The First Month Is Not Like the Rest: What to Look For in Early-Season Hong Kong Racing
The 2026/27 Hong Kong season starts on Sunday 6 September at Sha Tin, the first of 88 meetings, with the HKSAR Chief Executive's Cup over 1200m promoted to Group 3 status and HK$4.2 million in prize money. It is the one day of the year when almost every runner on the card is returning from a break, almost every rating is stale, and almost every punter is working from the same thin evidence base.
That combination is exactly why the first month behaves differently from the other ten. Below is what last season's opening six meetings actually showed, and what the barrier trials run since the yards came back tell us about the horses likely to hit the ground running this time.
A note on the data. The race analysis here covers the first six meetings of 2025/26 — 7 to 28 September 2025, 55 races, 677 individual runs. That is a real sample but a small one, so treat the percentages as directional rather than settled. The trial data covers 20 to 31 August 2026: 484 trial runs by 454 horses across Sha Tin, Sha Tin all-weather, Conghua and the Sha Tin City Circuit. It does not extend back to mid-July, so any horse that trialled in late July or the first half of August isn't captured.
1. Some trainers really do start faster than others
This is the clearest signal in the data, and it holds up when you check it against where those trainers finished.
Here is the first-six-meetings record from last season, alongside each trainer's eventual 2025/26 total. Those first six meetings represented 6.8% of the season's 88 fixtures, so a trainer landing meaningfully more than 6.8% of their annual wins in that window is genuinely front-loading their campaign.
The top of that table is not a fluke of a hot fortnight. Fownes, Newnham and Shum each banked about one-and-a-half times their "fair share" of season wins inside four weeks, and all three were still fighting for the championship in July. These are stables that bring horses back fit.
The bottom of the table is more useful. John Size took 2 wins from 45 runners in the opening month and finished the season on 51. Francis Lui had 1 from 27 early and ended on 54 — under 2% of his season's wins in nearly 7% of its meetings. Manfred Man and Ricky Yiu show the same shape. These are not weak yards. They are yards that build, and their September runners are frequently underdone horses being educated rather than backed.
Two further notes worth carrying into September:
Pierre Ng went 0 from 41 across those six meetings. A former champion trainer with a full complement of runners and no winner in a month is the strongest single illustration of the point.
Jamie Richards was the outlier of the opening weeks, with 5 wins from 23 runners at a strike rate above 21% and a healthy return at the odds. The HKJC's own CEO blog flagged it at the time, noting he had produced five winners from only 15 runners and two doubles in three meetings after a difficult previous campaign. Richards did not sustain that rate across the season, which is itself the lesson: a stable coming off a poor year with a reorganised string can spring an early-season surge that the market is slow to price.
Practical read: in the first month, upgrade Fownes, Newnham, Shum and Hayes runners at face value. Downgrade Size, Lui, Man and Yiu runners unless there's a specific reason — a sharp trial, a proven first-up record — to think otherwise. And watch for a stable that has changed something over the winter, because those are the ones the market hasn't recalibrated for.
The riders follow the same pattern
Zac Purton went 10 from 48 across the opening six meetings, a 20.8% strike rate, and the club noted he had made his usual fast start with seven winners from 23 rides after three meetings. Hugh Bowman was 5 from 30. Alexis Badel had only 4 wins from 34 rides but returned a substantial profit at the odds, because his winners came at prices.
More instructive is who didn't fire. Several capable riders — Robbie Kingscote (0 from 32), Matthew Poon (0 from 35), David Probert (0 from 23), Keagan de Melo (0 from 19) — went winless through the entire month. Book strength matters more in September than at any other point in the year, because the strong stables are firing and the weaker ones aren't, and jockey bookings track that.
3. Almost nobody wins twice
What about horses that won more than once early in the season?
The answer, from 55 races and 555 individual horses, is exactly one.
Beauty Crescent (A S Cruz) won a Class 3 1400m at Sha Tin on 7 September off a mark of 68, leading throughout under Y L Chung at $8.10. Three weeks later, on 28 September, he came back over the same course and distance off 75 — a seven-point rise — settled just off the pace this time, and won again at $6.30.
