Proportional Bet Sizes

[stebbo]

Hi All,

there seems to be two schools of thought on sizing bets according to the odds on offer..

crash:
Big overlay bigger bet

jfc:
Your Odds are increasing so your bets should be decreasing

I'll state up front that I do neither. However, were I to do so, I suspect I'd do the former... The larger the overlay the larger the bet.

Personally, I can't see the sense in reducing my bet just because other people don't like my selection. If I've done the form, and I make a selection that I think is the best horse in the race, why would I change my mind and my plan just because others don't agree with me?

Comments?

Cheers,
Chris.

[Dr Pangloss]

'afternoon Stebbo

In his paper "Computer Based Horse Race Handicapping and Wagering Systems" William Benter states:

"...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 generally 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 other words, ignore the market at your own financial peril.

I might add, for those unfamiliar with the work of Benter, he developed a computerised handicapping model for Hong Kong racing and reaped substantial profits from pari-mutuel win, place and exotic pools using fractional Kelly. He speaks with the advantage of profitable hands-on experience.

[puntz]

Dr Panglos,
does the book say anything about re-framing your market to 100 % if it's over 100 % ?
If so, interested what it may have to say .

[Dr Pangloss]

Puntz

The information is from a paper appearing as an addendum in Nick Morden's, "Winning Without Thinking."

My understanding is Benter framed a market to 100% before launching himself at the Hong Kong tote.

He had a few of advantages over the Australian racing scene.

1. Massive tote pools - superstitious punters
2. Discrete horse population
3. Longshot bias non-existent

He has since left Hong Kong but I understand more than one syndicate (and their computers) are now exploiting the tote pools there now.

[umrum]

dr. why is the longshot bias non existent in hong kong?

[crash]

With so many punters there Umrum, everything is a favorite [get with the program mate !].

Cheers.

[Dr Pangloss]

umrum

such a simple question, such a complicated answer, err..... maybe no answer.

First of all, reasons for a long-shot bias include:

1. Bettors derive utility simply from the hope associated with holding a ticket on a long-shot as it is more FUN to pick a long-shot to win over a favourite and this has more bragging rights. Witness the great mass of no-hoper 'tips' offerred on this Forum.
2. Bettors chose horses for irrational reasons such as colour, names, numbers etc etc.

A recent Japanese study found that racetrack market efficiency increased relative to increased turn-over. See http://pareto.ucalgary.ca/~wdwalls/papers/lvw.pdf

In other words, as more money is bet on an event the market trends toward efficiency. Ergo the prodigious Hong-Kong betting pools promote efficiency. Please don't quote me.

<font size=-1>[ This Message was edited by: Dr Pangloss on 2003-11-02 23:57 ]</font>

[puntz]

Wait on,
If I am to use a method of selections, and this method converts my rated selections to odds, then I have my own "framed market"
right ?
Then if my number 2 pik is 3/1 on MY market, and it is 8/1 on the public market, and I have faith in my method of selections,then who cares about bias ?
As long as I have 3/1 and bookie is paying 8/1, who cares ?
If I want to make 300 on that runner, I'll bet 100, right?
if they want to pay me 8/1, that is THIER problemo.
But, if I had 3/1 and the bookie is offering 2/1, no bet as far as I am concerned.
Right?

[Dr Pangloss]

Puntz

Let's reframe the scenario.

Say you rate a selection at 3/1 and the market offers 100/1 and you're staking fractional Kelly. Ignoring the market in this context will lead down the path to the poor house quick smart.

So the question is not whether to utilise the market or not, but rather, at what level should the market influence your bet, and by what amount. e.g. Rated 3/1 with 7/2 bet so why worry adjusting rating or bet size. Rated 8/10 with 12/1 bet you might want to adjust rating and bet size accordingly.

To ignore the market is to ignore the single most accurate predictor of a race outcome. But let the experts speak.

Benter's work was reviewed by Chapman in "Still Searching For Positive Returns At The Racetrack" and under the heading 'Extending Fundamental Modelling: Does the public have special insight' the following was written:

"Statistically, these results are absolutely clear. Even with an elaborate and sophisticated 20-variable pure fundamental horse racing handicapping model, important incremental explanatory power is captured if it is possible to include the public's win probabilities as an additional predictive factor."

Good luck.

[puntz]

Dr's quote:
Let's reframe the scenario.

