No race until Monday 5th with a selling race
at Warwick. This will be followed by a maiden at Ffos Las on Tuesday, a
Novice event on Wednesday at Nottingham (Takeaway & Style And Panache
possible runners) and a Maiden race on Thursday at Leicester.
This Preview will summarise the usefulness
of the FTO RAG (i.e. First Time Out Red-Amber-Green or FRAGs) files for
the trainers which are linked from their 2010 Season individual Pages.
The FRAGs are presented in three versions sorted by Date order, descending
FTO Estimate (the B2yoR rating) and ascending SP. The individual horse
lines are colour coded as follows :-
Dark Green = FTO winner
Green = Winner during the Turf Season
Amber (Yellow) = Placed during the Turf Season
Red = Unplaced in all runs (ran at least twice)
White = Only ran once.
The horse lines provide the usual information
about the FTO runs such as Date, distance, final placing and so forth.
The final four columns give overall 'Season' information about the horse
Best Estimate achieved in the Turf Season
The Run on which the peak Estimate was achieved
The Total number of Runs in the Turf Season
The Percentage of the Best Rating in the Season
that the Horse Achieved on Debut (How 'ready' it was).
We could propose a 'Null Hypothesis'
that there is nothing to be learned from FTO runs and 2yo racing
is entirely unfathomable. Debut runs by juveniles are affected by all sorts
of issues that are problematic. The horses will run to a wide variety of
performance levels within the actual amount of ability they have. The rating
achieved FTO will have little relationship to the final Official Rating
they have because demonstrated 'Form' over, at least, 3-4 outings is the
only way to assess the 'Class' of horses. Trainers just run horses when
they think they are ready and there is nothing to be learned from when,
and where, they start out. We would be proposing that 2yo racing is very
difficult to understand and FTO runs are, essentially, random and
there is no pattern to them.
Everything in the previous paragraph is cobblers
but we do need to prove it. Never, just believe people who come
up with plonking statements that you do not think they have researched
to any level. The broadcast and printed Media is full of this nonsense
and is misinforming the whole of Society with their insufficent checking
of stories. This is a hugely negative issue because you are better off
with someone who says the 'Do not know ' an answer. At least you then know
you need to research the issue further. If an 'Authority' like, for example,
the BBC News expects reporters to respond to questions with definitive
answers when they got off the plane 45 minutes ago then you cannot trust
The point of coming up with an initial hypothesis
lies in being able to predict what we should expect to see in terms of
Colour Patterns in the FTO RAGs. If there is nothing to learn we should
find a random mix of the four colours which 'Sorting' by any data variable
will not affect. If we see patterns we also need to be able to estimate
whether they suggest a real underlying cause or are just 'lumpiness' you
will see in any random series. Toss a coin often enough and you will find
runs of, say, 10 heads within the series, if very rare. That does not indicate
a biased coin or some underhandedness if the 50-50% split is upheld across
the whole series.
The following discussion is not going to go
into the maths for proving statistically significant results so do not
worry. We will work through the three different FRAG types and highlight
some ways that they might be looked at to identify patterns. More importantly
to propose reasons why some trainers show certains patterns. Not rigorous
proof but a starting point. One point to remember is that the 2009 season
FTO RAGs we are working with are part of a series of years of similar data.
The obvious question to research further is then 'Does the trainer show
the same pattern every year?'
Let us start with the Date Order files.
A worthwhile question would be whether a trainer batches the better horses
up in some way. That would then allow us to make an estimate of whether
a particular debut was likely to be competitive for the day and a possible
future winner. Initially let us look at a sample of smaller stables who
typically have 5-10 individual runners each year. These should be easier
to look at than 130+ Hannon horses with several different patterns probably
overlaying each other. Here are links to the 2009 Date Order files for
- Gay Kelleway, James
Given, Mark Brisbourne
and Alan McCabe.
Any evidence in those files that they might
choose their likely winners and run them first or in the earliest batch?
