Horse racing ratings are numbers that put every runner's ability on one scale, so you can compare horses that have never met. In Australia you will meet three kinds: the official handicap rating a state's handicappers set, speed ratings built from the clock, and private ratings that form analysts build from margins, weights and class. In NSW and Victoria one official point is worth 0.5kg, which is how ratings become weights in a benchmark handicap.
A rating only becomes useful as a price, and most of the judgement sits in that step. The form fields a rating summarises are laid out in how to read a form guide.
Official ratings, benchmarks and speed ratings compared
| Kind | Who sets it | What it measures | Units | Its weakness for pricing |
|---|---|---|---|---|
| Official handicap rating | The state's handicapping panel | The merit of what a horse has achieved | Points, 0.5kg each in NSW and Victoria | Built to set weights, so a handicap's weights cancel it |
| Speed rating | Form services, or you | How fast the horse ran, adjusted for track, distance, rating and weight | Points or lengths against a standard | Pace distorts the clock: a slowly run race gives good horses low figures |
| Private rating | Form analysts, or you | How well the horse ran against its rivals, from margin, weight and class | Usually points or kilograms | Each analyst's scale differs, so numbers do not cross between systems |
The official number is printed as the Hcp Rating on Racing Australia's form. Racing NSW's handicapping guidelines say a horse's rating is assessed within 48 hours of each run (in maiden races, only the winner is assessed). A win generally brings a rise of 1.5kg to 2.5kg (3 to 5 points), and an uncompetitive run may cost 0.5kg to 1.0kg.
Racing Victoria's handicapping page puts a winner's usual rise at 1.5kg to 3.5kg (3 to 7 points). Racing NSW's ratings assume a mature male; a filly or mare, or a younger horse, receives its allowance as a lighter weight rather than a lower rating, which handicap races explains.
Benchmark races turn the scale into weights. Racing NSW's benchmark system, introduced on 1 October 2009, gives each point 0.5kg of handicap weight. The weights Racing Australia published for Dubbo on 2 October 2026 show it at work. In a Benchmark 58 handicap, a gelding rated 58 was allotted 61kg, one rated 59 got 61.5kg and a mare rated 58 got 59kg, 2kg less. Benchmark races covers how the BM number sets each field.
Two more properties matter for pricing. Official ratings look backwards: Racing NSW's benchmark policy says merit-based handicapping reviews past performances and "does not anticipate improvement in performance". And Racing NSW's handicappers leave apprentice claims out of rating changes, because of the junior rider's relative inexperience.
A speed rating starts from the clock. Grade each run's time against a par for its track, distance and track rating, the way sectional times grades the last 600m, convert the gap to lengths, then allow for the weight carried. Pace moves the clock, so rate several runs before you trust a figure: the same horse can post a low one in a slowly run race and a high one in an evenly run race.
Why official ratings cancel out in a handicap
Racing NSW's guidelines describe the handicapper's job as allocating weights "with the purpose of equalising their chances of winning". So in a benchmark handicap the weights are the ratings in another form. Adjust each rating to a common weight, 2 points for every kilogram above or below it, and every mature male above the minimum weight comes out level. A mare or young horse comes out ahead by its allowance unless you add it back: the Dubbo mare adjusts to 58 + 2 x (61 - 59) = 62, 4 points clear of the gelding rated 58. In an illustrative Benchmark 76 for three geldings, using the 61kg-at-the-benchmark base the Dubbo weights show:
| Runner | Official rating | Weight in the Benchmark 76 | Rating adjusted to 61kg |
|---|---|---|---|
| A | 80 | 61 + (80 - 76) x 0.5 = 63kg | 80 - 2 x (63 - 61) = 76 |
| B | 72 | 61 - (76 - 72) x 0.5 = 59kg | 72 + 2 x (61 - 59) = 76 |
| C | 70 | 61 - (76 - 70) x 0.5 = 58kg | 70 + 2 x (61 - 58) = 76 |
That is why the official number, on its own, says little about who wins a handicap. It earns its place in three cases:
- A horse below the minimum weight. Racing Australia's weights pages print a True Weight beside the Weight. At Dubbo on 2 October 2026, a gelding rated 58 in a Benchmark 82 had a true weight of 49kg but carried the 55kg minimum: 6kg, or 12 points, more than its rating asked.
- Weight-for-age and set-weight races, where weights do not follow ratings, so a rating gap stays a real gap.
- Your own rating disagreeing with the handicapper's, which is where a handicap leaves room for a view.
What a ratings point is worth
| Measured in | One point is worth | Basis |
|---|---|---|
| Weight | 0.5kg | The official scale in NSW and Victoria |
| Lengths | About a quarter of a length to a full length | If a length is worth roughly 0.5kg over staying trips to 2kg over sprints: 0.5 / 2 = 0.25 and 0.5 / 0.5 = 1.0 |
| Chance | It depends on the field and the scale you choose | The table below |
Weight in horse racing explains why kilograms per length is a rule of thumb, never a fact. For chance, take a runner at 20% ($5.00) and give it one more point while its rivals stay put. If each point multiplies a runner's strength by a scale factor, its new chance = 0.20 x scale / (0.20 x scale + 0.80):
| Scale per point (illustrative) | New chance | New fair price | Change |
|---|---|---|---|
| 1.08 | 0.216 / 1.016 = 21.26% | $4.70 | +1.26 percentage points |
| 1.15 | 0.230 / 1.030 = 22.33% | $4.48 | +2.33 percentage points |
| 1.30 | 0.260 / 1.060 = 24.53% | $4.08 | +4.53 percentage points |
As a rule of thumb, then, a point is worth something between about 1.3 and 4.5 percentage points to a $5.00 chance, depending on how steep your scale is. Your own results, not a guess, should settle which scale fits.
