The 2025 Epsom Derby favourite went off at 2/1. Delacroix carried impressive Group class juvenile form and a pedigree profile the market deemed acceptable for Derby distance. He finished ninth, beaten sixteen lengths, never threatening.
Three months later, he’d won the Group 1 Coral Eclipse Stakes. Then the Group 1 Irish Champion Stakes. Two elite victories—both at ten furlongs, not twelve.
His pedigree suggested Derby stamina. His stride revealed something different.
Does dosage index work for predicting racehorse optimal distance—or do modern biomechanical methods provide more reliable predictions?
For over forty years, dosage index has been racing’s standard for predicting distance suitability. Developed by Dr. Steven Roman in 1981, it quantifies stamina potential through sophisticated pedigree analysis. Tools like Pedigree Query make dosage calculations accessible—breeders, trainers, and punters rely on these profiles when evaluating whether horses will handle Classic distances.
But pedigree tells you what a horse inherited from ancestors, not what the individual can actually do. Stride biomechanics offers a different approach—measuring the horse’s actual physical capacity through observable racing data: stride frequency, stride length, how efficiently the horse deploys energy.
One method analyses bloodlines. The other measures individuals.
This article tests both approaches against the 2025 Epsom Derby field to answer one question: which method better predicts racehorse optimal distance?
The comparison uses eighteen Derby runners with complete dosage profiles, stride predictions for sixteen horses, actual race results, and—critically—post-Derby performances that validated optimal distances through subsequent Group race careers. Not theory. Validation.
Both methods have value. One proves significantly more accurate for individual predictions.
Here’s what the data reveals.
Does Dosage Index Work? Comparing Pedigree vs Biomechanics
Quick Answer: Is Dosage Index Accurate?
For identifying stamina pedigrees: Yes – 69% of Derby winners (2000-2025) fell within the expected range.
For predicting individual optimal distance: No – horses with identical dosage profiles (e.g., DI 1.00) can have optimal distances ranging from 9f to 13f.
Stride biomechanics provides specific predictions: 9.5f, 12.6f, 10.7f – immediately actionable for race distance evaluation.
How Both Methods Work—And Why One Is More Precise
The Dosage Index Approach
Dosage index identifies whether a horse’s pedigree leans toward speed or stamina by analysing influential sires (called chef-de-race) across multiple generations. The calculation produces two numbers: a Dosage Index (DI) and Center of Distribution (CD).
Derby winners from 2000-2025 averaged DI 0.86. The typical Derby profile falls between DI 0.70-1.10. In this twenty-six year sample, 69% of winners fell within this range, with 31% outside.
So does dosage index work for identifying Derby horses? The statistical correlation suggests directional usefulness, but range-based prediction has inherent limitations.
If a horse shows DI 1.00, you know it fits the Derby statistical profile developed from historical winners. What dosage doesn’t tell you: will this specific horse actually be most effective at twelve furlongs? Multiple horses with identical DI values perform vastly differently at the same distance.
The limitation: Dosage identifies horses fitting population-based statistical ranges. It doesn’t predict optimal distance for the individual. And the profile remains fixed from birth—a two-year-old and five-year-old with the same pedigree receive identical numbers despite physical development.
The Stride Biomechanics Approach
StridePredictor’s models analyse racing biomechanics to predict a specific optimal distance for individual horses. Not a range requiring interpretation. A number: 12.6 furlongs, 9.5 furlongs, 10.3 furlongs.
The key difference: Specificity. When evaluating a twelve-furlong Derby, a horse predicted at 9.5f optimal is clearly unsuitable. A horse at 12.6f shows stamina credentials for the race distance. The prediction is immediately actionable.
Age-specific models: Separate predictions for two-year-olds, three-year-olds, and mature horses account for physical development. As horses mature and race, predictions update to reflect changing capacity. A two-year-old showing 10.5f optimal might develop into 11.8f as a three-year-old.
Individual measurement: Two horses with identical pedigrees but different stride mechanics receive different predictions. One might have the biomechanics for twelve furlongs. Another might suit a mile despite identical breeding. Stride distinguishes between them through actual physical data, not inherited probability.
Range vs Distance
The fundamental distinction: dosage identifies horses fitting statistical profiles developed from populations. StridePredictor predicts the specific distance where this individual horse will be most effective based on biomechanical data.
For Derby evaluation, that specificity matters. When dosage places multiple horses in the same 0.70-1.10 range, how do you distinguish between them? One might have genuine twelve-furlong biomechanics. Another might be optimal at 10.7f—close to the range, but insufficient at elite Derby level.
Of the 18 runners in the 2025 Derby, 12 horses (67%) fell within the expected 0.70-1.10 range. The 2025 Derby field tests whether dosage index works for individual predictions—or whether range-based methods struggle to separate horses with similar profiles.
The 2025 Derby Test: Does Dosage Index Work?
