dave, question, your pars times table

[freeneasy]

hey dave, richard here
i ve just recently started getting back into handicapping. i like to design my own systems and i partway have a system that is on and off. when its on its good but when its off its off.
the system has come up with some big longshots and iam pleased with that. i use 4 contenders for fields up to 9 and 5 contenders for fields 10 and over. a #3 contender paid $100 at holly, a #4 contender paid $207 at gulf, a #1 contender paid $65. the system had an $85 payoff, 2 or 3 60's, 40's, 50's on down the line to the favorite yawanna twist in the 8th at aqu. sunday.
the system sometimes gets hot then just takes a crap.
when using 4 to 5 contenders per race there should be a reasonable expectation that some big price horses will come thru so these wins ive mentioned are really no big deal as they should to certain degree be expectd. hell the first 4 or 5 odds betting choices in a race will produce its share of longshot winners.
but the thing that just keeps getting away is the fact that a system is a system is a system is a system. u may have all the best horses in a race and no matter how right a system may be they can all take a crap. reason being? well i asked yogi one day why the right horses lose and yogi said "well if u asked a horse why he lost a race when he shudda won the horse would probably tell u that he's only human, i mean if u were a horse u'd be human too" thats what yogi said yep.
i want a chart that shows all the class levels for each track and
ordered or ranked to show
the stongest or highest class level for that track
with the next strongest class level to follow that
with the next higest class level following that
and so on down the line
for example
there are like 30 different conditions that a horse running at santa anita can run under and i'd like to have a structured order that ranks each of these levels according to the strength of their conditions.
i cant do it dave. i cant determine the ranking order that each class level should fall under according to the strength of their conditions. i mean we all know that a nw3x is a stronger condition then a nw1x and that a nw2L is pretty much equal in class to a n1x but then u have the op clm nw1's, 2's and 3's. ok we all know a op clm nw3 is classier then a op clm nw1 but is that same op clm nw3x a stronger conditioned race then a nw2 other then md or clm? then u got the conditioned claiming races, nw2L, nw3L, n1y, n$y, and so on. and the starter hcp races who do u rank them above, under and or equal to.
i feel that the strength of these cless levels when put set down in their proper ranking order would add too and improve my system. if u have it, i want it. id like to make this clear tho. if u have charts that do display the ranking order of the class hirearchy's for the verious tracks i dont think i would want the class levels to be ranked or based on the par times for that class. iow if a $20 clm has a par of 1:10 then that class level would be considered a stronger class level then a $40 claimer if the par times for the $40 claiming level was a tick slower in 1:10:1. each class level has its own spacific condition seperating it from all the other conditions conditions see what i mean. and iam trying to find those differences in a ranked order.
right now, for me its all about strength vesus weakness and ranking all the different classes according to the conditions from the strongest down to the weakest conditions being carded without par times being the dictator of dominance. and this is basically what iam looking for dave
below is what iam up against.

g1
g2
g3
non graded stks and hdcs above $200k
non graded stks and hdcs up to $200k
restricted stks
overnight hcps

clsf alw
nw4x md clm
nw3x md clm
nw2x md clm
nw1x md clm

nw4L
nw3L
nw2L

starters alw
alw n1y nw race in specified time
alw n$y nw specified monies in specified time
alw n1m nw at mile or over in specified time
alw n$my nw specified monies at mile or over in specified time
alw nc no conditions
alw c multiple conditions
maiden
maiden clm

op clm nw3x
op clm nw2x
op clm nw1x

clm $100,
clm $80,
clm $60
clm $50
clm $40
clm $30
clm $20
clm $16
clm $12
clm $10
clm n2L nw 2 lifetime
clm n3L nw 3 lifetime
clm n1y nw race specified time
clm n1my nw race at mile or over specified time
clm n$y nw specified monies in specified time
clm n$my nw specified monies at mile or over in specified time

see right now i dont care about par times. right now i want this listing of all these race conditions ranked in their proper order of strength

well thats about it, hope u can get back

[Dave Schwartz]

i want a chart that shows all the class levels for each track and
ordered or ranked to show
the stongest or highest class level for that track
with the next strongest class level to follow that
with the next higest class level following that
and so on down the line
for example

We've got that. Below is the one from 2006 shown on our website. It may change in appearance a little from year to year.

http://www.horsestreet.com/BBSImages/Parsbook2006.jpg]


there are like 30 different conditions that a horse running at santa anita can run under and i'd like to have a structured order that ranks each of these levels according to the strength of their conditions.

