PDA

View Full Version : Price Predictor



fred_hill
04-04-05, 01:11 PM
What a wonderful topic .. Lets start a thread on this ..and thanks to Tommo for the inspiration to start this thread. It appears from the threads we may be underestimating the power and assistance this module may give us. On a personal basis, I use system analyser, but always do a cross reference using price predictor, and then make my final decision as to bet or not to bet. I use price predicter as the major influence but would be interested in how others are using this module. I will post a more definitive post on how I select later in the week when time permits and I get my thoughts straight on what I do. Look forward to more postings from Tommo on this subject as well .. fred

Far Canal
08-04-05, 12:32 AM
Hey Fred,

Not much action eh!!....well, here's something from left field that was put to me a couple of years ago by a prominent Vet and he used a Kiwi icon (Snell) so we'll substitute an Aussie icon.

"Herb Elliot was a great Aussie world record holder over 1500 meters, where would he get over the same distance to-day....probably would not even get a start, so why do you put so much importance on what a horse did 1,2,or 3 years ago, you should be concentrating on what that horse is doing now....after all, what the horse did 1,2,or 3 years ago is only an indication of what it could do to-day. There is not much difference between highly strung horses and humans in their prime, they are racing animals trained to win. Training situations change, as does technology including their food and quite frequently, let alone trainer changes, so concentrate now mate"

As I'm not a horse (I think) nor a racing human I figured that this Vet joker knew more than me so I concentrate on young, fit, horses that have not had the guts run out of them, so, 5 start form and system selector does me just fine with a passing glance to the others.

Wouldn't it great Fred if a few horses shared their opinion too.

Cheers,

F C.

john morgan
08-04-05, 06:54 AM
Hi All,

I take the opposite approach to F.C.

I believe the better class horses are trained to repeat winning patterns of previous preperations.

So in my approach I believe 10 - 15 start form is needed.

I look to see where better class horses (55 class and above in BS)peaked (won or finisshed within 0.5 lengths or achieved thier highest CIS Rating)in previous preps and try and determine if they are likley to peak today.

If a runner has run well in its last few starts I look back to it previous preps to see how long it held fittness for when it had run or run well last prep.

I think Brian Mayfield Smith said recently that he can't keep his best horses fit for more than 42 days.

Just a dferent approach for you to ponder.

Kayser32
09-04-05, 11:04 AM
Hi Everybody
I think this could be a good learning thread if everybody throws there thoughts in especially for relative newcomers to betselector like myself(had it for about 3 months - Its a fantastic product)
I use system selector mainly, though once ive got a selection I use price Predictor to analyse the form of the rest of the horses in the race.
Personally I try to use as few rules as possible - though someone on this forum has given me a lot of food for thought recently.
Most of my systems apply the BR ratings to the rules.
But one of my main personal criterias rightly or wrongly is that it must show some consistency over a long period of time. The first thing I did when i got Bet Selector was get 4 Years of data(Mainly Saturday metro files for SMBWA) I figure with a combination of rules regarding preparations, class of the horse and race as well as applying ratings, that patterns could occur over long period of time. The systems im using are ones that have shown a consistent S/R and profit each year over the 4 year test - These as I say rightly or wrongly are my aims for a system, They dont throw out a lot of selections generally between 5 - 10 selections a week, there not long shots so the profit isnt huge but any profit is a good profit- Ill take what I can get.
I look forward hopefully to reading everybodies ideas on there use of bet selector especially concerning rules(how many or how few? etc) and how much of a test period is long enough? or to much.

Especially from some of you more experienced guys your input is invaluable.

Cheers
Matt

YUCKMAN
09-04-05, 01:04 PM
I disagree with Far C,
Humans have cut a minute or so off the mile time, but horses have NOT!
The athlete of 2005 is a vastly differntly trained and tuned individual to the amateurs of Herb's era.

A horse running today over 2400 may not have started over similar distances or conditions for 7 or 8 starts- so you NEED to look back that far to ascertain it's ability over today's conditions.

I do believe horses ar creatures of habit- having been around a few (too many).
The smart ones remember the track they're running at, the jockey who kicked 'em in the guts- and even the barrier attendant who pulled it's ear to get him to go into the barrier.

Some like hot days, others wet days- they're only human, after all.

I digress(how's my spelling?)

Just check out how a horse performed 12 months prior- you'll be amazed to see history DOES REPEAT.
Enough to make our friend FAR C cough up his Resch's, perhaps!

DavidC
09-04-05, 01:12 PM
Hi Matt,

I use Bet Selector for systems only -- I use a different programme called "Bizzy Form" to do my formguide study.

I only have 15 months of data, including all metro meetings and a fair few Sunday country meetings, and naturally I wish I had more. But I've managed to construct about 40 system in all -- 20 systems for general metro racing (some based on distance, some by state, some by class, etc) as well as a system for almost every metro track. I've got separate systems for quinellas, trifectas, and first-fours, and half a dozen systems for country races. Most of my systems have around eight rules, including the combination of PR, BR and CIR ratings. I've found that almost all my systems are based on certain numbers of runners on "good" tracks over certain distances.

Your query about the appropriate test period for a system is valid. Most of my systems produce about one selection a week, yet one of them has had only a dozen bets. I can't stop myself from betting it though, because it's only been wrong twice. And one of those came 2nd!. Certainly not enough data to get too excited about, but it's good enough for me. I started betting on it for half a unit only, and the last three bets for the normal stake. What would you do in the same circumstances?

I think there may be merit in linking system success rates (strike rate and POT) to a staking plan. Surely I should be increasing my stake for a very successful system compared to a less successful one, shouldn't I?


David C

Far Canal
10-04-05, 01:33 AM
Hey Yuckie me ole mate,

You've just agreed with me "They're only human, after all"
And yeah mate, I do rely on repeat history.....but not half-a-dozen decades ago.

And Resch's, crikey ole sock, they have to drop that from a great hight ta see if there,s a little pop when it hits mother earth over there. And we won't go into humans cutting a minute off their mile times....thought Herb was a human not a V8.

F heehor C. :P