B2yoR would suggest enough to make it worth someone's while to look at
it further. A quick summary. Miss Kelleway had a good season overall with
three winners from 5 runners so that slightly hides the fact that the first
runner was a winner. In 2008 the only winner that made the FTO run for
her was her first runner. Which means the fact that she had an early runner
in 2010 might be a clue. Primo Lady won at 12/1 as her first runner this
Given's RAG is even more clear cut. This pattern
is not unusual and the Previews so far have made the point that he, effectively,
has two different types of 2yo runner. The point that Frank Carter (2yo
guru for the 'Authority' Racing Post) has not bothered to look into. Which
is why the Brocklesby preview got interested in an early runner for his
stable in Sky Diamond (as with an early runner for Kelleway). The Paddock
Review for that horse confirms that Given has picked out a solidly built
athlete and no problem thinking he can rate comfortably above OR70. He
is a nutter at present so the mental wildness stopped him showing his ability.
Mark Brisbourne does a solid job a low level
with his 2yos. in the 5 seasons 2005-9 he has had 1 or two winners each
year with the number of runners ranging from 5 to 13. In every year
the first runner of the season has been a winner. The second winner can
be a bit more difficult to spot and in the last 4 seasons the second runner
of the year has managed to place, at a low level rating, at best. If he
produces a runner soon it ought to be an OR64 winner of some sort, at least,
and the odd one will be better. After that it will be a rare season if
the only other winner is not a seller success of some sort later in the
Alan McCabe is a more difficult proposition
because in his fourth season in 2009 his career was still developing. The
number of 2yos he had was increasing and he was in the process of moving
to set up in his own stable having worked for owner Paul Dixon. The 2009
FRAG suggests that he chose a small number of competitive ones to run early
including his best, and only winner, Fratellino. The last four debuts over
5f were definite garbage. Isle Of Ellis & Gasparilla were also garbage
but would fit the bill of 'Tiny' ones having to run early before they looked
too ridiculous against the opposition. The first 6f debut was the placer
and then there was a break to later season and the 7-8f debuts started
with those capable of placing at OR60s level.
That would count as a working model because
his previous FRAGs were different but to a large extent because of a different
type of raw material. Mr McCabe has talked up the crop of 2yos he has for
2010 and ran the first of them last week with Barista getting put through
the mincer in a fast race at Southwell. Mr McCabe seemed a little disappointed
at the horse's inability to compete better. A mixed story and B2yoR does
not have a solid view on Barista yet. He is not a 'little one' and might
be ok. In Paddock Review he is ok in front of the saddle and questionable
behind. He stuck at it for a long time in a tough test and probably ran
ok. On the template above he ought to be an OR70+ type. See how he goes
For a bit of variety let us link to the FTO
RAGs by date for Bill Turner
and David Evans. This are
bigger 2yo stables than those above but have a different approach. They
get nearly everything they have ready for the first two weeks of the season
and a lot of their 2yos will have run before April ends. In that model
you might not expect the 'front loading' of the competitive 2yos
to show up like it has with the smaller strings. Mr Turner has less competitive
2yos and the best of them will go in the Brocklesby to start the sequence.
After that the early runners are more likely to include usable ones but
will include garbage. Difficult to spot the difference just on order.
In 2009 Mr Evans seemed to front load a number
of his best 2yos in the first 5 runners before tailing off. 2008 showed
a similar setup but 2007 was subtlely different with a few more of the
garbage types starting before mid-April and Vhujon's debut win blast. But,
looking at the overall record 2007 was a rotten season by his standards
with juveniles and the early garbage actually told you he was going to
struggle before April was out. In that context what to make of the seven
2yos he has run already in 2010? Style And Panache is a better one and
Bajan Bullet is probably not really a 5f type and can do better but difficult
to work up enthusiasm for the others.
