Harness done the "Old" computer way.....

[LottaKash]

Hey All,... Back in the early and mid 70's, I bought a book, and it's been long out of print.....It was based on data that was collected, and anzlyzed by computer, by 2 gentlemen that are parobably resting peacefully by now....Now, keep in mind, that computing in the early 70's wasn't what it was today....The compact desktop computers of today are miniscule by comparison to what they were using way back when.....They were room-size machines using magnetic tape drives versus the HDD's of today........So, when this book first came out, it was a revelation of sorts, as up until that time, there was literally nothing to help the Harness Handicapper of that day.......

I used the method suggested to a great degree, but being lazy and smarter than these guys, I wasn't too, too successful....should've been, but, wasn't......

I would like to put out some of the findings of this study, piece by piece, and see how it may enlighten or interest any or many of you, when compared to how we do and see things today.....I will try to disclose as many of the tasty little tid-bits, in an ongoing basis, perhaps daily if time permits......

The first piece, is probably the most crucial of the lot, as I believe, based on my own experience and application, that much, if not all of the value, still remains the same, even today......

The following were the impact values determined by the computer study, and some of the values may be eye-opening to some, as the order and significance is somewhat different when going from a small track races to a larger track races......A small track is a half-mile track, and the larger tracks are the 5/8's and 1-mile tracks......I don't want to argue with this study, but I feel that today's 5/8th's mile tracks straddle somewhere in between the 1/2 milers and the 7/8's and 1-mile tracks.....As if I'm not mistaken the 5/8th's tracks of yesteryear had longer home-stretch's than those of today....I say this as the early speed is of almost as much important on the 5/8's as the 1/2-milers......

Keep in mind, that for both the 1/2's and 1-mile track that the applicable factors cumulate to explain roughly 70% of the outcome of any given race .....And what of the remaining 30% of the races ??......The explanations or excuses that were considered and suggested are these:.....

1. Horses feel, and do better on some days than others.....
2. Beyond the horses' attitude, on any given night, it may have come out of its last effort Worse for the wear,(Bounce?), or may have caught a cold or virus, enough to take the edge off, but not enough to keep it in it's stall.....
3. Today's racing surface, in ways subtle or obvious, can affect the outcome of a race in ways not always apparent in the past performance lines.......
4. While rare, good drivers make some big mistakes, and equally remarkable, truly poor teamsters at times rear up and drive like Herve Filion when fate drops the perfect lie in their lap......
5. The pace of any given race can botch up the cleverest pre-race plans..

This is called racing luck........


Half-Mile Tracks.

Rank....................Factor.................... ....Value.........................

1. Best adjusted mile time in last two-races..................60
2. Adjusted mile time in last race................................35
3. Post Position.......................................... .............24
4. Driver (today)........................................... ..........23
5. Best finish (by lengths ahead/behind in the last two-races........20
6. Class Rating............................................ ............20
7. Horse's Win Percent........................................... ..17
8. Best finish position in last two-races.......................15
9. Days since last race.............................................1 4
10. Finsh (by lengths Ahead/Behind in last race..............12
11. Finish Position in last race.....................................10

Larger Tracks

Rank..................Factor...................... ......Value

1. Best adjusted mile time in last two-races...................38
2. Driver (today)........................................... ............28
3. Adjusted mile time in last race.................................27
4. Best adjusted last-quarter time in last two-races........26
5. Class Rating............................................ .............20
6. Best finish position last two races............................19
7. Post Position (today)........................................... ..18
8. Horse's Win Percent........................................... ....17
9. Days since last race.............................................. 14
10. Best finish (by lengths behind/ahead) in last two-races.....12
11. Finish Position in last race.......................................11
12. Lengths gained/lost in the Stretch in Last-Race..........10
13. Finish (by lengths ahead/behind) in Last-Race.............7

Well that's the start of this......Food for thought, anyway......Is this the way you rank things.?

Enjoy,

best,

[Ray2000]

Well that's the start of this......Food for thought, anyway......Is this the way you rank things.?

LK

I too have played around with "Significance Factors" in many, many versions of computer programs to rate the horses. (Actually you can plug those numbers you cited directly into the Excel spreadsheet I mentioned in previous posts to see how it effects the Odds line.)
I finally settled on the standard 1/5 second (1 length) as the unit of comparison.
and adjust the following factors to what I believe is their "lengths" equivalent. The horses are then rated by summing these factors. Then significance of these factors then turns out to be:

Mile speed trend......24 %
Days off...............9 %
Driver.................8 %
Trip/bonus.............7 %
Med switch.............7 %
Post Position..........7 %
Post Bias..............6 %
4ThQ Time ave..........6 %
PurseSwitch............5 %
Break% Multiplier......5 %
Win%...................4 %
LastFinish.............3 %
TMaster speed..ave.....3 %
Trainer................2 %
TMaster class ave......2 %

The mile and 1/4 mile times are adjusted for Track Speed Ratings, Daily Variance, racing wides.
The Med switch and Break % factors are normally 0 so the significance of others goes up a little.

