The Rating: Why You Need One Number

Times lie, horses don't — and what to do about it

Every horse crosses the finish line at a certain time. The temptation is to say: faster time = better horse. The temptation is wrong.

Times lie. Horses don't.

Theory: why one number

The same track on Monday and Wednesday can be seconds apart in speed — rain, temperature, wind, rail position, ground preparation, a dozen other things. A horse running 1:10.2 for six furlongs on a fast Monday might be running a 115-rated race; the same time on a slow Wednesday might be a 125-rated race. Use raw times and you get the wrong answer, every time.

Think about your own running for a moment. You jog your usual two-mile loop in 22 minutes through deep sand at high tide, then 18 minutes over packed sand at low tide. You didn't get fitter in four hours. The ground changed. A racehorse's clock is the same: most of what separates a 1:10.2 from a 1:11.4 is the day and the surface under it, not the animal on top.

This is why we rate horses, not clocks.

The projection method — horses as measuring instruments

The MWP rating uses the projection method, inspired by Jerry Brown's Thoro-Graph and refined over 20 years of work across Scandinavia, the Middle East, Asia, and Europe.

The principle: when you have a group of horses with established ratings from prior starts, those horses become measuring instruments for the race. The margins between them, combined with their prior ratings, calibrate today's race as a best-fit regression across all runners simultaneously.

Simplified example: Horse A has run 110, 110, 111, 110. Horse B has run 113, 113, 113, 112. Today B beats A by a margin worth about 3 points. Without looking at the clock at all: B ran ≈113, A ran ≈110, and a newcomer who finished 2 points behind A earned ≈108. The race calibrates itself.

What the clock is actually for: a starting point for calibration (subordinate to the projection), a mild sanity check (when projection and clock disagree, projection wins), and the points-per-length conversion, which varies with distance — a length is worth more in a sprint than in a staying race.

And none of this is a teaching exercise. The MWP rating is the number I handicap on myself, every day, across the jurisdictions in this course — what you are learning to read is the real working figure, not a simplified version of it.

Five principles of the MWP rating

1. Track speed is everything. The same track on different days is a different surface. The system reads the whole race day together.
2. Weight is baked in. A 60kg horse matching a 54kg horse's time has run the better race. The rating reflects it.
3. Turf and dirt are completely separate. Effectively different sports. All-weather and dirt are grouped (the data supports it).
4. All horses in a race get the same adjustment. Never adjust one without adjusting all — best fit for the entire race.
5. Historical ratings are revised. A known error contaminates everything downstream; fixing it compounds in accuracy.

The rating scale

The single most important property: 120 at Sha Tin = 120 at Ascot = 120 at Meydan. Surface-normalised, track-speed-normalised, weight-normalised — that is what lets you compare horses across countries and continents.

What the classes actually rate — Britain

The scale above is the abstract version. Here is what it looks like in the wild. These are real production numbers: every runner's MWP rating, averaged by the class of race it ran in, across Britain & Ireland from 2022 onward. (Hong Kong, the other pool this course leans on, gets the same treatment in the next chapter, where you meet your first Sha Tin card.)

Britain & Ireland (turf)

Britain & Ireland (all-weather)

The classes overlap. A good Class 4 winner (124) rates the same as an average Listed runner (126), which is why a well-handicapped horse stepping up in class can still win. The winner's figure sits about 10 points above the average runner in the same race, because the winner is usually the best horse having its best day. And the ceiling is real: the best Flat horses each year push past 155. The "champions push past 155" line in the scale above is not rounding up.

In Britain, Class 1 is the top tier and the figures fall as the class number rises, with Class 6 at the bottom. Hold onto that. It is not the same everywhere: Australia, later in the course, numbers its classes the other way round.

You do not have to memorise any of this. The point is the opposite: when the Handicapper hands you a horse's expected rating — the figure the model projects for today — you now know what that number means in class terms, and whether the horse is rated like the company it is keeping. Always start from that expected rating and adjust it with your own read, rather than guessing a number cold.

It is not an official rating

A warning, because the numbers look familiar: the MWP rating has nothing to do with official handicap marks. Every jurisdiction runs its own scale — UK Official Ratings, Australian Benchmarks, French valeurs, German GAGs, Scandinavian form numbers — and they don't agree with each other. Some are denominated in kilograms, some in pounds, Japan doesn't publish one at all, and even two scales that look alike are rarely calibrated against each other: a Scandinavian 90 and a German 90 are not the same horse in practice.

The MWP rating is a single universal scale built so a horse in Hong Kong and a horse in Sweden end up in the same currency. 100 sits roughly at "the most typical racehorse" — but the base number itself is arbitrary.

So if you also follow official marks, keep them in a separate drawer — they answer a different question. An official rating is a broad classification of a horse's ability, usually expressed as a weight (in kilograms or pounds) and set by the official handicapper. Its main job is narrow: to assign fair weights in handicap races, so the better horse carries more. It's useful for judging whether a horse has been treated leniently or harshly by the handicapper, and as a rough gauge of its class — but it is not a performance figure.

The MWP rating is. It measures each race on its own merit — how fast, and how good, that particular run actually was. The two numbers will broadly track each other, of course, because better horses run faster on average. But only a performance rating lets you follow a horse race by race: its improvement, its form cycles, the conditions that bring out its best and the ones that don't. (That reading becomes second nature once you've spent a little time handicapping with the scale — it stops being a number and starts being a story.)