Usually all conditions for older males are represented, with point adjustments for females, N2L, N3L, and statebred. There are no Starter pars (except California) because they are just useless from one year to the next.


i cant do it dave. i cant determine the ranking order that each class level should fall under according to the strength of their conditions.


By "each" I think you mean "every." We can't either. Well, we could but the book would be 1,000 pages and cost about $75 per copy just to print. You'd need a wheel barrow to haul it around and the cost would be around $500. Just impractical.

We treat N4L as "open" because there are just too few races and the differences are slight.

The non-winners in a period of time races in claiming fall into the open category as there is just no way to build all the different classes. IMHO, these subtle class differences are much more dependent upon the mix of horses in a given field.

In allowance they are called "classified" (CLF for short).


id like to make this clear tho. if u have charts that do display the ranking order of the class hirearchy's for the verious tracks i dont think i would want the class levels to be ranked or based on the par times for that class.


LOL - I build "par times." That's what they are for.



g1
g2
g3
non graded stks and hdcs above $200k
non graded stks and hdcs up to $200k
restricted stks
overnight hcps

clsf alw
nw4x md clm
nw3x md clm
nw2x md clm
nw1x md clm

nw4L
nw3L
nw2L

starters alw
alw n1y nw race in specified time
alw n$y nw specified monies in specified time
alw n1m nw at mile or over in specified time
alw n$my nw specified monies at mile or over in specified time
alw nc no conditions
alw c multiple conditions
maiden
maiden clm

op clm nw3x
op clm nw2x
op clm nw1x

clm $100,
clm $80,
clm $60
clm $50
clm $40
clm $30
clm $20
clm $16
clm $12
clm $10
clm n2L nw 2 lifetime
clm n3L nw 3 lifetime
clm n1y nw race specified time
clm n1my nw race at mile or over specified time
clm n$y nw specified monies in specified time
clm n$my nw specified monies at mile or over in specified time





see right now i dont care about par times. right now i want this listing of all these race conditions ranked in their proper order of strength



I do not feel that my pars will satisfy your every need. You want ranks by "strength" but you don't want to use pars. What would you suggest that we use to determine "strength of race?"

I know how I would do it.

I would use "strength of the individual race." For that you don't need pars. What you need is a database of past races that allows you to say "this horse came out of a '106' race and that one came out of a '104' race."


Dave

[TrifectaMike]

I know how I would do it.

I would use "strength of the individual race." For that you don't need pars. What you need is a database of past races that allows you to say "this horse came out of a '106' race and that one came out of a '104' race."

Dave

Horse Talk No Math (Not yet)

Why not put aside the quest for pars and ranking from the handicapping process entirely. Finding universal metrics for race performances as well as race predictors has been the holy grail everyone is striving to achieve.

Each race in my opinion is unique.

Let's look at an upcoming race. The entries have in nearly all cases run against a different horse that won the race. An entry can also occasionally be the winner in its last race. Looking at only the horse's last race as a singular event, we have as many singular events as we have entries (except in cases when one or more horses come out of the same race.

So, let's proceed. How can we tie these singular events "together". What if, we could define a "representative" winning horse encountered by all the entries?

We should be able to view this race as a race against the "representative" horse.

If there is any interest, I can suggest a way to model a "respresentative" horse.

Mike

[Sham]

Horse Talk No Math (Not yet)
....
If there is any interest, I can suggest a way to model a "respresentative" horse.


Reading your post, I was reminded of Huey Mahl's "Superhorse"....sure, I'd love to hear more.

[Dave Schwartz]

We should be able to view this race as a race against the "representative" horse.

