the season is luck: Who has it, who's due for more of it and who might be running out of it. It's uncomfortable to admit this, but randomness plays a huge role in shaping the standings, sometimes accounting for as much as two-thirds of the variation between teams' records. And with fewer dominant teams than usual at the break, this season has been ruled by fortune, good or bad, more than most in recent memory.
So what does luck actually look like on the field? Sometimes it's players putting together unexpected career years, though that's harder to untangle from smarter player development in the era of data-driven coaching and scouting. But the most tried-and-true indicators involve teams outperforming their expected records, either by winning more than their run differential suggests (typically thanks to clutch and/or bullpen-driven wins in close games), or by producing more favorable sequencing — clustering hits together to score more runs than expected, or scattering opponents' hits to keep them off the scoreboard.
Using FanGraphs' BaseRuns model, which estimates how many runs a team 'should' score and allow based on its raw stats — along with Pythagorean winning percentage to estimate expected record based on run differential — we can break down both types of luck. And once we identify the luckiest and unluckiest teams, we can then dig into whether there are any good buy-low futures opportunities.
After converting the run-based effects into wins (using a standard exchange rate), here are the teams that benefited most — and least — from luck in the first half:
Luck ranks (more = luckier): Close games 2nd | Offense 10th | Defense 16th
We don't mean to downplay the Blue Jays' revival, one of the best stories of the first half. And while Toronto did look like one of the most improved teams of the 2024-25 offseason, it has been hardly lucky in the performance of its headline acquisitions — Andrés Giménez, Anthony Santander, Max Scherzer, etc., all of whom are on track for many fewer WAR than expected. Instead, the Jays' largest source of good fortune has been in converting run differential to wins: at +17 through the break, they were tied for 15th in scoring margin despite having MLB's seventh-best record. Toronto had the league's second-best record (38-21) in games decided by three runs or fewer, including a 17-12 mark in one-run games, but was below .500 (17-20) in games with a margin of at least four runs — a split between close games and blowouts that fueled one of the biggest disparities between a team's actual and Pythagorean records.
Advertisement
Luck ranks: Close games 4th | Offense 14th | Defense 8th
It's bizarre thinking of the Orioles as having been 'lucky' to start the 2025 season — if anything, it's felt like anything that could go wrong would, especially early in the schedule. While Baltimore has been better recently, going 27-18 from late May through the break, its playoff odds remain microscopic. So what's been so lucky? According to the Pythagorean expectation, the O's have the fourth-biggest split between their actual and expected records, thanks to a record 2 games over .500 (12-10) in one-run games and 11 runs below .500 (31-42) in multi-run contests. (Even during that recent eight-week improvement, they ranked 14th in Pythagorean win percentage despite ranking fifth in record.) Add in a pitching/defense corps allowing 13 fewer runs than predicted, and the O's have more than six extra wins compared with what we'd expect from their season-long stats.
Luck ranks: Close games 1st | Offense 9th | Defense 21st
At 47-49 heading into the ASG break, the Halos found themselves sitting surprisingly close to .500 — a level they haven't reached over a full season in a whole decade (since 2015). Some of that, however, was driven by a huge disparity between their record in one-run games (18-11, fourth best in MLB) and multi-run contests (29-38, seventh worst). Sometimes, massive splits like these can be partly explained by elite relief pitching — but that's not the case for the Angels' bullpen, which ranked 28th in WAR, with closer Kenley Jansen and his setup men rating as average at best by ERA and/or FIP. Another aspect of L.A.'s good fortune was an offense that ranked 19th in scoring despite sitting slightly below that in wRC+ (21st) and hitting WAR (22nd). Although the Angels lost some ground in luck on the other side of the ball, allowing 8.6 more runs than BaseRuns would predict, the net effect was a team that exceeded its expected wins overall.
Luck ranks: Close games 7th | Offense 3rd | Defense 12th
For the teams above, the majority of their overperformance came from winning more close games than we'd expect from their records in multi-run games. And the Dodgers went into the break with a bit of a split there, but not really too much of one. (Yes, they went 17-11 in 1-run contests, but they also had the league's third-best record in games decided by two or more, at 41-28.) Just as large a contributing factor to their fourth-place showing in the luck column, then, was an offense that scored nearly 20 more runs than expected, thanks to one of the league's largest OPS differentials with two outs and runners in scoring position versus overall, plus a run-prevention unit that benefited from timely stranded runners. (Of their nine most frequently used starters, seven had a left-on-base percentage above the league average.) L.A. has been unlucky in other regards, such as injuries, and it is a team to be feared in the postseason, but some of this run-differential luck might regress by season's end.
Luck ranks: Close games 5th | Offense 17th | Defense 9th
Like with Baltimore, there's nothing that feels especially fortunate about the way Cleveland's 2025 season has played out: One of the best teams by record a year ago, the Guardians have struggled with their offense and once-dominant bullpen en route to a slightly below .500 record at the break and flagging playoff odds. But somehow, matters could be even worse if they were not 34-25 in games decided by three runs or fewer, versus 12-24 in all other contests — a record that ranked only ahead of the Rockies' abysmal 5-33 tally in such games. And although their luck broke even exactly on offense, Cleveland pitchers allowed 12.4 fewer runs than expected, headlined by LOB percentage rates either at or above average for all six of the Guardians' primary starters. So not only will Cleveland have to overcome a long list of teams ahead of it in the standings, but also it will have to fight off any slowdown from its luck wearing off.
