PARS HSH VS. CYNTHIA?

[HIGH ROLLER]

i am throughly confused. i have two friends who have pars from each of these companies. the pars and variants in many many cases differ in a large way.

how can i decide which is the better product?

[delayjf]

I would refer you to Dave S. he's the one that makes the pars. He puts a lot of work into them. The job is bigger than you might think. I have no idea exactly how Cynthia calculates their pars.

[codemonkey]

I've never seen the HSH pars, but I can tell you how Cynthia does it.

Pars are simply the average leading times for the various points of call for a particular track, surface, distance and class (TSDC). In the old days the technique used was to pin a set of times to the $10,000 claiming class and then just add/subtract a tick for each class below/above the $10K mark.

For the last several years, the Cynthia Publishing pars have been done by entering the chart times into a database and then deriving and verifying the averages using standard statistical methods. It is a team effort involving several people.

As the goal is to generate a consistant benchmark which is valid when compared track to track, surface to surface and so on, only "plain old ordinary" races are entered into the database - those races which have only older (4YO+) male horses and no unusual weight allowances.

Using standard deviation and curve fitting techniques, races with times which are too fast or slow to be considerd "typical" are tossed out until all races for any particular TSDC are within a normal bell curve. We then take the mean for each point of call, giving us "hard" numbers for the pars.

Sometimes there are not enough races at a particular TSDC to generate "hard" numbers. In that case the time is extrapolated from the surrounding classes.

Sometimes the times for a particular class, no matter how statistically valid, is out of line to the point that it's not believable, at which point the raw data is looked at to decide if it should be tossed out and exrapolated or left in.

These are the numbers published and used in the various programs.

A set of simple adjustments are then used to generate par times for other age and sex restrictions, etc. and comparing the par time to the race and pacelines gives you a differential for computing projected running times.

Hope that answers your question!

(And yeah, the reason I know so much about it is because I'm the database guy that crunches the numbers)

[TonyK@HSH]

It wasn't too many years ago that Cynthia Pub commissioned Dave Schwartz to create their PARS. In my opinion, his PARS are tough to beat.

tony

[HIGH ROLLER]

i was curious why hsh does not produce variants for the turf, i noticed cynthia does?

[Dave Schwartz]

Because the ADV approach is typically "self-correcting" on the turf since there are so few races.

When you go away from the northeast, most tracks run only one or two turf races per day. If they run one race and the race is 10 ticks off the track record the day will be a day whether the day is fast or slow.

Next year's pars may be a bit different. Stay tuned. <G>


Dave

[Jake]

I have posted several times on this, because I firmly believe the
HSH pars are superior par numbers. If you knew how often Dave and I have bumped heads privately in the past, you would see that this comes as an extraordinarily strong recommendation. Simple litmus test, buy both sets and test which ones match actual performance. Dave has been doing this for years and has it down pat, so I would suggest taking a look at his pars and seeing if they meet your own needs.

[kenwoodallpromos]

Tossed out, extrapolated, tossed out, compare, differential; sounds complicated to me.

[PaceAdvantage]

Originally posted by HIGH ROLLER
i am throughly confused. i have two friends who have pars from each of these companies. the pars and variants in many many cases differ in a large way.

how can i decide which is the better product?


Simple. Tell one friend to buy HSH Pars, and the other buy Cynthia pars.

Whichever friend wins more money at the end of the year, that's the Pars BOTH friends should buy next year.....


Got it?

[HIGH ROLLER]

well i will certainly step up to the plate and buy daves pars for 2005 if they have a turf variant. because for $130 i gotta get my monies worth.

[cj]

This is not meant as any reflection on Dave's product.

A $130 price tag doesn't guarentee anything in life, especially one's "monies worth." :D

[Dave Schwartz]

High Roller,

As much as I would like to accomodate you, since the Average Daily Variant approach does not work in turf races, we will not produce one.

Regards,
Dave Schwartz

[HIGH ROLLER]

don't understand why not, if people from the sheets, and other services produce them, why would it be so difficult?