Say you rate a selection at 3/1 and the market offers 100/1 and you're staking fractional Kelly. Ignoring the market in this context will lead down the path to the poor house quick smart.

So the question is not whether to utilise the market or not, but rather, at what level should the market influence your bet, and by what amount. e.g. Rated 3/1 with 7/2 bet so why worry adjusting rating or bet

**
Reply:
First of all, "adjusting the rate" is not applicable, it's past that. The "rate" is what I used to work out what horse to bet on.
It could be expirience, it could be any way what so ever one chooses.

So if my TAB: 1 is rated at 95 and out of 100 it is my 2nd pik. (My highest pick may have 98.)
But, the market I want to bet on, it does not show to be of the same favouritsm as per my selection method. So the market is paying 8/1 or, your 100/1

You say to set a limit, the only limit I can see to work a base on, is the variance between the difference of the 2 prices.

Means, if I am 3/1, and the market is 8/1, then probably worth a bet.

If on the other way you say, I am 3/1 and your market shows 100/1, then sure, question the discrepency.

So then, does "bias" have a percentage value, + or - ; and if so, accept fact and bet accpordingly?

[stebbo]

Hi Doc,

when taken over the averages of all races, then the public will arguably get it about right... However when looking at each and every single race, I don't know if I agree that the public is necessarily right.

Ignoring rated prices for the moment, if I select horse A as the most likely to win a race, and it starts at $20, and it wins, then I'm right, but there's two scenarios

a) I'm right and the public at large was wrong, or
b) I'm lucky

If I do this consistently enough on a specific class of race or with a specific type of horse, then "lucky" is less likely, and perhaps I have an edge.

There's also the issue of whom to listen to. If my selection is at 8/1 on the bookies, but $20 on the tote, what are it's chances...

28% because traditionally 28% of my selections win?
14% because 14% of my selections starting over $10 win?
11% because the bookie has it at 8/1?
5% because the tote has it at $20?

I have trouble accepting the tote price. This price is set by the public, 95% of whom are going to lose money over time. If you have a mate who can't pick the winner of a two horse race, would you listen to his picks?

Cheers,
Chris.

[Chrome Prince]

It comes down to whether your rated price is accurate or not - simple as that.

IF it's accurate and they bet 50/1 about it, then don't ignore the overlay. Even if it doesn't win, over time you will make money.

The public market is the most accurate guide generally - but it's certainly not accurate all of the time.

[crash]

My rules of the punt:

1. Market price is set by the punting public and nobody ealse.

2. Pick the horse you think will win. If you can't split two selections back both if price allows. If more than two, move on to another race.

3. Check the public [tote] price of your selection and bet the horse if it provides a price that will produce a profit at your SR.

4. Know your strike rate.

It's that simple and has kept me afloat for years.

Cheers.

[puntz]

but the market has to be checked and corrected, if it's 108 %, correct it to 100%
then check against your own 100% true odds market.

[Dr Pangloss]

puntz and stebbo

You have raised some good points that I will attempt to cover by further explanation rather than specifically as raised.

Profiting from horse-racing is the process of exploiting the difference between REAL ODDS (Market) and ACTUAL ODDS (rated expectation).

In arriving at a Rated Expectation most punters will adopt a regime of comparing VARIABLES such as Jockey, Track, Distance, Barrier etc etc all of which are considered PREDICTIVE FACTORS in their own individual right.

Benter's handicapping model originally considered 20 VARIABLES used collectively to predict a race outcome. I now quote as follows, "Adding in the public's estimates of the win probabilities as proxied by the log of the public's win bets" (end of quote) as the 21st VARIABLE greatly improved the model's predictive quality.

Furthermore, when weighing up the quality of each VARIABLE's 'predictability', it was discovered the original 20 VARIABLES collectively accounted for 78.6% of predictive value - the public's estimates accounted for the remaining 21.4% all by itself.

Leaving out the MARKET from betting considerations will not necessarily turn your system into a losing proposition - but the opportunity to improve overall results is available by simply eavesdropping on what the Public is thinking.

[stebbo]

Hi Doc,

your point is well made. I can accept that on principle we should take note of the market. The question now becomes how do we take note of the market?

