Expert Analysis

I come to praise Daryz, not to bury him, as Mark Antony in Shakespeare’s “Julius Caesar” did not quite say. But I feel I must point out that some of what was spoken and written about the colt in the aftermath of his second Arc victory at the weekend was an eensy weensy bit over the top. 

Yes, dual Arc winners are to be celebrated, undoubtedly. But they are also to be evaluated objectively and with a mind as to how history should properly view them.

The fact is that this year’s Arc had a remarkably bunched finish, with only about seven lengths between Daryz in first and Bright Light in last of 16. If Daryz is an all-time great on account of this performance, what does that make the others?

I think it is time for a brief run through of “How To Rate A Race”, numerically, before more subjective matters get factored in. Perhaps it will counter some of the more ill-informed remarks that handicapping is only ever “just a matter of opinion”. It doesn’t have to be.

Weight carried and ability (ratings) are expressed as pounds, so margins between horses need to be converted to the same and assimilated in order to establish the theoretical differences between the finishers by that measure.

Margins at the finish are time lapses converted for convenience into lengths (this has been the case in Britain for nearly 30 years now), in a manner which is neither consistent across jurisdictions nor necessarily even within them. You should, if possible, use the time lapses themselves, not someone else’s crude conversion of them to lengths.

There are different schools of thought on how this is best achieved, some of them remarkably basic. Ideally, calculations should involve not just time or distance, but time AND distance.

A horse winning at 5f in 55.0s should be treated differently to one winning at the same distance in 65.0s, never mind to one winning at 12f in, say, 150.0s.

The conversion I currently use is (83.33*DistInYards)/(OverallTime*OverallTime). The resulting “difference at the weights” for this year’s Arc can be seen in the following table.

Daryz was - based solely on the result at the line - in theory 1.7lb superior to Bay City Roller, 10lb superior to Minnie Hauk, and so on.

Next, race standards. We know what it usually takes in terms of ratings to win an Arc, finish second in an Arc, and so on, and we can apply that to the result in question. My ratings quoted in these blogs are on the level of WBRR/BHA, by the way.

An average Arc winner in the previous decade ran to 124, a second to 122, a third to 121, a fourth to 118, and a fifth to 115. You should apply a field-size adjustment to those figures (the more rivals you have to beat, the stronger the form, all other things being equal), but 16 runners this year was entirely normal, and so none is required.

An Arc winner would typically be rated 124, an Arc second beaten 1.7lb would imply an Arc winner with a rating of 123.7, an Arc third beaten 5.4lb would imply an Arc winner rated 126.4, an Arc fourth beaten 6.2lb would imply an Arc winner rated 124.2, and an Arc fifth beaten 9.1lb would imply an Arc winner rated 124.1.

You add the poundage beaten to the typical rating by-position for the race under consideration in order to get a figure for the winner, in other words.

If you weight those resulting figures by 1/finishing position you get a race standard of 124.3. The winner of an Arc with this overall time, weights carried, ages of participants, field size, and time margins between the principals, typically “should” be rated 124.3.

This may not have been a typical Arc, of course, and you can allow for that by considering the prior ratings of those principals.

In some races - those for two-year-olds, for instance - horses tend to improve on what they had achieved previously. In others, not so much. The Arc - for maturer horses, late in the campaign - is more in the latter category than the former.

Prior-rating calculations nudge up this year’s Arc, but only a little, to 124.5. That figure is the standard assessment of this year’s Arc result, based on tried-and-tested processes.

Those processes can be applied to all races, good and bad, here and elsewhere, with tweaks appropriate to specific context. They are algorithmic and repeatable rather than “just a matter of opinion” in other words.

Race standardisation, in some form or other, has been successfully deployed for at least 50 years now. Prior-ratings standardisation and sectional upgrading, which we will come to shortly, have been around for something like half that time.

To return to the 2026 Arc, that standardised assessment of the winner at 124.5 means also that Bay City Roller is worth 122.8 (he was 1.7lb worse than the winner at the line), Diamond Necklace is worth 119.1, and so on, all the way down to Bright Light in 16th on 114.0 (an improvement of 3lb on my previous rating for him).

If you are going to rate this year’s Arc on the bare result, you cannot pick and choose, ignoring horses who make it look worse in favour of those that make it look better. Daryz was an average Arc winner by that measure.

But you don’t have to stop there, and nor should you, even though others often do. The official handicappers, for instance, do not consider the evidence of sectionals in a consistent and across-the-board manner, despite occasionally claiming that they do.

Sectionals make Daryz’s 2026 Arc win look a fair bit better.

The race was run at a steady-ish gallop, one illustration of which is that Daryz got to halfway 1.0s behind when he did in last year’s well-run affair and yet ran 2.06s quicker overall (on what was faster ground). The resulting timefigure is at least 7lb below what you would expect an Arc winner to achieve.

Another illustration is that the winner’s 600m FS (speed in final 600m as a % of average speed for race overall) was 110.4%, where par is about 102.5%. Sectional upgrading results from the difference between those last two figures, ideally considering several sections and not just one.

The below is how things look for the Arc after the base ratings have been reframed to take into account the fact that sectional upgrades will boost all figures.

Two horses stand out as deserving the most upgrading, and they are Daryz and seventh-placed Varandir. Meanwhile, last-placed Bright Light reverts to a more credible figure.

Numbers can tell us only so much. Maybe a horse was hampered or boxed in, in a manner that is difficult to quantify, or maybe the jockey was motionless even at the line. None of those applies to Daryz in this instance.

A precise and objective numerical assessment of a horse’s performance still leaves scope for interpretation of what that means, in the wider scheme of things and for the horse itself. But such an assessment in the first place is crucial to understanding what went on.

My 128.6 for Daryz is higher than Timeform’s 128 basic rating (equivalent to 124 on my scale), if only a little higher than their 131 master rating (equivalent to 127).

It is also higher than the WBRR figure for the colt (126) going into an Arc, which precedent suggests they will take at face value, leaving that figure unchanged. That said, France Galop have boosted Daryz’s rating from the equivalent of 127 to 128. There could be some “decision by committee” friction there.

In summary, the numbers suggest Daryz was an “average” Arc winner judged strictly on the bare form, but a fair bit better than that one once the run of the race and the strength of his finishing kick have been allowed for. The latter is what handicappers should aspire to.

But is he an "all-time great” Arc winner, as some claimed in the heat of the moment on Sunday? I have my doubts, even now. Then again, they cannot all be Sea-Bird or Dancing Brave, can they?!