The Thursday Take is our weekly column: one racing talking point, put to the test against the Proform database — every British and Irish runner, every result. This week, after a Cheltenham weekend that was called Good and did not look it, a question that comes up every autumn.

The short version

  • The official going and the clock agree less than half the time. Of 33,608 jumps races declared Good since 2010, the race times said Good in 48%. One in four rode faster, one in eight two steps faster.
  • It leans with the season. A summer “Good” rides fast one time in three; a winter “Good” rides soft one time in three. On the Flat, “good to firm” rode slower than that 62% of the time.
  • It barely moves the price. On a Good day, horses that had already won on ground that rode fast returned A/E 1.02 at Betfair SP on 51,215 runs: above par in 12 years of 15, but worth about break-even after commission.
  • Knowing it rode fast adds nothing. Those horses read 1.02 on the days that rode fast and 1.03 on the days that rode good. The market moves with the ground on the day.
  • Trainers do not pull more horses on days that turn out fast. 9.2% of runners came out on those days, 9.9% on days that rode good. They react to the forecast and the walk, not the clock.

This week's talking point

Cheltenham's November meeting was declared Good all weekend. On the Saturday more than 20 horses were taken out for quick ground and the course watered; by Sunday morning there was an inspection for standing water. The official word did not change. When we read the ground from the winning times, every one of Saturday's races rode Good to Firm.

That raises an obvious question. If the official going and the ground can be that far apart, how often does it happen, and does it matter to anyone trying to find a winner? So we compared the two across every British and Irish race since 2010.

How to read the numbers. The official going is the description published before racing. The actual going is worked out after racing from the race times: how fast the winners ran against the standard time for each course and trip. It is a reading from the clock, so it is only known once the races have been run. Since 6 November jumps races have been read on a new scale (the old one called the ground Firm far too often), and every race in the archive was re-labelled, so every jumps figure here is on the new scale. A/E (actual over expected) is winners divided by the winners the Betfair SPs said to expect: 1.00 means the prices were right. Every runner counts, fallers and pulled-up horses included.

How often the official going matches the clock

Line the two up race by race and the gap is wide, and it is not random.

Declared going (2010 to 2026)RacesClock agreedRode fasterRode softer
Jumps, Good33,60848%27%25%
Jumps, Good to Soft16,75048%28%25%
Jumps, Soft21,28442%25%32%
Flat turf, Good26,52370%22%9%
Flat turf, Good to Firm25,26624%14%62%

A jumps race declared Good was Good on the clock less than half the time, and faster about as often as softer. One in eight jumps races declared Good rode two steps faster, Firm or quicker. On the Flat the bias runs the other way: a card declared good to firm rode slower nearly two times in three.

None of this means anyone got it wrong. The official going is a forecast, made in the morning with a stick and a pair of boots, and the ground changes during an afternoon. The clock is a hindsight measure, and it carries its own error: a course whose standard times are a little slow will always look faster than its going. Treat the gap as the size of the uncertainty, not as a score.

It leans with the season

The most useful pattern is the calendar. Here is the jumps “Good” by month.

Jumps races declared GoodRode fasterRode softer
July and August34-37%16-18%
October28%24%
November17%33%
December to February13-17%32-36%

In high summer a Good declaration rides fast one time in three. In winter it rides soft one time in three. November sits in between, and last weekend at Cheltenham was the summer pattern turning up in the autumn.

Courses differ too, though partly because of their standard times. On our record Fairyhouse, Clonmel and Sligo ride faster than declared most often; Tipperary, Thurles and Cheltenham least often. At Cheltenham a jumps race declared Good rode faster only 11% of the time, which is why Saturday stood out.

Does it change who wins?

Now the betting question. Take every chase and hurdle declared Good since 2012, and split the runners by whether they had already won on ground that rode good to firm or faster, and by how the race itself rode.

Declared Good: how it rode, and the horseRunsWinsA/EReturn at BSP
Rode fast, already won on fast ground15,6922,1801.02+2.1%
Rode fast, never won on fast ground50,0945,0960.99−7.6%
Rode good, already won on fast ground24,0493,2491.03+3.6%
Rode good, never won on fast ground100,23110,3570.99−6.8%
Rode softer, already won on fast ground11,4741,4150.99−9.0%
Rode softer, never won on fast ground50,8475,1721.00+1.8%

Two things stand out. First, horses already proven on fast ground do a fraction better than their prices on a Good day: put together, however the race rode, that is 51,215 runs, 6,844 winners, A/E 1.02, above 1.00 in 12 years of 15, and a return of about +0.3% at Betfair SP after commission. That is roughly 1.7 standard errors from par, a small and steady lean.