I don't currently have the expertise or the desire to frame my own markets. I think that this exercise requires more knowledge and information than I currently possess. I know that I need to acquire these skills, but they are a way off yet. The only set of rated prices that I have access to are not necessarily relevant to my selection method, and my simulations have shown that if I were to use them and bet proportionally I'd be worse off than flat stakes. So the question remains, how to utilise the info the market is offering?

One idea that I begun to formulate lately is to take ranges of SP's, and bet them as separate banks using flat stakes and a percentage of highest bank approach. I'm confident that each of the ranges is a "profit" center in itself, as this is one of the tests that I do on prospective systems. Approaching it in this manner allows bet sizes for each SP range to "progress" at their own rate, and perhaps at least try to take advantage of the market info.

Comments?

Cheers,
Chris.

[puntz]

Dr Panglos,
yes, that is part of my point, "what the public is thinking".
Next scenario, and it's what I look for.
Extreme case,
if then you use the exact same syetem/method of selection as you stated above, and I use the same system/method above,logicaly we should both to the exact same conlclusion.
If TAB No 1 then calculated 3/1
and if 12th runner's rates come to 25/1, and frame my market as such'
TAB 1 odds :3/1
2,3,4,5 etc
TAB 12 odds : 25/1
and you are a bookie, what stops you from swapping around the odds to 3/1 on 12
and 25/1 on 3, therefore altering public opinion to a point?
Note, I am only showing a principle, but many times, not all the time I find this to occur, a discrepancy in ratings and odds.
Becuase, if a highly publicised runner is, publisised, it seems to shorten, and yet, it loses.
Therefore, to a point one has to discard some public opinion and use a combo. of ratings and market to see what the others are doing.
Bias, I believe could probably be viewd at that point. So we are back to square 1, how much bias is permited?

[crash]

I can't believe what I am reading here.

You are all asking or trying to work out what the market is thinking [?] LOOK AT YOUR NEAREST TOTE SCREEN. THAT IS WHAT THEY ARE THINKING [it ain't rocket science].

If the market is offering 6/1 and you think 3/1, that is your edge to exploit. How do you work out 3/1 ? Practice and belief in yourself, your intuition and hadicapping methods [there ain't no exact formula or anybody you can call]. You will soon know if you are getting it right often enough. It's called profit. No profit ? Practice until you get it right [or take up stamp collecting].

Ever heard of the KISS principle? [keep it simple stupid].

Cheers.

Yowza, yowza, yowza !

<font size=-1>[ This Message was edited by: crash on 2003-11-03 22:24 ]</font>

[puntz]

I have no problem whatsoever with my punting.
I may have some issues, but definately no problems with understanding what I do.
All I am saying, are scenarios I have eluded, for that same very reason as you,
Keep_It_Simple_Stupid.
But I just want to know if these avenues I elude, are worth discussion and exploiting.
Nothing wrong with that.
mathermaticaly speaking, is there a way to measure bias? If there is, then it can be tested and if an advantage, it can be utilised.
I see nothing complicated about going beyond the norm.

<font size=-1>[ This Message was edited by: puntz on 2003-11-03 23:24 ]</font>

[crash]

Puntz,

I wasn't meaning to offend or highlight anybody, but there is no enquiry needed to work out what other punters are thinking and cetainly no need to ponder how.

There seems to be an obsession with trying to pin race selection methods down to a science on a spreadsheet [trying to measure everything and turn it into maths. it seems].

Track bias Puntz, if it could be measured [good question] would be useless, as bias is never a constant but like shifting sand beneath the feet: [rail out/rail in. Wet track/dry track and how wet or dry ? Windy ? How windy ? Field size ?]. What can be measured [constants] already is measured and most good form guides already have that info. in them. Systems are no more than various forms of Handicapping [without the Art].

There seems one discussion always sadly lacking in this forum and that is about sound hadicapping methods and statergy [?] It's not exciting stuff but it is the starting point to winning more races that any amount of the crystal ball gazing going on about maths solutions to variables.
The maths takes us so far and then intuition must take over the rest of the way and that happens to be most of it.

The more that is measured beyond known facts [starts at the distance for example], the more out of wack collective results get [ask any carpenter and he will tell you, " measure twice and cut once"]. Ratings are a good example of this cummilative addition of measured variables getting out of wack.

Selecting winners is a combination of sciense [good handicapping of the very few measurables amongst huge amounts of variables] and art [intuition] gleaned from the experience of observation and understanding of variables and how they relate to each other in the selection process.

