2025 Triumph Hurdle

[darlojim]

You’re right, the 3rd and 4th had solid form and they were used to produce the race ratings, and Stencil and East India Dock were far too good for them.

I also check out Timeform ratings too, the base rating rather than the race by race rating,

They’ve got

TIMEFORM up to 16th Feb inclusive

East India Dock 147p
Nietzsche Has 145 (retired)

Hello Neighbour 137p
Gibbs Island 134p
Stencil 133
Naturally Nimble 132p
Sauvignon 132
Willy De Houelle 131
Sony Bill 130

Lulamba 129P
Galileo Dame 129p (+ 7lb allowance = 136p)
Lady Vega Allen 128p (+7lb allowance = 135p)

Bacchanalian 127 (before todays easy win)
Murcia 127 (+7lb allowance = 134)
Saint Lucie 127 (+7lb allowance = 134)

Live Conti 125P

Sauvignon rated higher than Lulamba tells you all you need to know about ratings

[The Giant Bolster]

Do we think Murcia is definitely going to the Boodles? She finished a neck behind 133 rated Bacchanalian off level weights. Lady Vega Allen is rated 132 and has been beaten by Hello Neighbour twice.

If you draw a line through Murcia's seasonal debut (and the betting beforehand suggested not much was expected), doesn't she have more upside in the Triumph than most Mullins runners (including Lady Vega Allen)? She's very big on the BFX for the Triumph.

[ComplyOrDie]

Sauvignon rated higher than Lulamba tells you all you need to know about ratings

Haha, even as a ratings fan, although don't use Timeform much, this made me laugh.

How on earth have they come up with that?

[darlojim]

Haha, even as a ratings fan, although don't use Timeform much, this made me laugh.

How on earth have they come up with that?

Bonkers stuff, really is. Which is why its better just to form your own opinion on them.

Gibbs Island 134 - I had to fucking google who that yoke was

Yeah let him give Lulamba 5lbs jesus wept, he'd murder him

Regardless of all the 'P' Bollocks aswell..

Oooo hes got a little P, ahhh he's got a big P

Get in the bin

[Quevega]

Bonkers stuff, really is. Which is why its better just to form your own opinion on them.

Gibbs Island 134 - I had to fucking google who that yoke was

Yeah let him give Lulamba 5lbs jesus wept, he'd murder him

Regardless of all the 'P' Bollocks aswell..

Oooo hes got a little P, ahhh he's got a big P

Get in the bin

It's more bonkers if you don't understand the ratings system you're crabbing though.

[darlojim]

It's more bonkers if you don't understand the ratings system you're crabbing though.

I don't use them

Go on....

[Quevega]

I don't use them

Go on....

Neither do I, but this topic comes up every now and then on here.

Timeform have a base rating, which essentially is each horses best achieved rating, which it might never get near again.

Sauvignon will have got that for his listed win in France.

[darlojim]

Neither do I, but this topic comes up every now and then on here.

Timeform have a base rating, which essentially is each horses best achieved rating, which it might never get near again.

Sauvignon will have got that for his listed win in France.

So pretty useless unless they've ran that rating very recently no?

And point still stands with Gibbs Island as I ASSUME his latest run was his best

[Quevega]

So pretty useless unless they've ran that rating very recently no?

And point still stands with Gibbs Island as I ASSUME his latest run was his best

I'd assume that also, doesn't mean I agree with the rating either.
Most ratings are useless (to me personally), we're not going to disagree much on this topic, just pointing out the Base rating thing.

It depends how you use these ratings as to how useful they can be.

Like RPR's, I suppose in handicaps in particular there will be occasions when the horses Official mark is a lot lower than some rating service has rated the horse. Even though they are using different scales and measures, sometimes there is a clear difference, and I'd imagine that's where ratings devotees will consider their edge.

Katate Dori is a good recent example of this.

[thorne365]

Pretty sure they have both a master rating (what you're talking about) and a race rating. Early career they will be similar as you say - id be surprised if the master rating didn't decay over time though - it definitely loses the P/p so is subject to change.