You probably get the idea now that if you
have a framework for how you think the season is going to unfold you can
come up with any number of interesting theories as it goes along. Or you
could sit there with Frank, blinkers on, randomly picking out whatever
'stat' comes to hand, with no supporting Model but caring less so long
as you fill the space down to you, and waiting until some 'Form' establishes
so the Ratings guy at the RP can tell you what to think. It is probably
A lot of other areas to think about with 'Date
Order' files and looking at how long the 'tail' of duff ones is for major
trainers in later can be useful. Do some of the larger trainers have a
'Surplus Stock' period when they get a group moderate one running for major
owners? Particular trainers are interesting and Mark Prescott an idiosycratic
trainer where his record is variable from year to year but the first batch
of runners each year are the '2yos' and the rest the set-ups. But, the
first runner in any batch can often be the best of them. B2yoR used to
think of Roger Charlton as a classic 'Front Loading' trainer but the last
three years have produced variable results and that now looks a suspect
One item to make a point of though. Here are
the 2009 files for Richard Hannon
& Mick Channon (who
used to work for Hannon). They ran 214 juveniles between them in 2009 and
more than 7% of all of the 2yos that ran. That is a constant test during
the season where you have to face up to the 'What should I do about
the Hannon/Channon newcomer in this race?. The sloppy thinking in the
'Spotlight' features will be telling you that these are 'Major' 2yo trainers
and the stables are in form so you have to take these debut runners seriously.
What to do?
Try this Model. Both trainers are 'Peak STO'
and FTO wins are actually quite rare (although their Strike Rate is at
least average with FTOs). The debut successes are due to high class runners
and odd accidents in soft races. They each get their small number of debut
wins in a similar manner. Odd ones in early season (the first 3 weeks),
a small set with their 'Royal Ascot' 2yos and this is often in Mid-May
to early June but can be a bit earlier for Channon. They then get odd FTO
successes through the long tail of their other debuts. Usually for specific
owners (The Queen for Hannon and Jaber Abdullah for Channon). Aside from
those times just cross their debut runners off the list of possibles. Spend
the times this frees up trying to learn whether you can predict something
useful from the circumstances of the debut what type of 2yo this might
be for the STO runs and ownwards. With which let us move on to how the
FTO rating, i.e. the Estimate, might be used in this way.
The RAGs began with looking at trainer John
Dunlop's debuts and trying to see whether the FTO ratings achieved by his
runners were a good indicator of STO rating and best figure for the season.
Mr Dunlop prepares his 2yos thoroughly at home using his own work riders,
jockeys are not allowed at gallops, and they will be about the fittest
FTO 2yos you will see aside from someone like David Evans. Which means
they are not going to fail and underperform on the initial outing through
getting tired. They will lack racing nous which will take the edge off
the rating and leave some improvement to the second run.
The hypothesis you can try to work to is that
with that background the Estimate for the debut run tells you how good
the 2yo is, completely. One that rates 10 points lower than another on
debut really is a 10lbs less able athlete and so on. The highest raters
to a certain level will be future Listed+ winners and the non-winners wil
also be apparent. A big issue to understand is that this analysis is impossible
with ratings systems based on the traditional BHA handicapper methods.
The BHA guy has to over-rate maiden form because the race rating has to
be the prospective Official Rating (OR) the horses will have for nurseries
and not necessarily the absolute performance level they achieved.
A big reason for the B2yoR Estimates going off into their own methods and
trying different scales.
What should we expect to see in the FRAG ordered
by FTO Estimate with the 'Beau Ideal' trainer who prepares his juveniles
to a consistent level for debuts? You would want the FTO winners to be
at the highest ratings followed by the best of the winners. You would not
want to see winners sprinkled across the bottom half of the ordered list.
The FRAG includes a 'Best Estimate' for the season which is colour coded
by the same levels as the 2010 results ouput. You would want the Green
Coloured 'Bests' to be at the top and then the Ambers to filter into into
the red at the bottom. Demonstrating that the order of FTO ratings were
a good match to the final ordering of best ratings.
