Are Boston Marathon
Qualifying Times Fair?

If you're a 45-year-old woman who trained for months to hit a 3:45 marathon, are you being held to the same standard of effort as a 30-year-old man chasing 2:55? Or are some brackets simply getting a better deal? After analyzing 22 brackets through three different lenses, the answer turns out to be more interesting than either side of the usual argument.

The Question Nobody Answers

On April 20, 2026, roughly 30,000 runners will line up in Hopkinton, Massachusetts to run the world's oldest annual marathon. Most of them earned that starting spot the hard way: by running a qualifying time in another marathon, somewhere in the world, within the previous 18 months. The Boston Marathon is one of only two World Marathon Majors that requires you to qualify by performance rather than by lottery or charity.

The Boston Athletic Association (BAA) publishes qualifying times across 11 age groups and 3 gender categories. For the 2026 race, they tightened standards by a full five minutes across the board for everyone under 60, responding to record demand: 33,249 applications for roughly 24,000 qualifier spots. Even after the tightening, runners had to beat their published standard by 4 minutes 34 seconds to actually get in.

But here's the thing: in the BAA's century-long history of adjusting these standards, they've never publicly explained how they arrive at each bracket's specific time. Their official press releases reference “careful analysis of results data” without specifying whether they're optimizing for equal difficulty, equal selectivity, field-size targets, historical continuity, or something else entirely.

The Data Behind 30,000 Hopefuls

We analyzed every bracket in the 2026 Boston Marathon (11 age groups × 2 genders, excluding non-binary because the BAA itself notes insufficient data to set evidence-based standards for that category yet).

Data sources, all verified against primary sources:

• 2026 BAA Qualifying Standardsbaa.org
• Open Marathon World RecordsWorld Athletics
• Masters Records (M35-M80, W35-W80)Wikipedia / WMA / ARRS
• WMA Age-Grading FactorsWMA 2023 Appendix B
• 2026 Field Size MetricsBAA press releases

The marathon world records anchoring this analysis are extraordinary. Sabastian Sawe set the open men's record of 1:59:30 at London in April 2026, becoming the first man to break two hours in a record-eligible race. The women's open record stands at 2:09:56 (Ruth Chepngetich, Chicago 2024). For the 80-and-over bracket, Ed Whitlock's 3:15:54 (men) and Yoko Nakano's 4:11:45 (women) are the references; both are landmark performances for athletes in their ninth decade.

Three Lenses on Fairness

Every fairness analysis depends on what you compare to. We deliberately chose three anchors that capture different intuitions about what “equal difficulty” should mean.

What the Numbers Actually Say

Here's the first surprise. Under Framework 1, the median multiplier across all 22 brackets is almost exactly 1.50×. The typical BQ standard requires you to run 50% slower than your bracket's world record. But the spread tells the real story.

Notice how the blue bars all sit between 1.43 and 1.53 — a range of only 0.10. The red bars span 0.35, more than three times that range. For men, the “standard” difficulty relative to their world record is remarkably consistent. For women, it's all over the map.

A formal statistical test confirms what the eye sees. A Welch t-test comparing the two distributions returns p = 0.81 — no significant difference in means. But Levene's test for equal variance returns p = 0.036: the variances are statistically different. The averages are balanced. The consistency isn't.

What does a 1.50× multiplier mean in practice? It means that if Sabastian Sawe's 1:59:30 represents the ceiling of male marathon performance, an 18-34 man hitting his 2:55:00 BQ standard is running at 68% of that pace. Hold that proportion constant and you get the “fair” BQ for every bracket. The next chart shows the gap between what the BAA sets today and what each framework suggests.

The pattern is striking. Younger women's brackets (18-34, 35-39, 45-49) sit well above the zero line — they're relatively lenient relative to their reference records. Then everything flips at 70+, where women's brackets become aggressively strict. Men's bars hover near zero across the board.

Did the 2026 Tightening Help?

The 2026 race introduced the largest single tightening of qualifying times since 1990 — five minutes across the board for athletes under 60. The BAA's stated rationale: more applicants than ever, athletes getting faster, and a desire to set a standard that better reflects current performance levels. But did the tightening also improve fairness across brackets, or did it just shift everything uniformly?

The answer: the tightening was uniform within each gender, which means the relative structure of the standards barely moved. The 60+ brackets weren't touched at all. Women's variance in 2020-2025 (CV ≈ 6.8%) and 2026 (6.6%) are essentially identical. The 2026 changes responded to demand, not to fairness across brackets.

There's an interesting wrinkle here. By tightening only under-60 standards, the BAA implicitly steepened the “step” between the 55-59 and 60-64 brackets. Under the 2020-2025 standards, a 59-year-old man needed 3:35; his 60-year-old self needed 3:50 — a 15-minute jump. Under 2026, the gap is 20 minutes. That birthday cliff is now larger.

How Robust Are These Conclusions?

Before drawing conclusions, we stress-tested the main finding (women's brackets are 3-4× more variable than men's) against three alternative scenarios. If a single outlier or a debatable choice of reference record drives the result, we'd see the gap collapse under any of these perturbations.

Three takeaways from this:

The W80+ bracket alone accounts for nearly a third of women's bracket variance. Dropping it cuts women's CV from 6.6% to 4.5%. That's a single data point doing enormous statistical work. It's both the most miscalibrated bracket and the one most dependent on one extraordinary athlete (Yoko Nakano's 4:11:45) for its reference record.

The result is robust to alternative records. Substituting Sinead Diver's W40 marathon (2:21:34) as the W40-44 reference barely changes women's CV. Using the women-only WR (Assefa 2:15:41) instead of the mixed-race WR shifts CV from 6.6% to 6.4%. The variance gap is not an artifact of which records we anchor to.

Even with the most outlier-friendly choices, women's CV remains 2-3× higher than men's. The structural inconsistency in women's brackets is genuine.

The Honest Take

This analysis cannot tell you which framework is “correct.” That's not a data question; it's a values question. If you believe fairness means equal difficulty relative to current world records, Framework 1 is your lens. If you want to account for biological aging, Framework 3 is more defensible. If you think the BAA should optimize for equal selectivity, field-size diversity, or historical continuity, none of these frameworks fully capture that.

What the data can tell you is this: the BAA has the averages right. The mean difficulty is balanced across genders under every framework we tested. The criticism that “Boston is unfair to women” (or men) does not survive statistical testing. The consistency, though, is not balanced. Women's brackets are 3-4× more variable than men's, and the W80+ bracket is a genuine outlier regardless of how you measure.

If the BAA wanted to address one bracket on equity grounds alone, the data points clearly at W80+. Adjusting it by 30-50 minutes would bring it in line with every other bracket under either the WR or age-graded framework. The change would affect roughly a few dozen runners per year, but the symbolic message — that the standard for an 80-year-old woman should be calibrated to the same difficulty as a 30-year-old man — would be substantial.

The deeper question isn't whether the BAA is right or wrong. It's whether they've publicly committed to a fairness framework at all, or whether the standards have evolved through a combination of historical inertia, demand management, and ad-hoc adjustments. The data is consistent with the latter. A more transparent methodology — even if it leads to the same numbers — would let runners, coaches, and statisticians evaluate the system on its own terms rather than guessing at what it's trying to do.