In my view, at some point very early in the selection process, art must take over to narrow down selections, risk, value and profit potential etc.

If Racing could be reduced to mostly science, every Kid on the block with an Excel spreadsheet would be a millionaire !

Cheers.








<font size=-1>[ This Message was edited by: crash on 2003-11-04 08:46 ]</font>

<font size=-1>[ This Message was edited by: crash on 2003-11-04 08:50 ]</font>

[puntz]

crash, no, I am not offended, but I have time to ponder these things and experiment.
Not that much time, but it makes some interesting number crunching.
Yeah, it may be a form of science ( and fiction), but it is interesting to note, that some discepancies exist. There is no way however to do much about it init's raw form, but program can check stuff like this.
If a program can check these pesty little items of math, then I see no reason to not find out what is there.
Personaly, I could not be bothered checking every race before I bet, on bias issues.
It still does not explain though why one has a rated horse at 3/1 and the public market is 8/1, or 100/1
The question is, if I/you/anyone has 3/1 on their books, and PM is 8/1, then is that difference between the 2 a bet ?
I say it is, I say go to 15/1's
A variance or "bias" factor of 12.
If you want a factor bias of 97 ( 3/1 and 100/1), then would you still bet ?
These stats. I believe have not been done.
If it is a bet, and loses, then the kelly syndrome may come into play for the next bet. IF one chooses to use a kelly method.
If one chooses to use a kelly staking plan, will it hold? or, does it have it's place here on the punt long term?
I thinlk not, but I have not looked close enough on how it works it's sums yet.

[puntz]

If someone has the database.
Randwick Race Race 4,Sun 26 Dec 1999
Do your stuff, rate that race, and then see.
I had it 3rd on my list of selections that day.
It paid 14.90
The point is, if one uses a rate method and convert to odds, what do they get on this particular race ?



<font size=-1>[ This Message was edited by: puntz on 2003-11-04 02:14 ]</font>

[crash]

Puntz,

If you say 3/1 and the public [tote] say 8/1, one of you is wrong. Recognizing when the public is wrong and exploiting it is what the game is about. Sometimes you will be right and sometimes wrong. Even a public 100/1 is wrong now and then. It would be great to be able to spot that wrong conclussion though and bet bravely !

Cheers.

[puntz]

Yes, bet bravely...trust the system one uses to select.

[crash]

Right on Puntz ! Don't let anyone ealse control your money.

Cheers.

[stebbo]

Hi Crash,

I agree 100% that you shouldn't let others decide your selections. The intent of this thread for me was to determine if other people's selections (ie the market) should determine bet sizes.

In your first contribution, you stated if you rated your selection 3/1 and the market rated it 6/1 then great, you've got some real edge.... What about if the market rates your 3/1 at 30/1?

a) are you still as confident about your selection?
b) would you be tempted to change your bet amount?
c) if you were tempted, would you change it up or down?

At the moment I don't change my bet. I have been thinking of a multi-tiered bank approach which would give different bet sizes to the different tiers based upon performance, but haven't gotten around to modelling it yet.

However, I was flipping through some back issues of "that mag" last nite, and came across a quote from Alan Potts which basically says if a selection isn't worth your usual bet, then it isn't worth a bet at all.... Interesting stuff.

Cheers,
Chris.

<font size=-1>[ This Message was edited by: stebbo on 2003-11-04 09:19 ]</font>

[puntz]

Little left for what ain't controlled these days.

<font size=-1>[ This Message was edited by: puntz on 2003-11-04 09:28 ]</font>

[crash]

Stebbo,

I have faith in my betting decisions and I get it wrong like anybody ealse.

If the pulic say 30/1 and I say 3/1, I will assume I have overlooked something. After a good seach and I spot somthing I wont bet it.
But if I find nothing except what I think the public has missed [and they miss a lot]I would bet hard on such a huge overlay. I do it often [rarely on a 30/1 public mistake though] and it pays off often enough to put me well in front over the years.

I am doing it again today on Tumuric [Perth cup winner and perfect prep. with Shane Dye winging over from H/K for the ride for Freedman] who last won at big odds at Flem. 40/1 on 6/03 and he was in my book with the note. " has a habit of getting up at big odds" [I missed it but William didn't !] I will probably do my dough as I usualy do in the cup [last winner was a 3 horse boxed tri. with Rainlover] but then again...

