Two men race the same 10K. The 50-year-old runs 38:00. The 54-year-old crosses the line 75 seconds later, at 39:15. On the results sheet, the 50-year-old beat him. On the age-graded tables, the 54-year-old ran the better race — 78.3% against 78.2%.
Both men are in the same M50-54 award bracket. The one who finished behind still ran the more impressive performance for his age, and the trophy would never tell you that.
So this piece is about the bracket, not the trophy. Why "M50-54" is a structural lie the moment it's assigned, and what a continuous number fixes.
What a bracket actually assumes
An age-group award compares clock times inside a fixed window, usually five years. Whoever runs fastest in "M50-54" wins, regardless of whether they're 50 or 54. That's not a rounding choice. It's a claim that a 50-year-old and a 54-year-old are running the same event.
They aren't. The real age factor — the multiplier the age-grading tables use to account for age — moves every single year, not every five. The bracket just doesn't look at it.
The dashed staircase above is what a bracket effectively assumes: flat at the bracket's youngest age, then a hard drop at the boundary. The smooth line is the real curve underneath it. They only agree on the day someone turns the bracket's entry age. Every other day inside that bracket, the staircase is wrong, and it gets more wrong as the birthday approaches.
The trophy paradox, worked out
Here's the exact case from the top of this piece. Both men run a 10K, age factors from the Alan Jones 2025 Age-Grade Tables:
| Runner | Age | Time | Age-graded % |
|---|---|---|---|
| A | 50 | 38:00 | 78.18% |
| B | 54 | 39:15 | 78.34% |
Runner B finishes 75 seconds behind. He's still in the same M50-54 bracket. By the age-graded tables, he ran fractionally the better race — because the tables know he's 54, and the bracket doesn't.
Widen the gap and it gets clearer, not murkier. Push Runner B out to 39:30, still comfortably beaten on the clock, and he's now at 77.84% against Runner A's 78.18% — the gap in the other direction, but by a smaller margin than the raw 90 seconds on the clock would suggest. The age factor is always pulling the comparison back toward "how hard did each of you actually work", not "who has fresher legs from being younger inside the same box".
How big is the bracket gap, exactly?
I pulled the real factor at the youngest and oldest age in the 50-54 bracket, men and women, across four distances. The bracket assumes the "50" column. Reality, by the time you're 54, is the "gap" column.
| Distance | M50 factor | M54 factor | Gap | F50 factor | F54 factor | Gap |
|---|---|---|---|---|---|---|
| 5K | 0.8775 | 0.8495 | 3.3% | 0.8592 | 0.8204 | 4.7% |
| 10K | 0.8886 | 0.8586 | 3.5% | 0.8822 | 0.8425 | 4.7% |
| Half | 0.8860 | 0.8549 | 3.6% | 0.8753 | 0.8339 | 5.0% |
| Marathon | 0.8963 | 0.8635 | 3.8% | 0.8998 | 0.8619 | 4.4% |
Translate that into clock time and it stops being an abstraction. A 50-year-old man's 80% marathon is 2:48:10. The same 80% for a 54-year-old is 2:54:33 — a 6:23 gap the bracket never asks about. For women the gap runs even wider, 7:56 on the marathon between the same two ages. The bracket pays out the same trophy at both ends.
Why do brackets exist anyway?
Five-year windows aren't stupid, they're practical. A results system needs buckets simple enough to print on an awards podium and small enough that "masters" doesn't mean one giant category from 40 to 90. Race timing software, awards ceremonies, and 40 years of results-page convention all assume it. Nobody sat down and decided five years was the physiologically correct unit. It's the unit that fits on a printed results sheet.
That's a fine compromise for handing out age-group medals at a race. It's a bad tool for the question a masters runner actually wants answered: did I run better than I did last year, for someone my age, right now? A bracket resets that question every five years instead of answering it continuously.
What does the single number fix?
Age-graded percentage doesn't group you into a window at all. It looks up your exact age on race day and applies the factor for that year, not the bracket's floor. There's no boundary to sit just inside or just outside of. Turn 50 in October and your percentage moves a little on your birthday, the same amount it would move on any other birthday. Nothing resets, nothing steps.
That's also why your percentage can rise even when your finish time gets slower, something I've written about before. The curve keeps crediting you for age, every year, whether or not a race director drew a box around it.
Curious what your own races look like on the continuous curve instead of the bracket? The calculator uses your exact age, not the nearest five-year block.
Sources
- Alan Jones, Age-Grade Tables (road, 2025 edition — approved 2025-01-10 by the USATF Masters Long Distance Running Council) — retrieved 2026-07-20, https://github.com/AlanLyttonJones/Age-Grade-Tables
- RaceRecords, Methodology — retrieved 2026-07-20, https://racerecords.run/methodology