[AusRace] FW: The Co-efficient of Uncertainty (2) - a system

Fri Feb 24 07:39:03 AEDT 2023

Interesting Tony but a bit complicated for this simple brain.
Next assignment is to explain Mark Read's method of rating 'em!
Think he used a combination of times and weight carried, which sounds the
obvious way to go.

On Fri, 24 Feb 2023 at 00:26, Tony Moffat <tonymoffat at bigpond.com> wrote:

>
>
> Author Lowell Harbison (Munro) NZ - Part 8 Maths for Games
>
> As explained in Part 1 Mr Harbison obtained mathematical values for past
> and
> future runs for racehorses by using distance, weight, finish-1 in the
> sequence
>
> (a) Distance/ weight for future races
> (b) Distance/weight/finish position-1 for past form
>
> The variables obtained from these were used in an equation to determine a
> likely bet or series of bets.
>
> His workouts had schemes showing variables constricted to 1000m (he used
> furlongs, however, and pounds weight but offered multiples to use when
> converting - x200 and x.5).
>
> The use of the value 1000m (alright, 5 furlongs) was to show the 'output'
> of
> the horses exertion in the final part of the race. This output was not
> acted
> upon by any multiplier or divider, it seems it was simply a figure for 5
> furlongs (ok, 1000m then)
>
> Concerning the future race/past race scheme used in Part 1
>
> As a nod to the approaching decimal system (his words) the distance, weight
> and finish spots were used in a different way (than that explained in
> detail
> in Part 1) in a suggestion in his summary of his paper.
>
> It goes like this
>
> (a) FUTURE RACE-distance/weight over minimum - and he used 52 kilograms as
> a
> minimum - so 1600m with 59kgs becomes 1600/7 (7kgs over minimum -WOM) =
> 228.57 and again this value was subjected to the square root clause
> =sqrt(sqrt(sqrt(228.57) = 1.97
>
> (b) PAST RACE/FORM - 1600,59, 5 becomes 1600/7/5 = 45.71 then applying the
> square root clause = sqrt(sqrt(sqrt(45.71) = 1.61
>
> Again the decimal digit of the first part are divided by those (the
> mantissa
> or decimal digits) of the second to get a variable for an equation
>
> Today, the minimum is usually 55kg.
>
> NEW SUBJECT Price and position as a guide to ability in a future race.
>
> Mr Harbison wrote and emphasized this method of determining 'fitness of
> form' in a future race utilizing the finish position and starting price in
> a
> form race.
>
> A runner finishing 2nd at 9/1 is considered a good effort, under
> appreciated
> by most form followers. There are many examples in each race.
>
> Here's another, in a race just finished at Pakenham 23/02 Race 3 Maiden
> Shezatempest won after running 4 of 10 $101 in the form race. There was one
> other in that race to be considered under this system, #3 Chiara something
> which had been 5 of 9 at $17 last start. It placed at odds. Another, last
> prep and not considered, ran in an Oaks 3 of 11 at $16 in its form race,
> and
> this was too long ago, and too long in distance anyway. It ran second.
>
> I did cherry pick, back fit or some other cliché in that demonstration but
> the point is made - finish close up at odds is a form positive. Really, any
> finish, at any price, and anywhere closeup to the winner is a form positive
> but we need odds in this future race and this system gets them.
> Shezatempest
> is $34.30/5.30 - she was $8 place when I put my $15 on her. (Meganne, is
> that you?) I had $5 on Chiara Da Luna as a saver.
>
> Mr Harbison categorized his placed runners this way
> 1- showing better than 19.98%
> 2- showing better than 17.76%
> 3- showing better than 15.54%
> 4- showing better than 13.32%
> 5- showing better than 11.1%
> 6- showing better than 8.88%
> 7- showing better than 6.66%
> 8- showing better than 4.44%
> 9 -showing better than 2.22%
>
> The 9 includes all runners finishing 9 or worst (9-24th) Essentially if
> your
> runner has book percentage of 14.09 it would be ranked 4th and show an sp
> of
> 7.10. There are three horses shorter than it, 3rd and 2nd favourite, and
> the
> favourite, calculated at 2.50
>
> He categorized a runner at odds, (say 15 - 6.66%) and finishing 4th, it
> then
> earnt the value 13.32 (twice its earning). That differential (13.32-6.66)
> is
> used in pricing it to place in an upcoming race.
>
> The paper is based on NZ racing, so only tote odds are known and,
> agreeably,
> resulted as 7th favourite or similar (this was then, remember) It was a
> simple matter of reading the finish place and the tote rank (7th favourite
> in this case) and ticking a box or moving on to the next.
>
> In summary, the horses to consider are those who have form figures that
> flatter them, even though their sp prices consider them dunces. Presumably
> the 58 day since last race rule applies. Why 58? When 8 weeks is 56 days,
> may be international date line or something!
>
> In Pakenham R2 two runners were considered
> 15 Ardsley was 300 (in other words calculated from its price as 9, finished
> 6 @$61 - 9-3 =3 (*100)=300 -it was 148.50/22.90 so somewhat hopeful
> 11 Coast Princess was 200 (calculated from its sp to finish 5, finished 3
> so
> 5-3 = 2 (*100= 200) In Race 4 Brazen Lady 200 was scratched In Race 5 too
> many choices, 5214 featured, 5 won but short short In Race 6 10 & 9 were
> selections 2nd and 3rd and short In Race 7 1,2,3 were selections, placed
> but
> others disappointed In Race 8  11 and 7 selected, placed but short
>
> It appears to me that Mr Harbison has used too narrow a percentage gap
> between his selection criteria. In a book percentage of 100% he has a
> spread
> less than 2.22 to a top of better than 19.98. It appears, or it seems, to
> be
> a narrow band were dividends might be. Ok, 2.50 is 40% which is in the band
> 'better than 19.98% but so is 3.50, and 4.00, and 4.50,  and 5.00 and now
> you are into or near an 'open' race.
>
> For Pakenham today (see results above) the form winner average book
> percentage was 26.88 (winners prices totaled 1774.15% and there were 66 of
> them) =the spread however was $1.35 to $26 (74% to 3.85%). Normalised the
> book percentage was 22 (calculated back to 100% from a carry of 122%)
> Horses
> at that meeting running 7 th in their form lines 10.35/8.31 when normalized
> and that div spread  was $101 to $2.1 Mr Harbison used 6.66% as the mean
> for
> 7th whereas today it was 10.35% - perhaps that is the issue I have with his
> 'narrow' selection gap
>
> SUMMARY
>
> The first version of the selection system used distance divided by weight
> for future selections and distance divided by weight divided by finish
> position to allot a value to finishes in the form line of a runner
>
> This was amended later to distance/divided by WOM (off a 55kg base, 53kg in
> the paper) divided by finish position to ascertain a form line value for
> past runs.
>
> The author uses a form analysis where he compares a horses finish position
> to its expected finish position calculated from its sp - looking for a
> horse
> with a long price which finishes well, it performed better than punters
> thought it ought to and this information can be used to select a runner in
> the future race which may be at a long price and win, or place more often.
>
> Another way - probably
>
> Cheers
>
> Tony
>
>
>
>
>
>
>
>
>
>
>
> --
> This email has been checked for viruses by AVG antivirus software.
> www.avg.com
>
> _______________________________________________
> Racing mailing list
> Racing at ausrace.com
> http://ausrace.com/mailman/listinfo/racing_ausrace.com
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://ausrace.com/pipermail/racing_ausrace.com/attachments/20230224/81243072/attachment-0001.html>