Six Point Nine Per Cent
That's the share of Australian racehorses that ever see a Group 3 race or above, across two full seasons of data. The previous piece in this series found that the median horse doesn't cover its training costs below Group 3. This one asks the obvious follow-up: how many horses actually get there?
Not many. And the ladder between maiden and Group 3 is steeper, stickier, and more final than the prizemoney tables suggest.
The pyramid
The class distribution of Australian racing is bottom-heavy and stable. In any given season, 30% of runners never race above maiden level. Sixty-three per cent never get above Class 2. Only 9.3% reach black-type company (Listed or above), and 5.7% contest a Group 1, 2, or 3 race. The shape is identical across both seasons, which means it's structural, not a one-year artefact.
Put it this way: pick a horse at random and there's a one-in-three chance it's a maiden and a one-in-seventeen chance it's ever seen a Group race.
The funnel
Of all the horses that ran in a maiden race in the 2024-25 season, 57% had won at least one race by the end of 2025-26. The median winner needed four starts to break through. That sounds healthy until you follow the funnel.
Of last season's maiden runners, 1.5% reached Group 3 or above within a year.
The rest either won their maiden and plateaued (the largest group), lost their maiden and kept running (the second largest), or left the system entirely. The funnel narrows at every rung, and it narrows fastest between Class 2 and Listed. Three classes on paper. That's where most racing careers hit their ceiling.
The equilibrium
The typical class jump from one season to the next is zero. Across all horses that raced in both seasons, 39% moved up at least one rung, 37% stayed at the same level, and 23% fell. The point where upward and downward movement cancel out is BM 70, which is Class 3-4 in NSW terms.
Above BM 70, more horses fall each season than climb. Below it, more climb than fall. Not a coincidence. Class 3-4 is where the population settles because it's where the economics settle. Enough prizemoney to justify continuing if your costs are low. Not enough to justify paying a professional trainer.
The stuck
Two thousand, two hundred and eighty-two horses raced in both seasons and never got above a maiden in either. Ninety-two per cent of them were winless in the second season. They netted $690 per start.
Add the 1,021 horses stuck at Class 1 across both seasons, and that's roughly one runner in nine on the national programme going around at the same level, season after season, earning a fraction of their costs. These are the horses that make piece 1's Lorenz curve sag. There are a lot of them.
Where Group 3 horses come from
If the funnel is steep from the bottom, where do the horses at the top come from? A third of each season's Group 3 field is new to the dataset: young horses arriving straight into stakes company, typically well-bred two and three-year-olds whose connections never intended them for the lower rungs. Under 10% were a maiden or Class 1 horse the year before.
The path from maiden to Group 3 exists, but it's rare and when it happens it's fast. One explosive season, not a slow multi-year grind. The grinders mostly plateau at Class 2-3 and stay there.
The age ceiling
The data confirms what piece 3's earnings-by-age curve suggested: the climbing stops at five. Half of two and three-year-olds move up the ladder the following season. Forty per cent of four-year-olds do. At five, only 30% climb, and 30% fall, an even split. By six, just 22% are still moving up. The median age at every black-type rung is four. The median age at every benchmark rung is five.
A horse that hasn't got there by five almost certainly won't. The window is roughly three seasons, debut to age four. After that, the ladder is set and the economics are locked in with it.
What this means for piece four
Piece 4 said the median horse breaks even at Group 3. This piece says 93.1% of horses never get there. Put those two findings together and you have the structural reality of racehorse ownership. The vast majority of horses are running below the break-even line, with no realistic path to reaching it, on a ladder where the typical annual movement is zero.
The 6.9% that reach Group 3 are not the median horse. They're the exceptions the sires' premiership ranks by, the ones the average-earnings figures are built on, and the ones the syndication prospectus shows you. The other 93.1% are the reason the industry runs a $273 million owner deficit, and the reason this series started with a Lorenz curve.
Notes
Two seasons (2024-25 and 2025-26), all Australian thoroughbred races, same dataset as pieces 1-4. Class rungs defined by highest benchmark, restricted-class, or black-type race contested in the season (not won). Open and Quality handicaps are unranked for progression purposes due to 2YO/3YO set-weight classification issues (891 two-year-olds classified as "Open" by racing.com; documented in the analysis). Transition matrix covers horses that raced in both seasons. "Stuck" means same highest class in both seasons. Maiden graduation rate is 57% across the two-season window (horses that ran in any maiden in 2024-25 and won at least one race by end of 2025-26). Exit from the dataset between seasons is not necessarily retirement. All analysis is my own; queries and figures available to anyone who wants to check them.