Here is John
Dunlop's FTO Estimate RAG for 2009. A pretty good match if with some
caveats. The Yellows & Reds are a bit too mixed but looking at the
figures they are in a fairly narrow range. A really interesting point are
what are Many A Slip & Spanish Duke doing so low on the list as later
winners? Well, it is a case of John, you might be a Naughty Boy. Look at
the season winners higher up in the list and their 'Best' ratings were
achieved on one of their first three runs. The other, low debut rating,
pair reached their peak on the 4th & 6th outings. A check of their
careers show they were development projects targeted at showing their best
form in later season after they had got into nurseries and were going up
in trip. Could they possibly have been left a bit underdone for their early
debuts to assist with being targeted at Nursery futures off workable perches
compared with the other, all-out, Open Maiden winner types?
Try looking at this file for Walter
Swinburn, isn't it beautiful? Well, eye-of-the-beholder, and all that.
Look at that 'Best Estimate' column, unmixed Greens to Yellows then the
Reds. Exquisite. The trainer would say he is not a good trainer of 2yos
but the FRAG says he is a very consistent one. You can predict the ability
range of his 2yos to a pretty close level from how they go FTO. He gets
the debut winners with the better ones and the winners, of open maiden
prospect level, will stand out. The Garbage will be welcomingly anonymous.
In looking through for a bad example to counterbalance
the loveliness it proved quite difficult to find something really bad,
part Random. The majority of them were very encouraging for believeing
the predictive value of the first run. Perhaps the ratings work is coming
through. So, let us choose an important example to make a point and - Mark
Johnston. Now, Mr Johnston used to get more of the 'Lost Debuts' (complete
blowouts) than most larger stables and certainly not the fully consistent
picture with many trainers. His colour pattern is not dreadful and there
is an overall shape to it. But, there are later Green winners and high
raters mixed in right down to the lower reaches of the FTO estimates order.
You could list Pallantes Cross, Nave, Avonrose,
Whippers Love, Lowdown, Trailblazing, Karaka Jack, Bikini Babe & Take
Ten as good 2yos who achieved poor ratings first go. Something about the
Johnston preparation and race style leads to a higher incidence of this
sort of debut underperformance. The point to make is that, unlike many
trainers, you will have to trawl through the poor debuts for possible improvers
because the FTO ratings are more variable than the average.
It is a quiet moment before racing starts
for the day and B2yoR is at the course early. The 2yos for the first will
not be around for an hour yet so some time to fill. Hmm, let's push some
time round by reading through the 'Racing Post' Spotlights for the juvenile
events and underline the variations on the summary of "....the Market
should Guide...." for the debut horses. 50%+ of the summaries would
be typical with a lot of newcomers.
There is no polite way to put it but this
drivel is pointless arm-waving and should be banned. Ok, so a summariser
is not allowed to put "I have not got the first idea what to write about
this animal so will fill the space with something..anything...". But,
if that is what you think then you either are in the wrong job or need
to work harder to have a Model for 2yos races and a framework to approach
Let us see, in what way will the 'Market'
guide us? By the horse being odds-on?, Evens?, less than 5/1, more than
10/1? By there being a huge, visible plunge on the horse that takes the
price from 20/1 to 2/1? Or subtle trimming to 10/1 from 12s? How should
we value support in the Morning over support closer to the race? What about
in the final 5-10 minutes after the horses are on the way to the start?
How should we adjust our view for thin Markets at Gaff tracks compared
to more sturdy instances at Newmarket & Newbury? How, in the name of
Sam Hill, has anyone managed to summarise that complexity in one sentence?
Not only has our pundit tried to opt out of
telling us anything with the bullshit but they have not even told us how
to use the 'Market'. Perhaps we are doing the pundits a large disservice
and they have spent many hours poring over betting pattern data versus
final placings and 'Relative Performance Indicators' to assess how well
the 'Market' predicts what is going to occur. This seems vanishingly improbable
on the evidence available but even if they have the banal "...worth
a Check in the Market..." formulation is a useless, incompetent, way
to pass the knowledge on to the paying client.
