Sorry for the delay in responding. We had a scary moment here this morning and afternoon. Our internet service provider went down--no connectivity. We were in midst of figuring out a plan B when voila we resurfaced. Whew!!

Q1--Number of patterns?

There are 4 alternatives--top, pair, off, X--all of which are possible. So that's 4 to the 4th or 4 x 4 x 4 x 4 = 256, and then 2 surfaces, so 256 x 2 = 512, and then there are age groupings 2, 3, 4 and 5 and up, so 512 x 4 = 2048, and then we vary by time of year, a new sequence beginning every month, so 2048 x 12 = 24576. Now there won't be monthly groupings early in the year for the 2yos, so the 24576 patterns is give or take or few.

Q2--What % of these patterns yield more than a 15% chance of a horse running a new top?

Q2A--What % of all runners post new tops on a given race day.

I think Q2 and Q2A are related. Don't know the answer to either, do not have hard numbers and certainly Q2A would vary day-by-day, card-by-card. What I will say is that the younger the horses are, the more likely they are to run new tops. So my bet is 2yos show the most positive Thoro-Pattern probabilities, followed by 3yos, etc., and within that earlier in the year is better than later in the year. Common sense when you think about it. More lightly raced horses are more apt to run new tops.

Q3,4--In regards to the three Thoro-Patterns you provide, you ask two questions: "Should we pay less attention to the TP in GP Fleet's case because of the relatively small sample size? Should we be twice as influenced with the information on Deer Lake's pattern in relation to Elijah's Song based on the size of the sample?

Briefly the patterns cited are:

Deer Lake--Dirt 5yo & up Apr-Jun Prev Top -1
Off-X-Pair (2-1/2, 6-3/4, -0-1/2): 684--8-29-30-32 (starts-tops-pairs-off-x)

Elijah's Song--Dirt 4yo Apr-Jun Prev Top 15-1/4
Top-Off-Off (3, 4-1/2, 6-1/2): 280--9-24-36-31

GP Fleet--Turf 5yo & up Apr-Jun prev Top 0-3/4
Off-X-Off (3-1/2, 5-1/2, 2-1/2): 99-2-21-43-33

These three Thoro-Patterns come from sheets posted with figures for 5/7/5, not as they appeared on 5/7/5. For the questions raised the date difference is not relevant.


The answer to both questions is no. I need some help from the statisticians out there. The sample size in both cases is large enough to be significant. We're talking confidence intervals here and surely the larger samples have higher DEGREES of confidence, but 99 is a large enough size to be valid. In distributions, which the Thoro-Patterns relect, there are statistical techniques to determine degrees of confidence--that is how well the data reflect what's being measured. And it's not a direct realtionship. In others words "twice as influenced" while intuitively sensible does not hold here.

Q5--"Refer back to GP Fleet's TP, which predicts a 2% chance of GP running a new top; if you look at trainer Steve Flint's TG figure based trainer profile, of his 264 Overall starters, 16% ran a new top. What's a handicapper to do?"

Look at what his 5yo & ups do. That's most relevant group--54-7-19-48-26--and it's close to what the Thoro-Pattern shows. The overall includes the younger horses which are much more apt to run new tops.

The Thoro-Patterns for Buzzards Bay and Bellamy Road as posted with Derby day figures are:

Buzzards Bay--Dirt 3yo Apr-Jun Prev Top 5-3/4
Pair-Top-Pair (4-3/4, 3-1/4): 185--20-34-25-22

Bellamy Road:--Dirt 3yo Apr-Jun Prev Top 5-1/2
Top-Top-X (-0-1/4, -5, 4): 78--3-19-29-49

You cite BB's pattern as a peeve of yours because it's a "wash" pattern and you go on to state that "Bellamy Road had a similar pattern with a much smaller number of of starters." I disagree on both counts. BB's pattern is strong, 54/46 positive to negative and BR's is deteriorating, 22/78. But Bellamy Road is much, much faster and that's something the handicapper has to take into account.

I agree the individual patterns reign over the trainer profile and Thoro-Pattern in terms of analyzing a horse's prospective chances. But both the Thoro-Pattern and the trainer profiles provide information which may be relevant in specific instances. The trainer profiles point to trainer tendencies and the Thoro-Pattern provides a general overall snapshot of the horse's pattern. But both sets of measures have to examined closely to see if they're relevant to the particular horse and race in question.

I sense a general undercurrent, perhaps unconsciously, of I guess, disbelief--disbelief that horses don't run tops often. I mentioned this observation in the ROTW string and it's true. Only real good, healthy horses are able to replicate good, improving form over a period of time and even then at some point they hit a set-back level. Trainers are considered good if they win at a 20% threshold, and 15% or so isn't bad. Baseball hitters are good if they hit .300 and .275 isn't bad. You have to look at Thoro-Patterns in the same light, which is difficult and may take some reconditioning. But this game's only easy when Official lights up and you're holding the winning ticket.

Hope this helps.




















TGAB



Edited 1 time(s). Last edit at 06/08/2005 05:46PM by TGAB.