Paceline Selection

[shoeless]

It seems people liked the live discussion that was happening on Ray's
thread. I know this has been done before by FTL and Mitch but thought
maybe should be visited again. Let's face it this and getting contenders is the most important part of the Sartin process.
I am attempting to get feedback so the site can be more active if not thanks in advance for viewing this.


I know that people may not want to divulge their way of doing it but
thought I would give it a shot. I know this section is for programs and
probably this should be posted elsewhere but feel more comfortable here.

I use Pizzolla's ( I know elsewhere this can be a taboo subject) LASST
method described in his classic book Handicapping Magic. It is using the
last line unless one of these reasons appear

LAyoffs and their patterns

Structure of distance

Surface change

Trouble in the race



Feel free using any type of software to post your way of doing things here

Jeff

[DanBoals]

Hi Jeff,

I will jump in

I know Howard put out a book called The Pace Line in July, 1987 - and Ted is kind enough to keep it on this site.

In it he mentioned some things I found useful - like

a contender's second best paceline in the last 10 is often more predictive than the best

and

always use the LAST RACE unless there is a valid excuse to go back further

trouble line
off track
wrong distance
wrong surface
class level too high
first race of the track in the current meeting
bottom of form cycle
bad finish - but good early
poor early - but good finish

I don't know if Pizzola came up with these originally, but when his book Handicapping Magic came out in 2000, these were really old rules by then.

[MJS6916]

Hello Jeff and Dan,

I believe you both nailed how it's done.

If this helps any, in my last study of winners in 600+ events

I just let RDSS use the auto best perceptor last 3 setting......

....and the most recent line (# 1) showed up 46.7% of the time.

It was close enough to an even split between lines 2 and 3 back

at 29% and 24.3% respectively.

I'm curious to know if anyone else got an outcome

much different than that in similar studies.

Mike

[shoeless]

Dan, thanks for your input appreciated.


Mike, appreciate you sharing your study of best of last 3



Jeff

[DanBoals]

Hey Mike

That sounds correct to me.
I remember Howard wrote in one of the FollowUps or Manuals, that if you have to go back more than 4 or 5 lines, you should consider passing the race.

He talked to me about guys who were combining lines, which Synthesis manual entry allowed you to do, who were winning.

One thing that never seemed logical to me, was if a horse was coming off a layoff and you used its second line back - that might be from 9 months ago. But another horse that had been running recently, its 7th line back might be from 5 months ago. How do you reconcile use a 9 month old line while rejecting a 5 month old line. I have not done any type of regression testing on this, but I wonder what it would show.

When Howard told me about the combining of lines, I thought that would be a brilliant thing if you could master it, but if you made a mistake, that could be costly.

[MJS6916]

Originally Posted by DanBoals

Hey Mike

That sounds correct to me.
I remember Howard wrote in one of the FollowUps or Manuals, that if you have to go back more than 4 or 5 lines, you should consider passing the race.

He talked to me about guys who were combining lines, which Synthesis manual entry allowed you to do, who were winning.

One thing that never seemed logical to me, was if a horse was coming off a layoff and you used its second line back - that might be from 9 months ago. But another horse that had been running recently, its 7th line back might be from 5 months ago. How do you reconcile use a 9 month old line while rejecting a 5 month old line. I have not done any type of regression testing on this, but I wonder what it would show.

When Howard told me about the combining of lines, I thought that would be a brilliant thing if you could master it, but if you made a mistake, that could be costly.

Hey Dan,

I didn't read all of the follow ups,
but I do remember Sartin talking about
the risk of going further back into the PPs.

Better prices are found back there,
but you gotta have a good story
to justify going there.

To show one type of scenario which
shows up more in the low-end conditioned claimer ranks,
I have less of a problem going 4+ lines back
for horses who run 7+ times in the last 160 days
(didn't do a study, just guesstimating from memory)
if they just qualified for today's condition (which they couldn't get into
in the last 3 or 6 months or whatever)
in the last week or so.
Many times (in this situation), I've noticed with some outfits
their last 2 lines - 3 weeks or less leading up to
today's event didn't mean much; and the horse's best or 2nd best
in the last 150 days or so was useful anywhere in
their 3 to 8 lines back.

They're not easy to get winners from but
the lines are still good enough to match up against
today's competition.

Mike

[shoeless]

Dan and Mike, you both made good points here


Jeff

[shoeless]

Mike

I know this is a bit off topic but was wondering if you had any
thoughts which was better VDC or BL/BL

Jeff

[DanBoals]

Originally Posted by MJS6916

I have less of a problem going 4+ lines back
for horses who run 7+ times in the last 160 days
(didn't do a study, just guesstimating from memory)
if they just qualified for today's condition (which they couldn't get into
in the last 3 or 6 months or whatever)
in the last week or so.
Many times (in this situation), I've noticed with some outfits
their last 2 lines - 3 weeks or less leading up to
today's event didn't mean much; and the horse's best or 2nd best
in the last 150 days or so was useful anywhere in
their 3 to 8 lines back.

Hi Mike

Reading through the old manuals and followups, a lot of the PIRCO guys were solid going back to distant lines. That is a skill I have yet to master. For me, I seem to do a lot better picking recent stuff only and letting the program adjust surface and distance.

I am working on an AI Assistant that will hopefully be able to monitor what I am doing and then make note of when going back helped and when it hurt. I just do not have the time to pour over my records and even to keep all the records I need to keep, so am hoping an investment in AI will get a "second pair of eyes" that will help with the administrative tasks.

[MJS6916]

Originally Posted by shoeless

Mike

I know this is a bit off topic but was wondering if you had any
thoughts which was better VDC or BL/BL

Jeff

Jeff,

I look at Perceptor instead of BL/BL.

To me, Perceptor or BL/BL are good to
see if a runner I didn't have a high opinion of
may still have a shot in today's race.

VDC is helpful too, but according to my eye test
it's more late energy friendly (or 3rd fraction friendly)
than I would like it to be.
With VDC, I'm more interested in the number
than the ranking.

If I had to choose, I'm more confident using Perceptor.

I'll share some screenshots of what I noticed about
BL/BL and VDC as soon as I can.

Mike