If there is any interest, I can suggest a way to model a "respresentative" horse.

I'm all ears.

[freeneasy]

thanks for ur resonse dave
u have ranked in order for 3yo&up
stk grd
stk non graded
oc100
oc85
clf
oc75
c75
c65
str alw50
c60
nw3
oc65
c50
c45
c40
str alw35
c35
nw2
oc30
str alw35
c32
c30
str alw20
c27
c20
c18
c17
c16
nw1
my question is this. did u base the ranking order of this list on par times?
lets say clocks and watches were never invented and in looking at a racing form all we had to go on was horse A won his last race in a clm $16,0 race and horse B won his race in an alw nw1 so with all things being equal the real question for me is which horse took part in the stronger conditioned race? a or b?
i ask in this way because i see u have the clm $16,0 level ranked higher then the alw nw1 level which leads me to think that the structure of your class levels are based on par times and not the conditions that seperate each class level.
it would seem that the only way a $16,0 claimer classwise can be ranked ahead of a nw1 other then is if the par for $16,0 claimers is faster then the par for nw1's other then but the conditions for each of these levels tell me that the compitition entered for the nw1 will be stronger then the compitition enterd in $16,0 claimer.
if i had to decide between a $16,0 claiming winner in 1:09 racing today against a nw1 winner in 1:09 more then likely i'd take the nw1 horse 90-95% of the time because the nw1 horse would probably win over the $16,0 horse 65% of the time simply because the nw1 race will carry with it the stonger conditions and have had the stronger compitition.
i picked up this list of class rankings from 'todays racing digest' and this is basically what iam interested in
they list their class levels in 5 point incraments exam grade1 races are rated at 210 but dividing that by 5 breaks it down to and gives it a rating of 42. grade 2 races are given a class rating of 205 and broken down by dividing by 5 makes the rating 41 ect


3yo&up

42 grade 1 stakes & handicaps
____________________________
41 grade 2 stakes & handicaps
___________________________
40 grade 3 stakes & handicaps
non graded stks & hcps above $200k
_____________________________
39 non grades stkes & hcps thru $200k
_________________________________
38 restricked stks
overnight hcps
______________________________
37 clf alw
_____________________________
36 alw nw4
alw nw3 md clm
$100,0 op clm
$100,0 clm
________________________
35 $80,0 clm
_______________________
34 $62,0 op clm
alw nw3
alw nw2 md clm
$62,0 clm
______________________
33 $50,0 clm
_______________________
32 $40,0 clm
______________________
31 alw nw2
alw nw1 md clm
$32,0 clm
_____________________
30 $25,0 clm
_____________________
29 $20,0 clm
____________________
28 maiden
$16,0 clm
________________
27 $12,0 clm
$62,0 md clm
_________________
26 $10,0 clm
$500 md clm
___________________
25 $8,0 clm
__________________
24 $40,0 md clm
$32,0 md clm
_________________
23 $5,0 clm
$25,0 md clm
__________________
and they suggest to deduct 1 point for all 4yo
deduct 2 points for state bred md and alw races
deduct 2 points for all clmg nw2L
add 2 points to the clmg prices for strter hcps

basically these rankings are based on the typified strength of a race that comes with each set of eledgeibiity conditions.

todays racing digest only puts out this chart for del mar, sa, hol and gg and this chart wont conform to other lesser tracks because their nw1 races for exm might follow in the same pattern as the above chart but fail to stand up to the stronger nw1 horses at the stronger more higher level tracks.
i'd like to take a moment to say something here dave that this kind of chart if designed only for the purpose of distingushing the strenght of elidgebility conditions, when put together in the above foremat would only take up maybe 2 or 3 pages at the most. some food for thought.
but thats the basic idea iam after

[Dave Schwartz]

Par times are what I used.

I'd really like to know what you would suggest to use.

How do YOU determine "strength?"


Dave

[TrifectaMike]

In order to keep this process easily understood, let me describe the process
by example... cookbook approach. For the purpose of clarity, I'll make some assumptions.