This ratio works best for me.

[Sea Biscuit]

Hey Lottakash, I remember it was in the late 70s or maybe early 80s I can't remember now, I took up classes trying to learn the Fortran Language. After you wrote your program down you were given a machine something similar to typewriters of yesteryears which would punch out cards for the program you wrote. Then you take all the punched out cards which would be anywhere from 100 to 200 (depending on the length of the program) in proper sequence and feed it into a room sized machine and get a printout from the tail end of the machine. If you ever made a mistake with a dot or a comma or a colin when you are punching the cards it was back to the card punching machine all over again. And if you dropped all cards (which I did once) and get them all mixed up, you guessed it...... back to the card puncher all over again.

Aah.... those young uns of today have it so easy with their desktops.

Hey did I give away my age writing this post.. Hehe

[Stick]

Ray
Could you give us an example of how these % are broken down to create a line in an actual race from any track tonight?

[mrroyboy]

I remember the old IBM 360. Took up about two rooms!

Anyway guys I really like this impact stuff. It needs to be adjusted for modern conditions but it's still some good stuff. What do you mean by adjusted mile time?

[Ray2000]

Stick, Mrroyboy

I'll do the 2nd question first.
(I failed to mention one adjustment above, post position.) and I use TrackMaster Daily Var which I say works best if you consider it in tenths of a sec, not fifths. There's a thread in this forum discussing this.*

Adjusted times take into account actual finishing time in seconds, say a 2 minute mile= 120.0 seconds. Then the speed rating for the track and the Daily Variance and the starting post turns this into something like

120.0 - 4.4 (mohawk) + Var of -10 (fast track that day) - 7/5 (Post 7) =115.2

Times are increased for fast tracks and reduced by 1/5 second for each post.
(1st trailer is post position 1.5, 2nd trailer is 2.5 ... )



=============


As for getting an odds line from the numbers, the % numbers in my reply above are "how significant" that factor is not the numbers used for ratings

Those numbers are more like..

Speed trend..........115.1.......varies best to worst ~4 sec
Driver.................2.2 ......I use USDRx5 meaning a .300 driver reduces time by 1.5 secs
Post Position..........1.0.......0.2 to 2.0
4ThQ Time ave.........28.3.......varies best to worst ~1 sec

No need to list them all.......

Adding up the numbers for each horse, I then subtract the mean from the individual rating and divide by the Standard Deviation. I use that number to lookup the odds from a chart. I only play 8,9, and 10 horse fields so the average horse will end up 7/1, 8/1, or 9/1. If you want the chart it's in the sulky xls file at
http://members.localnet.com/~rayschell/ or PM me.


Stick

An actual race for tonight would be (My longshot pick) Tioga Race9
table attached..

but please note ..
the table combines final race time with 4thQ time because I use trends that I didn't want to get into here.
and after summing the factors I subtract the result from 200 so now the high ratings are best
my odds are not trying to predict posttime odds, I do that differently looking for overlays


*
http://www.paceadvantage.com/forum/s...ad.php?t=46006

[Ray2000]

Mistake on adjusted time example 120+4.4+1-1.4=124.0

You want to normalize the time to a standard track, no var, post 1

[banacek]

Sullivan and Adams "Predicting the Outcome", I believe. I used the harness method for ages (especially on 1/2 milers), still use some of the adjustments. Didn't like the thoroughbred one much. I still have them around here somewhere.

[LottaKash]

Originally Posted by banacek

Sullivan and Adams "Predicting the Outcome", I believe. I used the harness method for ages (especially on 1/2 milers), still use some of the adjustments. Didn't like the thoroughbred one much. I still have them around here somewhere.

"Correctamundo"......29 pts will be added to your score.........

best.

[Ray2000]

banacek

njcurveball mentioned that book in
http://www.paceadvantage.com/forum/s...ad.php?t=48129
haven't seen it
the book I liked the most on thoroughbred strategy from the seventies was



The Science of Winning .........Burton Fabrican



Wall Street vs horse track market inefficiencies

[banacek]

Originally Posted by Ray2000

banacek

njcurveball mentioned that book in
http://www.paceadvantage.com/forum/s...ad.php?t=48129
haven't seen it
the book I liked the most on thoroughbred strategy from the seventies was



The Science of Winning .........Burton Fabrican



Wall Street vs horse track market inefficiencies

Yeah, I've got that one too..loved the comparison between the market and the races. But the method was quite tedious..have the earlier Fabricand one too - Horse Sense I think it was called.