[Dave Schwartz]

Mr. Bronzino (That is you, isn't it?),

We are not talking about a "variant" here. We are talking about the "Average Daily Variant." The way the DRF ADV system works it is just not suitable for a surface that typically runs one race per day.

And as for comparing an annual set of pars to a $35 per-day/per-track service... I can honestly say that this would be comparing apples and caviar.


Dave Schwartz

[HIGH ROLLER]

I AM JUST ASKING ONE SIMPLE QUESTION. CYNTHIA PUBLISHING PRODUCES A TURF VARIANT IN THEIR PAR FILE, WHY DO YOU SAY THAT IT CANNOT BE DONE?

[Dave Schwartz]

HR,

This is getting tiresome. I am not saying it can't be done. I am saying that the concept of ADVs does not work.

The people at Cynthia are certainly entitled to their opinion - it may be different than mine - but I do not believe it works.


Dave Schwartz

[HIGH ROLLER]

the only reason i am asking is that many people i know use software that requires a turf variant. i would reccommend your pars if you would add a turf variant for the 2005 upcoming par file.

i would purchase them myself

[socantra]

I think what he's saying is that he doesn't believe in the concept, and he's not willing to 'rig up' something he doesn't believe in to sell you a copy of his pars and have you recommend them to others.

Or, put another way, make them up yourself. They'd be just as valid.

socantra...

[thelyingthief]

as a bemused bystander, i would ask the specialists why any given array of numbers will not with equal effect and efficacy produce a useful and adequate variant?

the supposed variance of surface consistency has always struck me as a largely mythical construct, besides: how often have i seen the outlandish disagreement of times of races run on the same day, of an equivalent distance structure, and not just occasionally to the embarrassment of the superior class, i cannot begin to tell you. a 12k claimer posts fractional and final times significantly better than a graded stakes run the same day; a maiden-claimed two year old betters the performance of its elders, never again to duplicate its effort; fillies suddenly take wing at classic distances: all of these heresies confront me every time i sit me down to figure at the numbers.

frankly, i have my suspicions that maybe any mathematical representation of the races, but very especially surface variants, are in fact little better than a numerology, like when they burned at the stake whoever failed to reckon properly the angelic population residing upon the head of some pin. which is not to say that statistical reasoning is unuseful, of course it is: but when two sides of an argument about equine performance begin plucking different truths out of the same mathematical hat, for all the world like a horse race is an ounce of water one has a better number for, i have to wonder at it. after all, the numbers need only be crude to suffice, and if robustness is any criterion, may be the only thing that DOES suffice.

making variants using class pars is hardly a science, anyway, and whichever of the services you use will still require you to learn how to use them. that is, how to implement a judgment. besides, if you want to make really good variants, project; which hardly requires any kind of sophisticated par, i can assure you.

[Tom]

Projection is totally subjective, wouldn' t you say?

[Tuffmug]

Everything is subjective!

[kenwoodallpromos]

I think one problem with variants is lack of large enough base from which to draw at least daily times and finding some kind of standard to compare them to. More true with the small number of turf races run on any 1 day.
I used to have the same doubts and problems until I figured out standards to use.
Complications also arise because of many circumstances of the individual horses during the race unrelated to the track surface which affect winners' finishing times. ( have suggested that using the place horse's finishing times may be a more accurate reflection of the track condition). That and the actual variantion in day-to-day surface conditions, specially in the dirt, make projections more than a week and at times more than the current day almost impossible unless the track has been running the same all meet.

[kenwoodallpromos]

The fact is, the closer you can accurately monitor the track speed, bias, and condition even down to race by race, the better you can handicap if you really know how to use the information and relate it to running style, pedigree, jockey styles.
I mainly use pars for a very general view of tracks whose speed remains fairly stable, or to get an idea ahead of time what the crews at individual tracks may be treying to do on "normal" days. For exzample, I found a closer running too slow at Fairgrounds when that track wasa running fast; when the horse was moved tro CRC I checked the pars and recent track speed and was able to cash a win ticket.
Also I knew Tom Schell's current horse, Air Cool, had a great chance of winning because I knew that its running style (closer) was a good match for the CRC track (slow, sometimes negative rail). It has won 3 in a row.