Trained for the Trip?
Exploring Distance Preparation in Flat Turf Racing
The idea for this analysis came directly from the 2025 Gold Cup at Royal Ascot. In the build-up, John Gosden was emphatic: Trawlerman was the horse to beat, not necessarily the most talented, but the one most guaranteed to get every yard of the trip.
That trip, of course, is no ordinary one. The Gold Cup’s two and a half miles demand a unique blend of stamina, conditioning, and tactical judgement.
At the same time, plenty of confidence surrounded Illinois, Ballydoyle’s up-and-coming colt. Talented, progressive, and reportedly thriving at home, he was reportedly one of Aidan O’Brien’s best-fancied runners of the meeting.
In the race, Trawlerman delivered exactly as promised, grinding out victory with relentless stamina. But it was the post-race interview that prompted this article. O’Brien was candid: he hadn’t trained Illinois properly for the distance.
Next time, he said, it would be different. That rematch hasn’t happened yet. But the comment lingers. Because it raises a vital and underexplored question:
How important is it that a horse has been trained for the trip?
How can we know whether a horse is being trained for a particular distance?
While some variables are public (past runs, ratings, jockey bookings), the goals of a training regime are largely hidden.
But race data can give us indirect clues. One such clue is the distance of a horse’s most recent run.
A simple assumption might be: if a horse ran over today’s distance last time out, it’s more likely to be prepped for today’s trip than one that did not. It’s not a guarantee, but it’s at least some observable signal.
This idea forms the core of our analysis: do horses who ran over the same distance last time out (“Same Distance”, or SD) perform better than those who did not (NSD)?
Method: Starting Simple
We looked at UK Flat Turf races, separating Handicaps and Non-Handicaps, as they behave differently.
Handicaps aim to equalise ability; non-handicaps often include unexposed or improving types, making distance prep potentially more revealing.
The core metric was a binary flag: same_distance, set to 1 if the horse ran over the same trip last time out.
This isn’t something we had to calculate, it comes ready-made from the historic_runners_insights table in Smartform, part of a suite of preprocessed indicators that simplify querying.
Here’s the SQL used to build the dataset:
WITH race_stats AS (
SELECT race_id,
COUNT(*) AS total_runners,
SUM(same_distance) AS same_distance_runners
FROM historic_runners_insights
WHERE race_type = ‘Flat' AND handicap = 1
GROUP BY race_id
)
SELECT h.*, rs.total_runners, rs.same_distance_runners,
1.0 * rs.same_distance_runners / rs.total_runners AS pct_same_distance
FROM historic_runners_insights h
JOIN race_stats rs USING (race_id)
WHERE h.race_type = ‘Flat' AND h.handicap = 1;
This gives us each runner’s record, whether it ran over today’s trip last time, and what proportion of its rivals did too.
Pulling this data into R, summarising win/place rates for same_distance horses was just:
sd <- full_data %>%
filter(same_distance == 1) %>%
summarise(
win_rate = round(100 * sum(won) / n(), 2),
place_rate = round(100 * sum(placed) / n(), 2)
)
We repeated this for non-same-distance runners – and here’s the resulting comparison:
| Race Type | Group | Runners | Wins | Win % | Place % | |
| Handicap | SD | 232,981 | 25,283 | 10.85% | 30.72% | |
| Handicap | NSD | 280,213 | 25,351 | 9.05% | 26.83% | |
| Non-Handicap | SD | 97,886 | 13,169 | 13.45% | 34.61% | |
| Non-Handicap | NSD | 190,105 | 18,168 | 9.56% | 26.76% |
The results are consistent and clear: same-distance horses outperform non-same-distance horses, and the effect is even stronger in non-handicaps. While it might not be conclusive proof of intent, it’s certainly correlated with preparedness.
Going Further: Race Composition
A useful extension is to examine whether this effect is sensitive to the composition of the field. In other words: does a horse’s same-distance prep mean more (or less) when many others in the field are similarly prepped? To explore this, we calculated the proportion of SD horses in each race, and assigned each race to a decile (from 1 to 10) based on that value.
So, decile 1 races had the fewest SD (same distance raced last time out) horses in the field, decile 10 the most.
Here’s a quick look at how we did it in code:
full_data <- full_data %>%
left_join(
full_data %>%
distinct(race_id, pct_same_distance) %>%
mutate(decile = ntile(pct_same_distance, 10)),
by = “race_id”
)
This allowed us to then group runners by both their SD/NSD status and the decile of their race, then calculate win and place rates for each subgroup.
Visualising the Effect
Let’s start with the overall win rate comparisons, for clarity:
The benefit of a same-distance prep is obvious in both race types. But things get more interesting when we examine how the decile of same-distance prep within the field interacts with performance.
Win Rate by Decile – Handicaps
Win Rate by Decile – Non-Handicaps
In both cases, the same-distance horses (blue lines) perform better than NSD horses across almost all deciles. In general, there are also higher win rates for SD horses when faced with less competition (lower deciles) of same distance last time out horses in the same field – particularly in non-handicap races.
What Does This Tell Us?
The concept of being “trained for the trip” is often speculative, something mentioned in post-race interviews or hinted at by analysts.
But here, we see a measurable and repeatable effect.
- SD horses are more likely to win than NSD horses.
- This effect holds in both Handicaps and Non-Handicaps.
- The effect exists regardless of how many SD runners are in the field.
- The effect is slightly more pronounced when fewer SD runners are present, and most strongly in non-handicap races
Final Thoughts
While no single variable is an end in itself, this analysis gives punters a tool to think in clearer terms about intent and readiness.
A same-distance last time out prep isn’t a guarantee, but it does suggest that the horse is being campaigned in a way that bodes well for today’s distance, which is often a significant unknown.
In future work, this could be extended to:
- Compare effect by course
- Account for recency of prep
- Account for last time out performance over distance
- Examine whether SD benefit varies by trainer
And of course, this same approach could be extended to Jumps racing, where distance variation is more extreme.
If nothing else, it’s a reminder that sometimes the simplest indicators — when measured properly — provide useful answers to the sport’s most opaque questions.
Colin Magee