How to convert ratings to odds in five steps
- Put every runner on today's terms. Adjust each rating for the weight it carries (2 points a kilogram), adding back any sex or age allowance if your ratings, like the official ones, assume a mature male. Then allow for anything your ratings leave out, such as the distance, the track rating or the speed map.
- Turn the gaps into strengths: strength = scale to the power of (points above the lowest-rated runner).
- Share 100% by strength: chance = strength / total of all the strengths.
- Take the fair price, 1 / chance. That figure is your rated price.
- Compare it with the market, and keep a record so you can test the scale over many races.
How to frame a market frames an eight-runner race with a points-to-chance chart instead of a scale, then tests the prices against the exchange. Ratings-based prices are weakest where the ratings are thinnest: maidens full of lightly raced horses, first starters and horses resuming from a long break, so price those races loosely or pass them.
Worked: three runners rated and turned into a 100% market
Example: An illustrative three-runner Benchmark 76 handicap, not a real race, kept small so every sum shows. The weights come from the official ratings in the table above, and the ratings below are your own, on the same 0.5kg-a-point scale.
| Runner | Your rating | Weight | Adjusted to 61kg | Points above the lowest | Strength at 1.15 a point | Chance | Fair price |
|---|---|---|---|---|---|---|---|
| A | 78 | 63kg | 78 - 2 x (63 - 61) = 74 | 0 | 1.0000 | 1.0000 / 3.4725 = 28.80% | $3.47 |
| B | 72 | 59kg | 72 + 2 x (61 - 59) = 76 | 2 | 1.15 x 1.15 = 1.3225 | 1.3225 / 3.4725 = 38.08% | $2.63 |
| C | 69 | 58kg | 69 + 2 x (61 - 58) = 75 | 1 | 1.1500 | 1.1500 / 3.4725 = 33.12% | $3.02 |
The strengths total 1.0000 + 1.3225 + 1.1500 = 3.4725, and the chances add to 28.80 + 38.08 + 33.12 = 100%. Each fair price is 1 / chance, such as 1 / 0.3808 = $2.63 for Runner B. You rate the topweight 2 points below the handicapper and Runner C 1 point below, so on your numbers Runner B, not Runner A, is the most likely winner.
Now the market. With illustrative prices from Bookmaker A of $2.40, $2.90 and $3.30, the market adds to 41.667% + 34.483% + 30.303% = 106.453%. Scaled back to 100% proportionally, the multiplicative option in the fair odds calculator, it rates A at 41.667 / 106.453 = 39.14%, B at 32.39% and C at 28.47%.
| Runner | Price | Your chance | Edge (price x your chance - 1) | Expected value of $10 |
|---|---|---|---|---|
| A | $2.40 | 28.80% | 2.40 x 0.2880 - 1 = -30.9% | 10 x -0.309 = -$3.09 |
| B | $2.90 | 38.08% | 2.90 x 0.3808 - 1 = +10.4% | 10 x 0.104 = +$1.04 |
| C | $3.30 | 33.12% | 3.30 x 0.3312 - 1 = +9.3% | 10 x 0.093 = +$0.93 |
The scale decides how big the edge looks:
| Scale per point | Runner B's chance | Fair price | Edge at $2.90 |
|---|---|---|---|
| 1.08 | 1.1664 / 3.2464 = 35.93% | $2.78 | 2.90 x 0.3593 - 1 = +4.2% |
| 1.15 | 1.3225 / 3.4725 = 38.08% | $2.63 | +10.4% |
| 1.30 | 1.6900 / 3.9900 = 42.36% | $2.36 | 2.90 x 0.4236 - 1 = +22.8% |
Every scale puts the $2.90 above your fair price for Runner B, but the edge runs from 4.2% to 22.8%, and even the flattest scale's 35.93% sits well above the market's 32.39%. Before backing a gap that size, ask what the market might know that your ratings do not, such as a gear change or a rider booking.
Risk: Betting involves risk. Your rated price is only as good as the scale and the ratings under it, and a runner you price at 38.08% still loses about six races in ten. There is no guarantee of profit; see responsible gambling for limits and support.
Mistakes that turn ratings into bad prices
- Backing the top-rated runner in a benchmark handicap because of its rating, when the weights were set to level it with the rest.
- Mixing scales: a private 80 and an official 80 are different numbers.
- Forgetting today's weight, or treating claims one way in one race and another way in the next. Pick a rule, such as Racing NSW's practice of leaving claims out of re-rating, and keep it.
- Treating a one-point gap as precise, when a single win can lift an official rating by 3 to 7 points.
- Forgetting that ratings look back, so a lightly raced or improving horse can be better than its number.
- Never testing the scale. Compare your rated chances with how often runners at each chance won over a few hundred races, and see sample size in betting for how many you need.
Ratings are your input: where B337 fits
B337 does not rate runners: it has no ratings engine, no model and no selections, so the ratings and rated prices are yours. What it shows is the market you compare them with. On the Terminal each bookmaker has its own column, 40+ of them across racing and sports, and the best available price on each runner is picked out, with Betfair back and lay, price history (flucs) and closing lines alongside.
On racing, the Terminal's EV overlays show expected value against an estimated fair price based on Betfair's market, a yardstick separate from your ratings, so the overlay and your rated price can disagree about the same runner.
Because B337 reads prices on a repeating cycle, what you see can lag a bookmaker's own price, so check it in your account before betting; expect some bookmakers not to price every race. A free account opens a limited view of the Terminal with live odds. Betfair is a trade mark of its owner, and B337 is not affiliated with it. For free and confidential support call 1800 858 858 or visit gamblinghelponline.org.au.
Risk: Betting involves risk. A rating describes past runs, the market can know things your ratings miss, and past results are no guarantee of future results. See responsible gambling for limits and support.