[Ray2000]

Yes it was. He called his method "The Principle of Maximum Confusion", looking for the race where crowd was most likely to make a mistake.

He also gave the answer to the age old question "How much should I bet?" keeping in mind the problem of the Gambler's Ruin.

Answer was proven by some guy at Bell Labs, ...Bet a percent of your capital equal to your advantage in the game.

[LottaKash]

Hey Y'all, the next tid-bit from this OLD Computer study shows the signifcance of Early Speed at ALL Harness Tracks......

"Since it is speed that unocks good early racing position, we looked at the imprtance of good early raacing position using the results charts from roughly 1,000 races both a half-milers and larger tracks. These charts were analyzed to determine the racing position of the eventual winner at the half-mile pole."

HALF-MILE RACING POSITION OF EVENTUAL WINNER OF THE RACE..................

Half-Mile tracks.........

....................Total
Horses....................................Winners.......................
Posn at..........
1/2 Pole.......... No.,.....%.........No.......%.......base rate ***

1. ...1st......930.....13.0.....348.....37.4.....2.89.....***
2. ...2d-3d..1860....25.9.....294.....31.6.....1.22.....
3. ...4th......930.....13.0.....111.....11.0..... .92.....
4. 5th-6th...1848....25.7.....129.....13.9.... .54.....
5. ...7th......891.....12.4..... 39....... 4.2..... .34....
6. ...8th+....723.....10.1...... 9........ 1.0..... .10....

.........7182.....100,0... 930......100.o.........................

"This table indicates that while good racing position at the half-mile POST is important at tracks of all sizes, it is critical at Half-milers. While racing as deep as 7th at the midway mark on larger ovals hurts a horse's chances, it is not a lost cause; and, if the animal is 6th, it has at least an average shot. But at the same time, any horse in front midway through a race at a half-miler has things its way almost *** 3-times as often as you would expect by chance alone."


HaHa....I typed a whole lot more of this and the results of the study on Larger tracks, but somehow thru all this typing I hit the wrong button and lost it all, so, for now I will submit the 1/2 mile-track portion of the study and later on will update with the stats for Larger-tracks......crap....

Enjoy,

best,

[Ray2000]

Lk
Too bad you lost the stats on larger oval tracks (I thought I was the only one hitting the wrong button) but
here's the numbers Mohawk publishes for Winning Trips

quote
May 1 to September 8, 2008. Includes 1104 races. Table shows the position
of the winner at 1/2 and 3/4 points. Percentages not indicated for
4th through 11th as all races had full fields.

1. 2. 3. 4. 5. 6. 7. 8. 9 .10 11
- - - - - - - - - -- --
1/2
376 161 126 122 119 91 71 29 10 3 0
% 34.1 14.6 11.4
3/4
440 200 119 133 81 66 34 26 6 3 0
% 39.9 18.1 10.8
end quote

Being upfront early is of course critical to winning, the hard part is predicting who is going to fill those spots (at the half pole) in any particular race.

I'd be interested in how handicappers here go about making this prediction. Some may rely on driver tendencies, some on class drops, some on horse past performance lines, some will only consider race-sharp horses (thinking nobody sends a lame or sore horse "down the road"). etc.

Only fair to give my answer, so the method I use (computer uses) is to subtract the call at the half from the starting post (1st trailer is post 1.5) for all good pp lines, then take the average. If > 1.25 then he's a leaver, if < -1.25 then he's a closer, in between he's a stalker. My method ignores the driver and fitness and that's its weakness and the reason for my question.

Any thoughts?

[Stick]

Here is my question. Do horses that race up close have a better chance to win or do horses that are in form leave or make a brush towards the front because they are in form? In other words, when horses are dull do they tend to stay in the back or just sit along the rail? Then when they start to get back in form, the driver pushes the horse to an up close position. The stats are showing us that horses up close do better. But are they up close because they can win and the driver makes a move knowing that?

Take this horse's made up lines post/1q/2q/3q/stretchcall/finish
6-6-7-6-5-2
3-5-5-6-7-8
1-2-4-5-6-7

Let us say that all races were just rail rides.The last race was a positive driver change.The top driver is up again tonight with an inside post. This horse will show no way to predict an up close racing position at the half or 3q. He leaves, gets the two hole tonight, and wins. He was 2nd at the half and 3q and is another winner in an up close position.The fact that his form started to change and he showed an improved race will not lead us to a numbers search that indicates he will be up close tonight. That being said, any thoughts on how up close position can be analyzed and used in the PP to predict some winners would be a great help.