Every Hong Kong race, rated for you.
Every runner at Sha Tin and Happy Valley rated before the off, locked 4 minutes before each race, and sent to your phone on Telegram. Every result published, losses too.
By starting your trial you agree to our Terms.
£0 today, then £9.99/month · no card needed to start · cancel anytime · 18+
Cancel anytime from your account.
Next meeting
Sha Tin · Monday 19 October
11 races · first race 12:30 HK / 05:30 UK
Add to calendar: Apple / Outlook · Google
Free Race 1
See the product before you pay.
Sha Tin · Monday 19 October · Race 1 · 12:30 HK / 05:30 UK LIVE · not locked yet
| Top Rated | #1 Star Brose Model 21.5% | Market 13.6% · Odds 6.0 |
| Threat | #4 Kingly Demeanor Model 12.5% | Market 21.0% · Odds 3.9 |
| E/W 3rd | #2 Golden Fortune Model 19.0% | Market 10.8% · Odds 7.6 |
Live ranking: it updates as odds move and locks 4 minutes before the off.
Open the full Race 1 card Every race gets this treatment for members.
18+ | Gamble responsibly
Latest meeting
Happy Valley, Wed 14 Oct: how the ratings did.
| Won | Finished top 3 | |
|---|---|---|
| Top Rated | 4 | 6 |
| Threat | 0 | 2 |
| E/W 3rd | 1 | 4 |
The winner was in our top 3 in 5 of 9 races. Every result, winners and losers, is published.
18+ | Gamble responsibly
Members-only Telegram
Every locked rating, as it locks.
Sha Tin Sundays start early UK time. The rating comes to you.
No need to keep the race card open. Each rating lands in the channel within a minute of being locked — 4 minutes before the off — followed by the result and a short summary once the meeting’s done. Same ratings, same honesty: every race posted, win or lose.
Members only · Join the channel → · Want a taste first? The free channel gets Race 1’s locked rating and every result.
How it works
One race card. Three stages.
Early ratings
Every runner ranked on form, draw, jockey, trainer and running style — where they raced early, how much ground they made up late and how fast they ran the final 400m of their recent starts, not just where they finished — before the odds settle.
Live race card
Our model’s win % next to the live HKJC market %, blended into Top Rated, Threat and E/W top-3 tags. Where the model and market disagree strongly, we flag it.
Locked & recorded
Ratings lock 4 minutes before the off and are never changed. Each result is recorded against them.
What members see
A real race card, with the real result.
Sha Tin · Race 11 · Sunday 27 September 2026
| No | Horse | Model win % | Odds | Gap | Tags | Finished |
|---|---|---|---|---|---|---|
| 6 | CIRCUIT CHAMPION | 2.0 | -4.3 | TOP RATEDE/W TOP 3 | 2nd | |
| 3 | AEROINVINCIBLE | 5.8 | -3.2 | THREATE/W TOP 3 | 1st | |
| 7 | COMPLETE UNKNOWN | 6.9 | -2.7 | E/W TOP 3 | 12th | |
| 2 | THE RED HARE | 9.7 | -1.0 | 7th | ||
| 10 | FORZA TORO | 17.0 | +2.4 | 9th | ||
| 1 | LUCKY SAM GOR | 25.0 | +3.2 | 3rd |
Top 6 of 14 runners. The Threat won; the top-rated horse finished 2nd. Shown after the race with official placings; odds are final HKJC win odds. Gap = model % minus market %.
Our record, in full
Every rating on the record.
Races rated
51
Top Rated finished top 3
26 of 51
E/W top 3 picks placed
71 of 153
Top Rated winners: 7 of 51.
Who’s behind this
Hi, I’m Lee.
I’ve followed Hong Kong racing for years. A visit to Happy Valley, where I was lucky enough to be invited, turned that interest into a real passion.
I built HK Racing Model PRO because I wanted to show the full picture, not just the winners. Every rating is locked before the race and stays on the record, whatever the result.
I don’t guarantee winners. What I do is make race day easier: the race card does the legwork for you, with ratings for every runner, the main threats, each-way contenders, and how the model’s view compares with the market price — shown for information, not as a value signal.
Lee · @HKRacingPRO
Straight talk
These are model ratings, not tips or guarantees. From 7 October 2026, Top Rated blends our model with the market. Tested on 854 races of the 2025/26 season the model never trained on, the blended top pick won 28% of races — about the same as the favourite (28%); the model on its own won 23%. The test used final prices. Our ratings lock 4 minutes before the off, when the market is less sharp, so live results are expected to come in below the test. No version we have tested beats the market price, and we’ll say so plainly until one does. How we test