Second, and more important: knowing the ground rode fast adds nothing. The proven fast-ground horse reads 1.02 on the days that rode fast and 1.03 on the days that rode good. If the hidden fast ground were catching the market out, the first number would be the bigger one. It is not. The prices on the day already move with the ground, whatever the official word says.

So the gap between the official going and the clock is real, it is large, and you cannot make it pay simply by knowing about it.

Do trainers pull more horses when it rides fast?

Last Saturday's non-runners made the headlines, so we checked. On jumps cards declared Good since 2018, 9.2% of declared runners were taken out on the days that rode fast, 9.9% on the days that rode good and 9.6% on the days that rode softer. Trainers do not withdraw more on the days the clock later says were quick. They decide on the forecast and on walking the course in the morning, which is all they can see. Cheltenham's wave of withdrawals was a reaction to quick ground reported in advance, not the normal pattern.

Where to see the ground as it rides

The clock does not have to wait until the evening. On our Fast Results page, RaceMetrics Pro and Proform Premium members get a line for each British turf meeting saying how the ground is riding from the winning times, from the second race of the day. Tested on the 2026 season, it named the going the overnight figures settled on about 70% of the time on the Flat (the official going, 45%) and 59% over jumps (official, 41%). It is a description of the ground, not a betting signal, and the numbers above are why: the market moves too. Our daily going-accuracy posts on social media do the same job after racing, course by course.

Build it yourself

The actual going is a reading made after racing, so in the System Builder it can only be used to look back, never as a rule for finding today's runners.

  • Race Type is one of Chase, Hurdle, and Official Going is Good.
  • Open the Breakdowns tab and pick Actual Going. Every row comes out at A/E 1.00: the market prices the ground as it rides.
  • For a version you can use before racing, set Official Going to Good To Firm and add Going Wins (actual going) at least 1: horses that have already won on ground that rode good to firm or firm.
System Builder, chases and hurdles, 2018 to 17 November 2026RunsWinsA/EReturn
Declared Good, all runners157,41517,5041.00−1.2%
  of which rode Good to Firm18,2732,1431.00+0.2%
  of which rode Firm15,8921,8311.00−3.9%
  of which rode Good87,4109,6491.00−3.3%
  of which rode Good to Soft32,6783,5661.00−2.2%
Declared Good to Firm, already won on that ground1,4092591.01−6.6%
Declared Good to Firm, not yet4,0435371.00+14.8%

The System Builder starts in 2018, so these counts are smaller than the figures above. One limit to know: the going record rule looks at the declared going, so on a day declared Good it counts wins on Good ground, not on fast ground. There is no rule yet for “has won on fast ground” on a day declared Good. And note the last line: more profit with no more winners than the prices said, which is a couple of long-priced winners, not a pattern. That is why we read A/E first.

The take

The official going is wrong a lot: a jumps race declared Good is something else on the clock more often than not, and in a predictable direction for the time of year. But it matters far less than it seems, because by the off the market has caught up with the ground. Horses proven on fast ground are a small, steady lean on a Good day, worth knowing as a tie-break and not much more. If you want to use the gap, use it to read the ground during the afternoon, not to beat the price.

Last week we asked whether the race-fit horse wins on fast autumn ground. Next week: another talking point, another test.

Questions we get asked about this

How often is the official going wrong?

Of 33,608 British and Irish jumps races declared Good from 2010 to 2026, the race times said Good in 48%. 27% rode faster and 25% rode softer. Flat races declared good to firm rode slower 62% of the time.

What is the actual going?

A reading of the ground made after racing from the winning times, against the standard time for each course and trip. It is a hindsight measure and carries its own error where a course's standard times are off.

Does a Good declaration ride faster in summer?

Yes. In July and August a jumps race declared Good rode faster 34-37% of the time; in December to February 13-17%, and softer about a third of the time.

Is it profitable to back fast-ground horses when the going is Good?

Not by much. Horses that had already won on ground that rode fast returned A/E 1.02 on 51,215 runs on jumps cards declared Good (2012 to 2026), about +0.3% at Betfair SP after commission. Knowing the race rode fast did not improve it.

Do trainers take out more horses when the ground rides fast?

No. On jumps cards declared Good since 2018, 9.2% of declared runners were taken out on days that rode fast and 9.9% on days that rode good.