That is worth sitting with, because it explains the whole shape of the early season. Only 119 of the 555 horses even got two runs inside those four weeks. Of those that did, 7 won. The handicapper is quick, and a September winner typically comes back carrying five to seven pounds more within a fortnight or two. Doubling up requires a horse whose improvement is running ahead of the assessment, not merely level with it.
What Beauty Crescent had going for him was that he won his second race from a different position in the run. He led first-up and sat handy second-up — a horse with tactical range, which is precisely the sort that copes when a rating rise means a tougher assignment.
The more reliable version of this angle is the second-up run.
Splitting horses by how they performed on their first appearance of the season:
Horses that finished second or third first-up won 19% of their next start. Horses that finished fourth to sixth won none of 24. The gap between "just missed" and "mid-pack" is far wider than it looks on a results page, and it is a cleaner filter than chasing the first-up winner into a higher weight.
4. Early speed is worth more in September than it is in March
This is the strongest structural bias in the sample.
Measured a different way: horses positioned in the first three at the opening sectional point won 14.8% of the time. Everything else won 6.8%. That is more than a two-to-one advantage on raw position.
The bias is sharpest on the Sha Tin all-weather, where leaders and on-pace horses won 5 of 18 starts while swoopers went 0 from 29.
It is mildest at Happy Valley, where the tight circuit compresses everything anyway.
Why it happens is straightforward. In September the fields are full of horses at 90% fitness who don't have a finishing sprint yet, the tempo is often modest because nobody wants to bury a first-up horse, and the ones with natural gate speed simply get a soft lead. By December, when everyone is race-fit, the pace goes on properly and the closers start hitting the line.
One caveat on the draw: horses from barriers 13 and wider won 6 of 40 in the sample, which looks striking. But that is 40 runs, concentrated in big-field Sha Tin sprints, and it's the kind of number that reverses on a larger sample. Treat it as noise until it repeats.
The market is sharp but not sharp everywhere
Favourites won 17 of 56 early-season races, a 30.4% strike rate, but lost about 16 cents in the dollar at starting price. The crowd is broadly right about who wins and slightly too keen on the top of the market when the form is thin.
Sorted by price:
The $5 to $9 band was the only profitable one, and the third-favourite specifically produced 13 winners from 52 starts. The dead zone is the $9 to $30 range: 6 winners from 256 runs. In an early-season race, a horse at $12 is usually there because the market genuinely can't see a case for it, and in September it's usually right.
Note also that older horses outperformed. Runners aged roughly six and a half or above won 12.3% of their starts against 7 to 8% for the rest. Seasoned campaigners have been through this cycle before and their trainers know exactly how much work they need.
6. Reading the trials from the winter break
The trials in the analysed data run from 20 to 31 August 2026 — eight sessions, 50 batches, 454 individual horses. Before the shortlist, three things about how to read them.
Batch quality distorts raw times badly.
In the 31 August Conghua Batch 6, the eighth horse home ran 1.40 seconds quicker than the median 1200m time at the venue, with a comment reading "limited response when asked." That batch simply went fast. Meanwhile the winner of the 28 August Sha Tin all-weather Batch 4 recorded a time nearly two seconds slower than par. Raw time against a venue standard is only meaningful once you know whether the batch was hot or slow.
The comment carries as much information as the clock. A horse recorded as being under a hold, hard held, or drawing away is telling you something the finishing position doesn't. Conversely, "led all the way to score" in a five-horse heat is close to meaningless.
Two trials beat one, especially when the second is better.
A handful of horses in the sample showed genuine improvement between trials — Russet Glow went from weakening after leading on 25 August to a comment of "moved better than before" and a much sharper time on 31 August. Circuit Champion went from racing wide and finishing eighth on 21 August to hitting the line strongly at Sha Tin on 30 August. That direction of travel is more useful than a single good hitout.
The standouts
Filtering for horses that were faster than the venue/distance/going standard, finished in the front half of their batch, and drew a positive comment, this is the group that stands out:
Clear class, doing it easily
Ka Ying Rising is the obvious one, and worth flagging that he won the corresponding Chief Executive's Cup at last year's opener at $1.00. Two trial wins in ten days, both with time to spare, in a horse chasing a 21-win streak. He is not a betting proposition at those odds, but he sets the tone for the meeting.