Eighteen runners, two distance prediction methods, one question: which identifies horses racing at their optimal distance versus those racing beyond their effective range?
StridePredictor optimal distance predictions are taken from each horse’s most recent race before the Derby, provided the race returned reliable pace data. Races with significantly distorted pace receive low confidence ratings and aren’t used for predictions.
The Predictions
Dosage’s view: Twelve horses carried DI profiles within the typical Derby range (0.70-1.10), all showing pedigrees consistent with Derby distance:
- Midak (0.74)
- Lambourn (0.76)
- Green Storm (0.76)
- New Ground (0.78)
- Stanhope Gardens (0.80)
- Tuscan Hills (0.82)
- Tennessee Stud (1.00)
- Tornado Alert (1.00)
- Damysus (1.00)
- Pride Of Arras (1.00)
- Nightime Dancer (1.08)
- The Lion In Winter (1.09)
The range couldn’t distinguish between them. All fit the statistical profile developed from Derby winners.
StridePredictor’s view: Only four horses showed optimal distances at twelve furlongs or longer:
- Tennessee Stud (12.9f)
- Lambourn (12.6f)
- Lazy Griff (12.0f)
- Al Wasl Storm (12.0f)
Every other horse predicted below 12f, suggesting peak effectiveness at shorter distances than Derby requires.
Testing whether dosage index works requires comparing these predictions against actual results. The comprehensive field analysis below shows both methods side-by-side.
The Complete Field Analysis
The table below shows all eighteen Derby runners ordered by stride-predicted optimal distance, demonstrating how both methods evaluated the same field:
| Horse | Stride Optimal | Dosage (DI) | Finish |
|---|---|---|---|
| Tennessee Stud | 12.9f | 1.00 ✓ | 3rd |
| Lambourn | 12.6f | 0.76 ✓ | 1st |
| Al Wasl Storm | 12.0f | 0.41 ✗ | 15th |
| Lazy Griff | 12.0f | 0.60 ✗ | 2nd |
| Nightime Dancer | 11.3f | 1.08 ✓ | 8th |
| Nightwalker | 11.1f | 0.67 ✗ | 12th |
| Pride Of Arras | 11.1f | 1.00 ✓ | 17th |
| Rogue Impact | 11.0f | 0.64 ✗ | 13th |
| Sea Scout | 10.7f | 1.74 ✗ | 11th |
| Tuscan Hills | 10.7f | 0.82 ✓ | 16th |
| Damysus | 10.7f | 1.00 ✓ | 18th |
| The Lion In Winter | 10.3f | 1.09 ✓ | 14th |
| Green Storm | 10.2f | 0.76 ✓ | 7th |
| Stanhope Gardens | 9.6f | 0.80 ✓ | 5th |
| Delacroix | 9.5f | 1.18 ✗ | 9th |
| Tornado Alert | 8.7f | 1.00 ✓ | 6th |
| New Ground | No prediction | 0.78 ✓ | 4th |
| Midak | No prediction | 0.74 ✓ | 10th |
Key:
- ✓ = Dosage within Derby profile range (DI 0.70-1.10)
- ✗ = Dosage outside Derby profile range
The Results
The top three finishers all clustered at the top of stride predictions:
| Horse | Stride Optimal | Dosage (DI) | Finish |
|---|---|---|---|
| Lambourn | 12.6f | 0.76 ✓ | 1st |
| Lazy Griff | 12.0f | 0.60 ✗ | 2nd |
| Tennessee Stud | 12.9f | 1.00 ✓ | 3rd |
StridePredictor identified all three podium finishers. Dosage correctly identified Lambourn and Tennessee Stud as suitable but underestimated Lazy Griff—his DI 0.60 fell below the typical range despite finishing second at 50/1.
For the top finishers, does dosage index work? Yes—two of three fell within the predicted range. The real test comes with horses dosage flagged as suitable but who failed at Derby distance.
The divergence appeared with horses dosage flagged as suitable but StridePredictor eliminated.
Three prominent in the Derby betting demonstrated this clearly—and their post-Derby Group race careers provided definitive proof.
When Methods Disagreed
Delacroix (2/1 favourite):
- Dosage: DI 1.18 (outside typical range but Sea The Stars won with DI 1.67)
- Stride: 9.5f optimal
- Derby result: 9th, beaten 16 lengths
- Post-Derby: Won G1 Eclipse (10f), 2nd G1 International (10f), won G1 Irish Champion (10f), 4th G1 Champion Stakes (10f)
Four Group 1 performances, all at ten furlongs. Optimal distance where stride predicted, two and a half furlongs short of Derby requirements.
Damysus (16/1):
- Dosage: DI 1.00 (perfect Derby profile, within historical average)
- Stride: 10.7f optimal
- Derby result: 18th (last place)
- Post-Derby: Won Listed Nureyev Stakes (10f), won G3 Darley Stakes (9f)
Perfect Derby dosage—identical DI to Workforce (2010 Derby winner by seven lengths) and Tennessee Stud (12.9f optimal, third in this Derby), though CD differs: 0.20 vs 0.23. Yet his biomechanics supported 10.7f, not twelve. Post-Derby victories at nine to ten furlongs validated the stride prediction despite ideal pedigree profile.