Ask yourself how many times timidness takes over on a bet when faced with a large overlay, only for it to get up and win ?

If you gave someone 3/1 odds on the throw of a penny for a set bet of $20, how many would take the bet ? Everyboby ! If you offered the same deal but the min. bet was now $1000, how many would take the bet ? Hardly anybody. That exacly the deal with overlays [I'm not suggesting betting more than you can afford to loose].

Normal human tendancy is to bet smaller the larger the odds or overlay. Hard habit too kick but it's not rational profit wise, in fact it's a crucifix.

Cheers.

<font size=-1>[ This Message was edited by: crash on 2003-11-04 11:39 ]</font>

<font size=-1>[ This Message was edited by: crash on 2003-11-04 11:43 ]</font>

[Chrome Prince]

On 2003-11-04 11:31, crash wrote:
Normal human tendancy is to bet smaller the larger the odds or overlay. Hard habit too kick but it's not rational profit wise, in fact it's a crucifix.


Hi crash,

Never a truer word was spoken!
Self doubt tends to creep in - did I do it right?
Surely I must have missed something?

[Dr Pangloss]

Stebbo

Goodness me - you started a bushfire of a thread here. Keeping things in context, the example given of Benter's Hong-Kong racing exploits was useful because it specifically addressed the original question of adjusting bet size. Instead of prattling off racing anecdotes it seemed to me that the 'feet over the fire' experience of Hong Kong pari-mutuel betting would be indicative in weighing up the value of 'market' information. How to then apply this knowledge to a 'system based' approach remains a vexed question not to be ignored.

puntz

I think the bias you are refering to should be handled thus. Firstly, an allowance for the 'take' (15% in Australia) needs to be made to the market price, and secondly, adjustment made for any inherent 'long-shot bias' before the 'real' market price can be discerned and acted upon.

crash

Cripes - there were times when reading your posts I thought this guy has definitely NOT read the proceeding posts. I concur with the spirit of keeping things as simple as possible - BUT NOT ANY SIMPLER. The original question posed was simple enough, but anything less than a vigorous statistical analysis (performed in Benter's and Chapman's papers) on real live betting experiences would hardly suffice.

I see no advantage in the process of dumbing-down the general punting populace, except of course to the Bookmaking fraternity.

[kenchar]

Crash.

Normal human tendancy is to bet smaller the larger the odds or overlay. Hard habit too kick but it's not rational profit wise, in fact it's a crucifix.

NO TRUER WORD HAS BEEN SPOKEN

Cheers

[crash]

Dr Pangloss,

I made sure I read your post very carefuly as it is true, like all of us I too am guilty of occassionly 'skimming' posts but usualy the problem is being unable to view and respond at the same time as for some reason my 'topic review' does not work.

'Simple' is a subjective word. In relation to my 'Kiss' principal you are unfairly useing innuendo to relate 'simple' to mean stupid ['dumbing-down']. Obviously then, in keeping with the rationality you have used, 'complex' means intelligent [or perhaps even eludes to punting prowess] as I suppose adding 'Dr' to a forum name inspires the same thought from readers [?].

Your romance [correct me if I am wrong and I am not using the word unkindly] with Benter style statistical analysis as a way forward in racing is spot on thinking. If only we had the computing power and the very big bucks, it would work a treat ...in Hong-Kong.

Comparing Hong-Kong raceing with the Australian experience is like comparing Derby day with a Dapto dog meeting [actualy was there by accident a couple of weeks ago]. Apples and Oranges would't you think ?

All would agree that meaningful stats. require a large volume imput of variables. The greater the variance of possible error the larger the volume needed. As for trackwise [or the horses running on them] stats. Oz and H/K cannot be compared as we have hundreds of them to consider and Horses runs on various numbers of them. As for large volume public betting stats. in Oz., where is the volume ? The only 'volume' here in Oz. are flies in the outback and outdoor dunnies.

We simply lack large public betting volume and a very small amount of track numbers in Australia for statistical analysis to be anything more than a general guide to intuition or a way to arrive at [and flog] dubious stats. and rating figures.

The XL road to riches in Oz racing is strewn with victims and very few success stories and of those all of them would have done well on the Punt regardless of spreadsheets due to superior intuition of the game. Kenchar lives off the punt [the only contributor here who does I think] and I know he uses the 'Kiss' principal.

