The runners who beat the time we predicted for them by the most — measured against the prediction the site was showing on the morning of the race, with the course and the day taken out. Open to the whole field, not just the front of it.
Liem Ormberg
32.1% under their prediction. Worth 32:59 once that field's day is taken out — 6.2× the normal week-to-week swing for a runner at that speed.
| Runner | Predicted | Ran | Under | Score | |
|---|---|---|---|---|---|
| 2 | Halim MohamedLewiston High School · SO | 22:49 | 19:54 | 9.4% | 4.13 |
| 3 | Aeron SpradlinSkowhegan Area High School · FR | 35:44 | 27:29 | 18.8% | 3.63 |
| 4 | Chase VanDerburghWindham High School · SO | 43:07 | 32:57 | 16.1% | 3.12 |
| 5 | Brody DowslandMt. Desert Island High School · FR | 29:50 | 23:39 | 14.4% | 3.10 |
Ranked by score, not by the raw percentage — a slower runner has more room to move, so the percentages are not comparable on their own. The explanation below says how that is handled.
Zachary Daniels
13.3% under their prediction. Worth 28:19 once that field's day is taken out — 2.6× the normal week-to-week swing for a runner at that speed.
| Runner | Predicted | Ran | Under | Score | |
|---|---|---|---|---|---|
| 2 | Joshua LongLake Region High School · JR | 26:19 | 22:09 | 8.7% | 2.56 |
| 3 | Logan BussellSkowhegan Area High School · JR | 23:11 | 20:47 | 6.8% | 2.46 |
| 4 | Sam ZhangSkowhegan Area High School · JR | 26:12 | 23:09 | 8.2% | 2.46 |
| 5 | Jake LittlefieldNokomis High School · SR | 25:30 | 22:46 | 7.4% | 2.36 |
Ranked by score, not by the raw percentage — a slower runner has more room to move, so the percentages are not comparable on their own. The explanation below says how that is handled.
Kylee Hayes
16.3% under their prediction. Worth 28:54 once that field's day is taken out — 3.1× the normal week-to-week swing for a runner at that speed.
| Runner | Predicted | Ran | Under | Score | |
|---|---|---|---|---|---|
| 2 | Dakota MoultonMarshwood High School · FR | 30:35 | 24:58 | 11.3% | 2.37 |
| 3 | Harper OakesOld Town High School · SO | 28:43 | 24:44 | 9.3% | 2.11 |
| 4 | Arianna AlleyNarraguagus High School · FR | 26:21 | 23:24 | 6.5% | 1.90 |
| 5 | Alexis BrittEllsworth High School · FR | 25:35 | 23:01 | 5.3% | 1.71 |
Ranked by score, not by the raw percentage — a slower runner has more room to move, so the percentages are not comparable on their own. The explanation below says how that is handled.
Lee Maloney
13.7% under their prediction. Worth 25:49 once that field's day is taken out — 2.9× the normal week-to-week swing for a runner at that speed.
| Runner | Predicted | Ran | Under | Score | |
|---|---|---|---|---|---|
| 2 | Emily SaundersMaranacook Community High School · SR | 31:43 | 28:15 | 9.2% | 1.87 |
| 3 | Lily DouglassLewiston High School · SR | 30:02 | 27:05 | 7.5% | 1.61 |
| 4 | Annabelle BachmanFryeburg Academy · SR | 25:52 | 23:22 | 4.7% | 1.51 |
| 5 | Grace GaumontLake Region High School · SR | 33:06 | 29:11 | 6.9% | 1.36 |
Ranked by score, not by the raw percentage — a slower runner has more room to move, so the percentages are not comparable on their own. The explanation below says how that is handled.
How the award is decided
Every runner on this site carries an index — a projected 5K on a neutral course, worked out from the races they have already run. Index Breakers goes to whoever beat theirs by the most.
The prediction has to pre-date the race
The index here is rebuilt from scratch using only races run strictly before 2026-10-12. It is the number the site was showing on the morning of the race, which is the only number anyone could have been trying to beat. Include the race itself and the model quietly drags the prediction toward what actually happened, and nobody can beat anything.
The course and the day are taken out
If a whole field beats its index by thirty seconds, no model can tell whether the course ran fast or everybody had a good day — those are the same observation. So each runner is measured against the rest of their own field, with themselves left out of that baseline, so a runner who destroys the field cannot flatten their own bar by sitting in the middle of it.
Stated plainly: this measures who outran the day, not who ran fastest. A whole team running out of their minds together will not sweep the list, because they moved the day with them.
Fast and slow runners are compared fairly
A raw percentage sounds fair and is not. Measured across this week's field, the best breaks run about ±2.2% for runners around 19:17 and ±5.2% for runners around 33:32 — slower runners simply have more room to move. Rank on the raw percentage and the award goes to the slowest qualifying runner every single week, by construction.
So each break is divided by how much a runner of that speed typically varies. A 19:17 runner going 4.4% under scores the same as a 33:32 runner going 10.3% under. That is what makes this open to the whole field rather than nominally open to it.
Who qualifies
2 or more races before the window. A runner with one race has no index worth beating — that is most freshmen in early September, and it is right rather than unfortunate: there is no claim to make about someone the model has barely seen. They qualify the moment they have a second race.
What the numbers mean
- We predicted — their index on a neutral course, with the ± showing how well we knew them. A wide band means a thin record, and beating a wide band is less of a claim than beating a narrow one.
- They ran — the actual clock, on the actual course.
- Worth — that run with the course and the day removed, which is the number being compared to the prediction.
Built on 8,743 results run before the window and 1,244 run during it, across 18 races. Ratings are estimates with stated uncertainty, not official times or seeds.