The Lion In Winter (7/1):
- Dosage: DI 1.09 (within Derby range)
- Stride: 10.3f optimal (Dante Stakes, pre-Derby)
- Derby result: 14th, beaten 28 lengths
- Post-Derby: Best performances at 7f-1 mile across six runs, including two G1 placings
His Dante Stakes prediction showed 10.3f optimal—nearly two furlongs short of Derby requirements. But all horses “stay” eventually. The Lion In Winter completed twelve furlongs at Epsom. He just couldn’t do it competitively.
The clue was there earlier. As a two-year-old, his Acomb Stakes revealed 7.5f optimal—a speed signature right at the seven-furlong breakpoint indicating sprint/mile biomechanics, not Derby stamina. Post-Derby career validated this: best performances came at seven furlongs and one mile, exactly where his two-year-old form predicted.
Dosage suggested Derby distance. Stride revealed a miler aimed incorrectly.
The Pattern: Individual vs Population
Three horses answer the question definitively: does dosage index work for eliminating unsuitable runners?
The pattern was identical across all three. Dosage profiles suggested Derby suitability. Stride identified shorter optimal distances. Derby performances confirmed stride. Post-Derby Group careers proved the predictions correct.
When racing at distances matching their biomechanical predictions, all three competed at elite level. At Derby distance beyond their optimal range, all three failed. This wasn’t about class or ability—they proved their quality in Group class company afterward. It was about distance suitability.
The critical insight: Dosage placed Damysus (10.7f optimal, finished last) in the same statistical profile as Tennessee Stud (12.9f optimal, finished third) because both inherited similar pedigree characteristics—both showed DI 1.00. The range couldn’t distinguish between them.
StridePredictor separated them by more than two furlongs. That specificity was the difference between identifying a genuine Derby horse and backing a 16/1 shot who finished last.
Modern racehorse distance prediction requires individual measurement, not population-based statistical ranges. Pedigree tells you probability. Biomechanics tells you capacity.
FREQUENTLY ASKED QUESTIONS
Does dosage index work for Derby predictions?
Yes, dosage index works for identifying stamina pedigrees. Analysis of 26 Derby winners (2000-2025) shows 69% fell within the typical DI 0.70-1.10 range, with 31% outside. However, accuracy for individual horses varies significantly—multiple horses with identical profiles finish in opposite positions.
What are the main limitations of dosage index?
Dosage identifies population-based statistical ranges, not individual optimal distances. The profile remains fixed from birth despite physical development, and it cannot distinguish between horses within the same range.
Is there a better alternative to dosage index?
Stride biomechanics analysis provides specific optimal distance predictions based on individual physical capacity rather than inherited probability. It accounts for age-specific development and distinguishes between horses with identical pedigrees but different biomechanical capabilities.
Modern Distance Prediction: Why Specificity Matters
The 2025 Epsom Derby answers definitively: does dosage index work? Yes, for identifying stamina pedigrees.
The question is whether range-based population statistics provide sufficient precision for individual predictions. Dosage index correctly identified horses with stamina pedigrees. StridePredictor went further—predicting specific optimal distances that separated genuine twelve-furlong horses from those best suited to shorter trips.
When dosage placed twelve horses within the same 0.70-1.10 range, stride biomechanics distinguished between them with precision. Damysus and Tennessee Stud carried identical DI profiles (1.00). Stride separated them by more than two furlongs—the difference between a Derby horse who finished third and one who finished last.
That specificity isn’t academic. It’s actionable. Delacroix went off 2/1 favourite with acceptable dosage. StridePredictor eliminated him at 9.5f optimal. Post-Derby Group 1 victories at ten furlongs validated the prediction. The information existed months before the Derby—before trials, before the market knew, before prices reflected reality.
2026 Derby Predictions
Stride predictions for the 2026 Derby and Oaks fields are updated as horses race through the season. Spring Derby trials reveal distance suitability through form. Stride biomechanics reveals it through measurement—often months earlier.
The first-stage Derby 2026 entries have already been published — stride analysis of 27 lightly raced entries is available in Derby 2026: Trainer Intent.
Want 2026 Derby stride predictions before trials prove them and prices move?
Get distance predictions for Classic contenders, updated analysis as horses progress through trials, and biomechanical insights the market hasn’t priced in yet.
Related Articles
Epsom Derby 2026 Predictions – Full stride analysis of current Derby contenders
Racehorse Stride Biomechanics Explained – How stride predicts optimal distance
Epsom Oaks 2026 Predictions – Stride analysis for the fillies’ Classic
The 7-Furlong Breakpoint – Why some horses develop speed while others develop stamina