When given more importance than deserved as an aid to handicapping, I feel this whole Stats. thing is no more than a form of pathetic snobbery. A wanna-be way to 'rub sholders' with punting success. Quite sad really [and lonely].

The 'public knowledge' bet [and what a last min. plunge that was] in the M/Cup came dead last and only one 'favorite'[odds-on] home in the last race in a card of 10 for the day. 'Public knowlege' score card: 1 out of 10. [Go straight to jail and do not pass go or collect $200].

Cheers.



<font size=-1>[ This Message was edited by: crash on 2003-11-05 11:06 ]</font>

[puntz]

Dr Panglos,
that figure of 15% is roughly what I was looking for, not the actual figure itself,but at least you for one has understood my question and responded.
Thanks

[stebbo]

The XL road to riches in Oz racing is strewn with victims and very few success stories and of those all of them would have done well on the Punt regardless of spreadsheets.

Personaly I think the whole Stats. thing here is no more than a form of pathetic snobbery. A wanna-be way to 'rub sholders' with punting success. Quite sad really [and lonely too].

Hi Crash,

with all due respect, I reject the notion that stats can't help with the punt. Statistical analysis of winning factors has helped me enormously with the punt, and has turned my punting around from losing to winning.

Statistical analysis of staking has allowed me to structure my banks accordingly, and has shown me why even with a successful selection method I sent two banks broke.

Could I have done all this without stats? Perhaps, but I think it would have taken a long time, and I probably would have run out of punting money before I got there.

Cheers,
Chris.

[becareful]

Sorry Crash but it is definitely possible to make a profit in OZ using nothing but statistical analysis of form and price. I almost never go to the track, never take any notice of the punting mags or even the racing section of the newspaper (usually read it for interest sake but it has zero influence over my betting), never listen to what the w*****s on SKY are talking about between races, etc. All my betting is based on statistical analysis of the form and the general consensus of the public as reflected in the TAB price - and I am doing quite nicely thank-you very much!

Despite what some others here seem to think there are a number of "truths" in Australian Racing (at least at the moment):

1. Longshot bias is alive and well and can be used to your advantage.
2. The TAB price IS a good indicator of a horses chances of winning.
3. Certain form factors ARE also good indicators of a horses chance of winning.
4. When 2 & 3 above agree you have a good bet, when they disagree you know to stay away.
5. Excel is NOT powerful enough for serious analysis work on large amounts of data - look at the punters with REAL databases (and the correct knowledge/training to use them) as opposed to Excel and you may find a bit higher success rate!

Of course this is not the only way - certainly there are some very successful punters who probably couldn't use a spreadsheet or database to save their life and wouldn't know a real statistic if it hit them in their face - getting down to the track every day to make contacts, gain inside knowledge, etc, etc, etc is also another approach that obviously works - to me it just seems like a lot of hard work. Each to their own I guess!

Bottom line is you need to find an approach that works for YOU but don't dismiss another persons approach just because you don't understand how it works.

<font size=-1>[ This Message was edited by: becareful on 2003-11-05 11:01 ]</font>

[Dr Pangloss]

crash

Thanks for your latest post, the spirit of which still leaves a lot to be desired, but nonetheless was far more appealing than the previous flippant attempts.

I am in no way advocating to drop everything in favour of developing one's own multinomial logit handicapping model - there are many ways to skin the punting cat. But unlike punting intuition or gut feelings a purely mathematical handicap model CAN BE MEASURED FOR PERFORMANCE and duplicated by all comers.

The Benter experience can not be simply dismissed on the basis of your myopic condition. More to the point, it conclusively demonstrates the value of 'Market price' input when executing the punt (remember the original question).

I liken the statistical snobbery remark to those who insist the Earth is flat - in punting terms I truly wish there were more of you.

[La Mer]

[quote]: On 2003-11-05 10:38, becareful wrote: Sorry Crash but it is definitely possible to make a profit in OZ using nothing but statistical analysis of form and price.
***************************************
La Mer ... I know of someone who has built their analysis and betting around the statistical teachings of Roger Biggs and they appear to be doing Ok, as they attend the track on a regular basis and gave their job away to concentrate on their betting.