Fast, front-running, and suited by the September pace bias
Cross-reference this group against Section 4. Horses that showed gate speed in a trial and are drawn to use it are the ones the early-season bias most directly rewards.
Held together — the ones the market may underrate
This third group is the most interesting for betting purposes. Each ran a competitive time without being asked for full effort, or ran wide and still finished off. None of them "won a trial", so none will lead the trial-watchers' lists, but the underlying performance is stronger than the placing.
Also worth noting: Self Improvement, who won the G3 Korea Sprint for Manfred Man last season, trialled second on 21 August with a "fought on well; just missed" comment. And Jubilant Winner, one of Jamie Richards' early-season winners last September, trialled twice in the window and finished well on both occasions.
The obvious caveat: the trial file doesn't carry trainer or jockey details, so these names need to be cross-checked against the actual entries and stable when the fields come out. A sharp trial from a Size or Lui runner means something different from a sharp trial from a Fownes or Newnham runner, given everything in Section 1.
7. The early-season checklist
Putting it together, here is what actually moves the needle in September:
1. Stable form cycle first.
Fownes, Newnham, Shum and Hayes bring them ready. Size, Lui, Man and Yiu don't. This single filter is worth more than any other in the opening month.
2. Early speed over late speed.
Leaders and on-pace horses win at roughly 12.7%, swoopers at 4.6%. On the Sha Tin all-weather, closers won nothing at all last September.
3. Second-up off a placing, not off a win.
First-up placegetters won 19% next start. First-up winners face a rating rise; first-up mid-packers won nothing.
4. Price discipline.
The $5–$9 band was the only profitable one. Almost nothing won between $9 and $30.
5. Trials read for effort, not for placing.
A horse held together in fourth in a fast batch is a better prospect than a horse that won a slow one.
6. Respect experience.
Older, seasoned horses outperformed their younger rivals by a clear margin first-up.
7. Look for what's changed over the break.
New stable, new yard, reorganised string, a returning rider. Richards' September surge came directly off a winter of upheaval, and the market took weeks to catch up.
None of these signals is individually secret. Gear changes are published, trial times are public, and stable form is on every racecard. The edge in the opening month comes from combining them quickly enough to act before the market consolidates — a horse with a sharp held-together trial, natural early speed, an in-form September stable and a price in the single figures is a far stronger proposition than any one of those factors alone, and confirming all four across a full card is a research job rather than a hunch.
That's the exact groundwork the HK Speed King Race Lab, Pace Lab and Trial Lab engines are built to do: a systematic, multi-factor read across trial evidence, sectionals, pace projection, stable form and market signals for every runner on the card, ready before the first race jumps.
Sources and further reading:
- [2026/27 season opening, 6 September 2026 — Hong Kong Tourism Board](https://www.discoverhongkong.com/eng/events/event.id87782.season-opening.html)
- [Excitement builds ahead of opening to 2026/27 season — HKJC CEO Blog](https://ceoblog.hkjc.com/19487)
- [Caspar Fownes wraps up Hong Kong trainers' title — The Straight](https://thestraight.com.au/caspar-fownes-wraps-up-hong-kong-trainers-title/)
- [Incredible 2025/26 season approaching its close — HKJC CEO Blog](https://ceoblog.hkjc.com/19297)
- [Hong Kong racing season continues to unfurl with excellence and achievement — HKJC CEO Blog](https://ceoblog.hkjc.com/17728)
- [Hong Kong's 2025/26 season sees record overseas betting turnover — Thoroughbred Daily News](https://www.thoroughbreddailynews.com/hong-kongs-2025-2026-season-sees-record-overseas-betting-turnover-of-hk34-4-billion-3-2-billion/)
- [Hong Kong 2025/26 season preview: trainers to follow — Idol Horse](https://idolhorse.com/horse-racing-news/hong-kong/hong-kong-2025-26-season-preview-trainers/)
Race and trial data: first six meetings of the 2025/26 season (7–28 September 2025) and barrier trials 20–31 August 2026
-Trial comments courtesy HKJC
Some interesting insights from the King🚀🏇