With which, what can we learn from the FTO
RAGs in terms of how much the 'Market' might know? Could we assess how
much it might know about which are the good ones and competitive on the
day? No bullshit here so they are not a detailed analysis of this. They
are a lightweight way to get an initial handle on this question and what
level of SP might be a guide. If you had a swift shuftee at the FRAG before
a race you would know far more than the RP pundit.
Here is an interesting example - Hughie
Morrison. Lord Whatever-he-is-to-be (he is the Heir, isn't he?, not
a spare) is not really a 2yo trainer and thinks that it probably is not
proper racing. But he is pragmatic enough to get involved and is a good
trainer so he gets the winners he should. Look at that, the SPs range from
11/1 to 66/1 and more than a hint that the 10/1 to 14/1 range are the ones
to be interested in for the day and as winners. Budva was 14/1 at Southwell
and Hot Diamond 20/1 in a tougher Goodwood race so you could argue that
their SP levels are actually equivalent if you wanted to fiddle it further.
Mr Morrison has had a single debut winner
in each of the last three years either in late September or early October.
His slower approach meaning his better ones are ready to win by that late
in the season. In 2009 it was useful Pastoral Player at Newbury at 12/1.
Later Listed winner Queens Grace was 11/1 on her starting effort. In 2007
the debut winner was later Group winner Simulation who was only 6/1 at
Newbury. In 2008 the FTO success came with State Banquet who was 14/1 at
Salisbury. That was the shortest SP of any of his runners that year.
To try to summarise that usefully as a Rule-of-Thumb
to back up the Paddock Review how about this. Mr Morrison runs a few competitive
2yos in Late May as a 'Front Loader'. Because it is relatively 'Early'
season for him they can bomb out but will win in time. The majority of
the others to run will be 3yo set-ups plus the odd one that will win at
Southwell during the early winter. By September he will know which are
the best he has and will have had time to get them properly ready. The
'Market' will tell you which are the good ones because they will be 12/1
rather than 20/1+ in a solid Newbury maiden. A really top one like Stimulation
will get to single figures. But, this will not be the "Having it
riggght offf.." GAMBLE that Pundits and Racing TV types want to believe
in all the time. It is a subtle lowering of the SP in response to a non-Betting
stable being confident that this one is 'Ok'. But laced with a modicum
of British restraint & understatement. [As an aside one of the things
that 'Racing For Change' might like to look at is the peddling by the Media
of the childish fantasy that every race is stitched up in advance
by clever insiders with huge brains who have plotted this for the last
12 months and are now all Millionaires while the rest of us are still festering
wherever we were before because we are inadequates with small brains and
no access to the levers-of-power. B2yoR is a firm believer in the first
approximation to the Cock-Up before Conspiracy assessment. Couldn't plot
his way out of Paper Bag....]
How about this as a record for Luca
Cumani? Similar to Mr Morrison he is not much fussed about the 2yo
palaver and only 3 debuts at single figure SPs. The shortest was the FTO
winner Seta (by 10 lengths) and another was one of his other two winners.
The one at 4/1 was in a thin Ffos Las race but placed on both runs so is
presumably OK. What about his first runner Start Right then who was his
other winner but was 33/1 FTO? He appeared to be a development project
who was being brought along slowly as a 'Nursery' type.
You should be getting the idea that if the
'Market' knew everything, as you are constantly told, the debut
winners would be at the shortest SPs, the better types should bunch up
in the lower half and the moderate to garbage in the upper half of longer
SPs. Here is the John Gosden
file for 2009. What are those FTO winners and later successful horses doing
at the longer SPs? Why are there more Reds at the shorter SPs than the
longer ones? You could probably prove that the colour pattern was RANDOM
with the appropriate Stats test. Take Azmeel out and it almost certainly
is random. How much is the Market really guiding us in this stable's case?
How well are you being served by the Pundits who suggest it is a terrific,
and reliable, guide?