[quote]On 2003-11-05 10:38, becareful wrote:
All my betting is based on statistical analysis of the form and the general consensus of the public as reflected in the TAB price - and I am doing quite nicely thank-you very much!
************************************
La Mer: Likewise, I developed my own ratings approach is soundly based on statistical analysis and it has and continues to serve me very well indeed

[quote]On 2003-11-05 10:38, becareful wrote:
1. Longshot bias is alive and well and can be used to your advantage.
2. The TAB price IS a good indicator of a horses chances of winning.
3. Certain form factors ARE also good indicators of a horses chance of winning.
4. When 2 & 3 above agree you have a good bet, when they disagree you know to stay away.
*****************************************
Le Mer: Longshot bias does exist and is a very useful tool in more ways than one, but I disagree with Becareful (in part) when he states that the TAB price is a good indicator of a horses chances of winning. This is not always the case as it was yesterday in the Melbourne Cup where a number of runners, such as Big Pat, Holy Orders, County Tyrone were all 'well unders' on the totes and did not in any way represent their chances of winning the Cup.

I happened to lay one of the three for the biggest amount I've every done so on Betfair and came out a big winner, as I laid it at a price that was approximately half the price it should have been according to my ratings. It is worth adding that the price I laid it at was still approximately twice the odds available on the tote.

[quote]: On 2003-11-05 10:38, becareful wrote:Of course this is not the only way - Each to their own I guess! Bottom line is you need to find an approach that works for YOU but don't dismiss another persons approach just because you don't understand how it works.
****************************************
La Mer: I agree totally with these comments.


<font size=-1>[ This Message was edited by: La Mer on 2003-11-05 11:26 ]</font>

[puntz]

STEBBO'S
Personally, I can't see the sense in reducing my bet just because other people don't like my selection. If I've done the form, and I make a selection that I think is the best horse in the race, why would I change my mind and my plan just because others don't agree with me?

REPLY:
so it's back to the original question, as qouted above.
This is the same dilema I face, and wantred to be sure on a "middle ground".
I think then,over time with a gathering of some stats.,how much to allow before the punt on a overlay is suitable to increase bet size.
If, let's say the variance is 50% difference, then selections with that variance on average, do they win according to one's own method of selection.

Crash
I only use the spreadsheet to calculate, accurately. But a rough guess might do. At the track though, usualy I am prepared with my selections and MY prices as a guide scribbled on the form guide.
So, if I have 3/1 on my form scribbles ( and they end up in the bin on the way out)
and bookie is 8/1, it's a bet for me.
If I have 20 units on it, or 200 units, regardles of bookies price, that's what I have decided before getting to the track!
What can be simpler?
If it's an extreme case, my 3/1 is 100/1 on the tote, well, it depends.
Perhaps quality or class of race.

[stebbo]

Hi All,

there's been some very interesting contributions to this thread, and some good food for thought. Following on from some of the ideas, I've done some analysis on a system that I have been developing.

I have come up with the following stats - Sorry Crash, but I couldn't help myself :smile:

For ExpSR (expected strike rate) I have taken the average strike rate for the range and used that to calculate expected strike rate. For "average" I calculated the mathematical mean (as opposed to the midpoint of the range)

<pre>
SP NRaces ExpSR ActSR
All 1049 10% 22%
[ 1.. 4) 239 34% 39%
[ 4.. 5) 128 22% 22%
[ 5.. 6) 138 19% 26%
[ 6.. 7) 69 16% 25%
[ 7..10) 170 12% 15%
[10..15) 148 8% 13%
[15..25) 93 5% 9%
25+ 64 2% 5%

<= $7 581 23% 30%
> $7 468 6% 12%
</pre>
.

It's clear that the market's opinion of my selection does affect the strike rate of my selection, but only after a certain point. For SP's up to $7, there is no effect on strike rate for this system and above $7 the strike rate is always above the expected strike rate. My conculsion therefore is that perhaps there's enough edge in this system to take advantage of the overs, and increase the bet.

Cheers,
Chris.

[puntz]

stebbo,
as catalyst to a market, if your 7 is a point X, then it can go either way.
The other way is to use it as measure, to hone selections. Yet stick to original bet size, but that is choice, or, get a tape measure, measure thikness of wallet, bet accordingly
:smile:

[crash]

Dr. Pengloss,

Thank you for your admirable restraint and such a speedy diagnosis of miosis, though I might seek secondary opinion [perhaps from Dr. Turf].

