Hi lenny.
My data goes back to 1999. I you would like any checks, I would only be too happy.
Might even sling you a system or two.
Rgeards Enjay

Stebbo, Lenny et al

For me PRICE is the supreme indicator (factor, parameter etc etc) and is indeed the only thing one can buy or sell (BetFair) within the punting marketplace (ok a First Four ticket might buy a dream). Even if your not glued to the Tote screen Stebbo you are still a slave to the price flashing on the screen - provided your selection finishes in the money.

I have made reference before to Benter's Hong Kong foray and his findings that PRICE was the most influential factor amidst 21 factors that he applied to handicapping a race. Given the Hong Kong market is considered relatively free of the Longshot bias I would conjecture that markets positively burdened with longshot bias (ie Australian race markets) are even more prone to the influence of PRICE with respect to profitable outcomes.

Allow me to offer two examples.

My place system hinges on favourites sub SP 2.5 with a couple more filters (ie field size). As PRICE exceeds 2.5 so do returns and profit evapourates into loss. PRICE is critical.

Better Ratings offer a positve return no matter the PRICE - but simple price filters not only increase returns but rid the results of long losing soul destroying ROOs.

To ignore PRICE is like going to Church on Sundays and ignoring God - it is the reason you are there and you simply must pay homage in one degree or another or suffer the consequences.

Hi Enjay,

Thanks for the offer! I will keep that in mind now, and probably send you some ideas to test. Your offer for some system ideas would be great, too!
~~~~~~~~~~~~~~

Hi Dr Pangloss:

Nice to hear from you. Your posting was both useful and amusing. It confirmed to me the essential nature of prices. For me, I can't seem to avoid an obvious relationship - the greater the price obtained, the lower the system strike-rate. BTW thanks for the tips.

What is Benter's foray you mention? Is it referred to somewhere on the net? Also, I know I'm a bit green, but what is the longshot bias you refer to?

One other thing - are you an Aussie? I only ask because you mention Betfair - from what I have read on the net, our Government is trying to ban us using it. Any suggestions for avoiding this tedious inconvenience?

Cheers,

~Lenny


[ This Message was edited by: Lenny on 2004-01-16 19:16 ]

[ This Message was edited by: Lenny on 2004-02-14 03:13 ]

please.... no more longshot bias posts... now all we need is JFC to come out of retirement to join in the fun..... :roll:

Lenny,

In a nutshell "longshot bias" is where punters tend to overbet on longshots and underbet on short priced favourites so that longshots pay less than they should and favourites pay slightly more than they should. If the TAB market was accurate you would expect that all price ranges should return 85% to the punter (allowing for the 15% TAB cut) so you should see a loss on turnover of 15% if you backed every runner in a particular price range. If you look at the actual results for ALL runners in a particular price bracket what you find is something like this:

Runners $2.00 and under return around 95%
Runners $2.00 - $4.00 return around 90%
Runners $4.00 - $15.00 return around 85%
Once you get over $15.00 the return declines - when you get to the real rough chances ($100 and over) the return is down to about 50%

The above figures are a rough indication - the actual percentages vary a bit depending on whether you look at metropolitan racing only or all races, whether you use bookie starting price or TAB prices, which TAB, etc but the general trend is exactly the same.

_________________
"Computers can do that????" - Homer Simpson

[ This Message was edited by: becareful on 2004-01-17 09:49 ]

Lenny

William Benter spent about five years fine tuning a handicapping model applied to Hong Kong races (no bookies). After achieving outstanding success he disclosed his methodology for academic review that entailed,

"A logit-based technique and a corresponding heuristic measure of improvement are described for combining a fundamental handicapping model with the public's implied probability estimates. ....This paper aims to emphasise those aspects of computer handicapping which the author has found most important in practical application of such a system."

Allow me to continue to quote. "It can be presumed that valid fundamental information exists which can not be systematically or practically incorporated into a statistical model. Therefore, any statistical model, however well developed, will always be incomplete. An extremely important step in model developement, and one that the author believes has been generaly overlooked in the literature, is the estimation of the relation of the model's probability estimates to the public's estimates, and the adjustment of the model's estimates to incorporate whatever information can be gleaned from the public's estimates."

In his academic review of Benter's 21 variable handicapping model (tote odds being the 21st variable) CHAPMAN reports in his paper, 'Empirical results from 2,000 Hong Kong Races', that - "Statistically, these results are absolutely clear. Even with an elaborate and sophisticated 20-variable pure fundamental horse race handicapping model, important incremental explanatory power is captured if it is possible to include the public's win probabilities as an additional predictive factor."

The bottom line is that the Starting Price (or closing SP) contains vital information that will impact on wagering returns. With respect to PPP it seems to me, amongst other things, that market information may dictate when to bet, and when not to bet.

Thanks to BC for the long-shot bias explanation.

It remains a mystery to me that someone with such articulate and mathematical skills would base any conclusion to a model on 2,000 races! (Benter not Dr. Pangloss)

That being said, all that would hypothetically remain, is to use a spreadsheet or like, that automatically downloads the updated TAB prices or IAS or other, and the spreadsheet would then incorporate this information into the already formulated preset calculations and update said calculations.

According to Benter's findings this would hypothetically make the model complete.

A cell in the spreadsheet could automatically highlight "BET" or "NO BET".



This is the headache for any racing analyst using data per se, even with the price criteria which is a very strong criteria, there are OTHER fundamental criteria which can not be measured or have never been measured and escape even the most astute mathematicians.
Of course one can always factor "luck" into the equation of value and price, but I've become increasingly aware that there are "x" and "y" factors which are unknown and which escape me for the time being.




[ This Message was edited by: Chrome Prince on 2004-01-18 16:11 ]

Hi All,

Stebbo: Please, forgive a punting pagan his sin of ignorance... :smile:

Becareful: Thanks for taking the time to clear up "long-shot bias" in detail. I still feel that longshots are a potentially most profitable area if you can stand the long run of outs. Perhaps I'm taking up the baton from a long dead argument, but the gut is talking here...

Dr. Pangloss: thanks also for taking the time to post that information on Benter's research in detail. It is certainly very interesting, and I'm going to see what the 'net has on Benter for my own research and interest. I have always found the pre-post market to be remarkably perceptive - your post adds meat to that bare-bones idea. Cheers.

In another post I mentioned I wanted to determine what is the expected spread for an "accurate" set of form ratings. Perhaps there is still legitamacy in this idea.

~Lenny

Hi Enjay, I don't suppose that offer would be open to me too??? I've got quite a few system ideas that I'd like to put through a long term test.

Hi Sportz,

I've got just over 2 years worth of data if you'd like something checked... email me offlist....

Cheers,
Chris.

__________________
Permanence is an illusion