Let's consider a typical race with the following assumptions.

1. All horses entered in the race, raced at the same distance in their last
last race, and todays race is also at the same distance.

2. The last race for the entries are independent.

3. The measurements we utilize are made consistently, and are reproducible.


Okay, let's move on to the relevant data.

Record the 2nd call and final time for each horse's last race. This data is
the race data, and not the horse's data.

So, if we have eight entries, we have eight data pairs:

Horse(n): 2ndCall(n) finalTime(n)

Now, let's make use of these data pairs to determine a simple linear model.

Fit the data to the following model:

FinalTime = b0 + b1*2ndCall (This model applies only to this race)

Time for another assumption... this one involves the 2ndCall times.

Let's assume the 2ndCall times follow a Triangular Distribution. A Triangular
Distribution is a continuous distribution defined by a lower limit, upper limit
and mode. ( The limits and mode is only useful for this race)

Okay, what have we accomplished?

We have modeled a horse which is "representative" of a horse encountered
by all the entries in the race.

We can take samples from the 2ndCall Triangular Distribution and plug them into the FinalTime model,

FinalTime = b0 + b1*2ndCall

to arrive at a FinalTime distribution.

Where does this lead us? Ahh, that's the fun part. Use your imagination!

Mike

[Dave Schwartz]

Mike,

Please try again as I did not get it.

"Upper limit" and "lower limit" I understand. What's "mode?"


Dave

[freeneasy]

dave i will be busy till later on in the day. talk to u then

[TrifectaMike]

Mike,

Please try again as I did not get it.

"Upper limit" and "lower limit" I understand. What's "mode?"


Dave

Mode is the value that occurs most frequentkly in the dataset. Since, we are dealing with small samples, the mode most likely has to be estimated using the mean of the data set. For the Triangular Distribution :

Mode ->3*mean -upper limit - lower limit

Mike

[Dave Schwartz]

I have done something like this in the past. It just never performed as well as my own definition of "strength."

[freeneasy]

all iam talking about is the eligibility conditions set down for each race that qualifies a horse to enter said.

A) if the eligibility conditions for a race is that ur horse must be enterd for a claiming price of $25,0 then in order to qualify ur horse to run in this race the horse must be entered for a claiming price of $25,0.

B) if the eligibility conditions for a race are for state bred $25,0 claimers then in order for a horse to be qualified to enter this race he must be entered for a claiming price of $25,0 and as well have been bred in said state

C) if the eleigibility conditions for a race is $25,0 clm n2L then in order for a horse to be qualified to run in this race he must be entered for a claiming price of $25,0 and as well must never have won 2 races in his carreer

D) if the elidibility conditionsfor a race is $25,0 clm n3L then in order for a horse to be qualified to run in this race he must be enterd for a claiming price of $25,0 and and as well must never have won 3 races in his carreer

E) if the eligibility conditions for a race are S$25,0 n2L then in order for a horse to be qualified to run in this race he must be enterd for a claiming price of $25,0 as well as be bred in said state as wel as have never won 2 races in his carreer

F) if the eligibility conditions for a race are S$25,0 n3L then in order for a horse to be qualified to run in this race he must be entered for a claiming price of $25,0 as well as be bred in said state as well as have never won 3 races in his carreer.

6 different conditions that a horse can qualify under. each added condition makes it harder and harder for the 'supposedly' better horses to enter into weaker conditioned races giving the weaker and weaker horses a chance to win a race. so which elgibility condition allows for the strongest compitition?
the open $25,0 claimer.
trainers can enter their horses with numerous wins as well as enter a horse with 1 win in 25 starts but if
horse A has 1 win in 25 starts, has been running his races in state bred $25,0 claimers which have never won 2 races lifetime and is getting nowhere
and is running against
horse B that has been running in open $25,0 claimers and has 4-5 wins in 25 starts and has also been getting nowhere then
which horse is meeting the horse that has been running in the weaker less competitive races? B and
which horse is meeting the horse that has been running in the stronger more competitive races? A
and in comparing only horse A to horse B only, which horse is taking the class hike and which horse is taking the class drop? A is stepping up to B and B is stepping down to A
so which of the 6 eligibility conditions should rank as having the strongest conditions to run against? and which condition would rank as having the 2nd strongest eligibility condition to run against? and so on