It's worth looking through both lenses. Think of the official mark as the view zoomed out — a horse's broad class at a glance — and the performance rating as the view zoomed in, where the development and the detail come into focus. Just don't treat them as the same scale: a 120 MWP rating is not an attempt to line up with a 120 official mark — and as we saw above, those official marks don't even line up with each other from one country to the next.

Distance and weight — already in the number

Two horses can run the exact same time and still not have run the same race. One may have carried more weight; one may have done it over a different trip. The rating settles both, translating each into lengths before it converts to points. The MWP tool is built so you never need to remember any of this — every rating is already adjusted for the weight carried and the distance run.

Still, three rules of thumb are worth carrying in your head, for the times you want to size up a race before you open the tool. They hold up well at any track, in any country.

1. Weight costs more lengths the further it is carried. A horse carrying more weight than a rival — "giving it weight", in racing's phrase — and still finishing level has run the better race. And a kilogram does more damage the longer the trip: it costs about half a length in a sprint, building to a full length at a staying distance.

This barely changes from one track to the next — it is mostly just the geometry of the distance.

*2. A length is worth more rating points in a sprint. A length of daylight on the track is always the same gap physically — but it is not a fixed amount of ability. In a sprint the field is bunched and one length is a real gap; over a staying trip that same length is only a small slice of the race. So when a length is expressed on the rating's points scale, it is worth more points the shorter the trip* — very roughly, a sprint length is worth about double a staying length in points. (The exact figure shifts with the track — a faster surface packs more points into each length — so hold onto the direction, not a number.)

The flip side is the habit worth building: the points are the fixed currency; the visible margin is not. A 2-point gap is the same 2 points of ability whether the race is 1200m or 2400m — but it shows up as a tighter finish in the sprint and a wider one over the staying trip. When you size up a race, trust the gap in points, not the gap in lengths.

3. The one number to keep: about 1¼ points per kilo. The two effects above pull in opposite directions — a length is worth fewer points as the trip lengthens, while a kilo is worth more lengths — and they very nearly cancel out. What is left is the single most useful figure in this chapter: a kilogram is worth roughly 1¼ rating points, at more or less any distance and any track. A horse that carries 4kg more than a rival and only dead-heats has run about 5 points the better.

And if this section felt fiddly and hard to keep straight — you are right, and here is the relief: you never have to do any of it. Every number in the MWP form, and everything the Handicapper works out for you, already has the distance and weight conversions baked in and calculated. Your job is the part the machine cannot do — decide what you think each horse will run today, from your own form study, the statistics you read, and whatever else you use to sharpen the estimate. The conversions in this section run quietly underneath all of that. They are explained here only so you understand what the numbers already mean.

[QUOTE: "Times lie. Horses don't."]

Recommended reading

• [LINK: "Alan's piece on the limits of time-based methods" — https://www.formnerds.com.au/blog/6f30024f-6c43-490b-8435-0d9cc965ae65 ; "Why the clock misleads even when it is accurate — pace, wind, runups, gate placement and track bias. The deep version of 'times lie' — read this one first."]

Reading a sequence: the new top and the bounce

A single figure is a measurement; a run of them is a story. Two shapes matter most: a new top — a figure clearly above anything the horse has done before — and the bounce, the regression that often follows it.

Here is the part most form books get wrong, and the data is blunt about it. The figure does regress: a horse that jumps to a career best usually runs a bit below it next time. But that is a fact about the number, not about the bet. Across the full MWP database, horses coming off a new top and sent off at a short or medium price out-perform the crowd's own estimate — they win about 5% more than the market's margin-free probability, and the gap has grown in recent seasons. (A flat bet on them still loses the takeout, like any short-priced bet; the point is they are underrated, not that they print money.) The crowd over-fears the bounce and shades the price too far. A new top from a horse on the way up is usually a real new level, not a fluke to fade.

So read a new top two ways at once: expect the figure to ease back, but don't assume the horse is a lay. The only new-top runners that reliably lose are the ones the market has already pushed out to long odds. Keep the reading lesson (the number regresses) separate from the betting lesson (the short-priced ones are, if anything, underbet).

Practical: read the pattern

A single rating is a measurement. A sequence of ratings is a story. Work through the rating sequences below and say what each horse is telling you.

Quiz: rating theory

More practice cases

Real horses, real rating sequences — what a "new top" looks like, and how the figure regresses afterward even when the horse stays well worth its price.

Reading our form going in (Horsepower): a 10-point leap to a career-best 141.8, now 3.5 to follow up. Expect the figure to ease back — but the live danger is LA FORZA, an unbeaten made-all type alongside. Read it as "the 142 won't repeat," not "lay him": across the database, short-priced new-top horses are, if anything, underbet.

Reading our form going in (Mr Energia): the 143.0 is a career best, ~5 points clear of his next-best (138.1), and he's 2.8 stepping up to 1650m on dirt in a deep handicap. The price asks you to bank on the outlier repeating, with no margin if it doesn't.

Reading our form going in (Blazing Wind): only a +3 new top — but sent off 1.7. Here the trap is the price, not the jump: at 1.7 you're paying for a career-best to repeat with zero safety margin in a 14-runner sprint.

Up next: the rating is one lens on a race. The next chapter is about handicapping as a craft — using that lens, knowing its blind spots, and starting to find your own.