I Think the main difference between TF and rpr is that the TF is more evidence based. Taking your example above:

- Gibbs Island has an OR if 134 so safe to assume that the TF master rating is current and 'fair'. The 'p' is there to indicate that he could still improve (won last race by 4 lengths so not easy to argue against I'd say)
- lulamba is rated on the basis of the last race performance seen (haven't looked at any of the numbers so I don't have an opinion on that particular race) but the 'P' indicates that it is expected to be able to seriously improve given the right conditions. I'm sure there is a loose interpretation of P/p that when applied put most ratings in similar ballparks

I personally much prefer the TF approach to attempting to separate actual from potential - the rpr rating seems to rate the race as much as the horse at times.

In the past I've tried to build my own (only very generic across the whole population) using the following methodologies:
- manually rate each race based on prize money/course/average winner ratings. Subtract from race rating based on distance between and a few other metrics for each runner.
- use a ranking model (based on halo player ratings) to score matchups within each race and run across all races to get a global horse rating.
- try to use market implied probability in the place of match ups in the method above

The first 2 worked alright (passed the eye test and were roughly as predictive as OR ) but we're not good enough to bet automatically with (which is what I wanted as opposed to a tool to form opinions on single races / selections). The 3rd was useless which surprised me a bit.

[Quevega]

Pretty sure they have both a master rating (what you're talking about) and a race rating. Early career they will be similar as you say - id be surprised if the master rating didn't decay over time though - it definitely loses the P/p so is subject to change.

I Think the main difference between TF and rpr is that the TF is more evidence based. Taking your example above:

- Gibbs Island has an OR if 134 so safe to assume that the TF master rating is current and 'fair'. The 'p' is there to indicate that he could still improve (won last race by 4 lengths so not easy to argue against I'd say)
- lulamba is rated on the basis of the last race performance seen (haven't looked at any of the numbers so I don't have an opinion on that particular race) but the 'P' indicates that it is expected to be able to seriously improve given the right conditions. I'm sure there is a loose interpretation of P/p that when applied put most ratings in similar ballparks

I personally much prefer the TF approach to attempting to separate actual from potential - the rpr rating seems to rate the race as much as the horse at times.

In the past I've tried to build my own (only very generic across the whole population) using the following methodologies:
- manually rate each race based on prize money/course/average winner ratings. Subtract from race rating based on distance between and a few other metrics for each runner.
- use a ranking model (based on halo player ratings) to score matchups within each race and run across all races to get a global horse rating.
- try to use market implied probability in the place of match ups in the method above

The first 2 worked alright (passed the eye test and were roughly as predictive as OR ) but we're not good enough to bet automatically with (which is what I wanted as opposed to a tool to form opinions on single races / selections). The 3rd was useless which surprised me a bit.

Yep they adjust the ratings for racecards/days, for handicaps and allow for penalties, conditions of the race etc etc.

I prefer timeform for the same reasons, and they've been at it (especially using the clock/timings) plenty long enough.

[darlojim]

Is there much evidence of it being useful in this case that their juveniles and improve vastly at different rates and they've all had varying number of runs. Surely the results are very skewed and not consistently on the mark?

Just thinking out loud

[thorne365]

I only have anecdotal evidence so also thinking out loud but:
- aren't juvenile ratings usually inflated when they step into open company
- improving at different rates is a problem for any method but one that tries to measure potential instead of assuming it should do better

The above should mean that they're (TF ratings) both useful in the juvenile season and especially the subsequent season where they should be on a more level scale with the horses they're now competing with (unlike ORs and I ASSUME rprs)

[That Horse]

Oooo hes got a little P, ahhh he's got a big P

Are the gay porn links back on the site again?

Also, it'd be pretty funny if this Gibbs Island wins the triumph now

[Kevloaf]

For anyone with the price they have on Lulamba

and no price on East India Dock


Would you rather your slip on Lulamba with the ratings as they are

or leave the ratings as they are, but have your Lulamba price on East India Dock



I know which I would