well to start
1) A is stronger then B C D E F
2) B is stronger then E and F
3) C is stronger then F
4) D is stronger then C E F
5) E
6) F is stronger then E

so i know that A is has the strongest eligibility condition here as A is stronger then B C D E F
what else do i know?
B is stronger then E F but is B stronger then C D
C is stronger then E but is C stronger then E F
D is stronger then C E F but is D stronger then B
E is not stronger then A B C D F. E is the weakest of the group
F is stronger thenE but is F stronger then C

so i know i have
1) A
2)
3)
4)
5)
6) E
___________________________

1) A
2) B or D
3) B or D
4) C or F
5) C or F
6) E
_______________________________

and as best as i can figure the ranking order i think the preference might be
1) A open clm 25,0
2) D clm 25,0 n3L
3) B Sclm 25,0
4) C clm 25,0 n2L
5) F Sclm 25,0 n3L5
6) E Sclm 25,0 n2L
A has 1 condition attached to it. B C D have 2 conditions and E F have 3 whatever thats worth and lest we forget theres the optional claiming nw that would also need to be thrown into the mix
but you see what iam saying here dave? theres no par times involved for this type of ranking. just ranked eligibility conditions that are representative of certain types of horses allowing these horses to run in certain types of races with 1 set of conditions being the strongest to run in and another set of conditions being the weakest to run in and all the rest being in between. basically its just class rankings according to the conditions of elidibility that iam looking for

[Dave Schwartz]

I appreciate what you are saying. However, you have pretty much stated the obvious.


In rank order:

$25k, open
$25k, N4L
$25k, N3L
$25k, N2L


However, where does $25k, sN4L fall?

Certainly below $25k, N4L but depending upon the state $25k, sN4L might fall below sN2L.

The logical question to ask is "What does it take in terms of speed ratings to win at $25k, sN4L?"

Isn't that logical?

How would YOU answer the question. And tougher... "How does a $25k, sN4L shipping from track A compare to a $25k,N3L at track B?"

Tougher still: "How does a $25k, NW@1m-90 days compare at track A with an open $25k at track B?"

How do you answer these questions without pars?

[freeneasy]

i fully understand what your saying dave that there are tracks, lesser tracks, where an alw n3x might grade out to be equivolent to an open $25,0 claimer at a more major track and that may be more in the vicinity of track class and i include this type evaluation in the system that ive been working on but for right now this is not the area of design that iam concentrating on. iam looking for a chart that lays out the equivolency value for all eligibility conditions and eligibility conditions only, found at major tracks.
most major tracks, sa, hol, dm, aq, sar, bel, cd, gulf, kee, will use justabout the same eligibility conditions as each other and its the strength that each condition, that is each eligibility condition in and of itself, carries that iam interested in finding a ranking order to.
i dont know if this is possible without the use of pars but i think i have a few ideas i want to start in a new thread that may bring this more to light. i know your doing what u can to carry me thru and in appreciation if my system turns out to be a winner i promise to split all my winnings with right down the middle 60-40 :faint:

[Dave Schwartz]

i know your doing what u can to carry me thru and in appreciation if my system turns out to be a winner i promise to split all my winnings with right down the middle 60-40

That's half right?

[bucktron]

TrifectaMike attached is something similar to what you are talking about. It is a table of final time Multiple Regression equations that are automatically created and used in a Home Brewed System. Because of the factors used, it is more of a rating standard than a par time table.

[TrifectaMike]

TrifectaMike attached is something similar to what you are talking about. It is a table of final time Multiple Regression equations that are automatically created and used in a Home Brewed System. Because of the factors used, it is more of a rating standard than a par time table.

Thanks for sharing. I like what you've done.

Have you attempted to use odds or odds grouping as a classifer?. For instance predictors for horses vs odds or odds grouping.

Thanks,
Mike