Luck ranks (more = luckier): Close games 30th | Offense 16th | Defense 27th
The Aaron Boone-era Yankees are a fascinating, ongoing test case in whether luck-based stats like these can ever truly regress back to where we'd expect them to be. On the one hand, I (and others) have been complaining about New York's lack of fundamentals for years, citing them as a primary reason the team doesn't always have the record its talent and underlying stats would predict. However, this year's splits are so pronounced that some kind of positive regression seems inevitable. (An eight-win luck shortfall in 96 games is truly enormous.) The Yankees were just 13-18 in one-run games in the first half, and 18-26 in those decided by two runs or fewer, while they were a league-best 35-17 in games with a margin of three runs or more. Performance in blowouts is a strong predictor of future success, and New York's improvement potential is further bolstered by the fact that pinstriped pitchers have allowed 17.3 more runs than expected via BaseRuns. While some of New York's bad luck may be unfixable, a differential like this is extreme.
Advertisement
Luck ranks: Close games 23rd | Offense 29th | Defense 28th
When it comes to historically good or — in this case — bad teams, it takes luck on top of talent (or the lack thereof). So, although the 2025 Rockies are legitimately one of the worst teams in MLB history, and they'll be a threat down the stretch to challenge the 2024 White Sox's record for most losses in a modern season (at 121), Colorado also rode some misfortune to get to its 22-74 mark at the break. The shortfalls came across the board: The Rockies were the only team in the first half to lose at least two wins of luck from close games (they were predicted to be a marginally less-horrendous 24-72 based on their run differential), two wins of luck from offensive sequencing (their OPS was 17 percent worse with two outs and RISP) and two wins of luck from bunching too many hits on the other side of the ball (no team has a lower rate of stranding runners). This Colorado team would have been in the worst-ever conversation regardless of its luck … but that luck is not doing the Rockies any favors, either.
Luck ranks: Close games 29th | Offense 12th | Defense 19th
Unlike Colorado, Texas is an example of a team where neutral luck would make a huge difference to the complexion of its season. As things stood at the break, the Rangers went in with a disappointing sub-.500 (48-49) record and playoff odds around 20 percent. But they also had MLB's sixth-best record (22-14) in games decided by four or more runs, despite its sixth-worst record (26-35) in games decided by three runs or fewer — including 14-17 in one-run games specifically — which helps explain why their Pythagorean record of 54-43 would have had them sitting in the American League's first wild-card spot at the break instead of 3.5 games back. The rest of Texas' luck profile is fairly neutral on balance, but its Pythagorean gap makes the team an easy stretch-run improvement pick.
Luck ranks: Close games 28th | Offense 27th | Defense 3rd
Another of the most disappointing teams in MLB's first half, the Braves have been held back by injuries as much as anything else, but their poor luck metrics are also leaving them in a much worse position than they would be otherwise. The All-Star hosts were six games underwater in their Pythagorean differential at the break, with a huge split between their record in close games (13-24 in one-run games, 17-32 in two-run games) and blowouts (25-21 in games decided by three or more runs). And while their pitchers redeemed them with one of the league's highest strand rates, their batters hit .214 with two outs and RISP, scoring a whopping 22 fewer runs than expected from their overall stats. With the Braves buried in the wild-card standings, will any improvement be far too little, far too late?
Luck ranks: Close games 22nd | Offense 11th | Defense 30th
It's no surprise that the Red Sox are a baffling team on multiple levels — that's just who they always seem to be. The same club that traded Rafael Devers to the Giants in mid-June, then promptly went 16-9 (including a 10-game win streak going into the break), was also among baseball's unluckiest in the first half.
Boston went 12-19 in one-run games (fifth worst in MLB) but 41-26 (second best) in multi-run contests, helping fuel a 2.2-win gap in Pythagorean versus actual records. The Red Sox also allowed a staggering 38.2 more runs than BaseRuns predicted they should have, by far the worst mark in the league — Washington was second worst at -28.8.
So, even with an offense that scored 7.8 more runs than expected, the Red Sox came out more than five wins short of projected. As much as their recent dominance feels inexplicable, this is an indicator that it might have more staying power than it seems once the luck straightens itself out. (If the Red Sox were a team that made any sense, that is.)
Neil Paine is a freelance writer whose work also appears regularly at ESPN.com, NASCAR.com, The Philadelphia Inquirer, Sherwood News and his eponymous Substack. He is the former Sports Editor at FiveThirtyEight, and was also an analytics consultant for the NBA's Atlanta Hawks.
Betting/odds links in this article are provided by partners of The Athletic. Restrictions may apply. The Athletic maintains full editorial independence. Partners have no control over or input into the reporting or editing process and do not review stories before publication.
(Illustration: Kelsea Petersen / The Athletic; Matthew Grimes Jr. / Getty, Tim Warner / Getty, Katherin Skeean / Getty)