Read carefuly [and by all others here too] you would have noticed that I did not dismiss Benter [meaning stats.as that is now how it has been interpreted] as you conveniently suggest and I quote myself,'when given more importance than they deserve, stats...', but questioned the model and stats. generaly as to their overemphasis [beyond stats. that deal with facts of course] in impotance for Oz. conditions. Perhaps miosis is more widespread than you think.

I note you have elevated me to the dizzy heights of somewhere beyond 'flippant'. I feel wonderful but nowhere near what you must feel breathing the rareified air kept exclusively for 'Doctors' of the punt [do you bulk bill ?].

Cheers.

PS. And by the by everybody, I use a spread-sheet too, but obviously not to the extreames [my missed point in my past post] of science used by some.



<font size=-1>[ This Message was edited by: crash on 2003-11-05 17:42 ]</font>

[crash]

Stebbo,

You have quoted me correctly except for the quote that you base your arguement on "Stats carn't help with the punt" [I just never said it]. Points given for an outragious try though ! Pity about all your hard work.

As for the rest of you ratbags [I can only think so fast, even though I have made it to beyond 'flippant'], I will get around to a comment eventualy, just dealing with the easy replies first [Stebbo is new and a bit green still] as I'm sure you all well know.

Cheers.

<font size=-1>[ This Message was edited by: crash on 2003-11-05 17:40 ]</font>

[stebbo]

You have quoted me correctly except for the quote that you base your arguement on "Stats carn't help with the punt" [I just never said it].

Hi Crash,

apologies for the misconception. It was not my intent to misquote. The stats can't help with the punt was italicised purely for effect. The statement paraphrased what I thought was being said in your original, unedited post, and it appears that a few others read it the same way.

And there's no pity at all about my "hard" work... For starters the hardest part about it was typing it into this crappy editor, and it was quite an informative exercise none the same.

:smile:

Cheers,
Chris.

<font size=-1>[ This Message was edited by: stebbo on 2003-11-05 18:04 ]</font>

[kenchar]

Hi all,
Interesting thread.

For any that are interested I back from odds on to infinity,within reason the bigger the odds the better.If my gut feeling tells me the nag can win back it regardless of odds, and what ever I bet on the first race I maintain that bet.

Except in rare cases as per the other week when I changed my bet to have $50 E/W on a 40/1 shot, but then spent a lot more on tri & first4.(I must admit the 40/1 was a bit out of my usual range)

My excel is a notebook and pen,when the plus side on the day is enough I go home. :smile:
Sometimes it doesn't go to plus :sad:

When I get home I get my other excel(bigger notebook and different pen) and keep my records.

I have just subscribed to a new service which is showing a 50%+ strike rate and 30%+ pot. They are all shorties,but 7 days a week. I am going to run it in conjuction with Propun specials as a seperate betting method with a seperate bank, will update in a couple of months to see if it is worthwile.

My own thoughts, if it's worth backing and you think it can win regardless of odds put the money on,if your'e not sure leave the race.

I will quite often let 10 races go with no bet.

Many years ago I drove from Sydney to Newcastle AND NEVER HAD A BET FOR THE DAY.

Cheers

[crash]

Becareful,

Of course it may be possible to make money just from stats. analysis alone, but greater profit exists with some intuition added [quite a bit actualy]. You should try it sometime as it is even quiet enjoyabe.

Puntz and Kenchar, I basicaly use spreadsheets as you do as they have their place as just another ingedient in the cake mix, but never THE ingredient as often suggested in this forum.

Stebbo, Oh yeah the old slip in some 'italicised' quotation mixed with genuine quote trick. The old 'Strawman' gambit [build a Strawman and then proceed to burn it down, 'proving' your own barrow], gets flogged to death in this forum and easy to spot.
Your stats. model by the way at least shows that what I was saying about the bigger the overlay, the bigger the bet was spot on. Didn't need a stats. model though to work it out [long ago].

Being serious for a moment [what was this forum about ?], I stand by what I [really] ment about stats. and that is they are just another tool in the toolbag, but unlike most other tools, it can get a bit too greasy.

Cheers.





<font size=-1>[ This Message was edited by: crash on 2003-11-05 23:25 ]</font>

[Chrome Prince]

On 2003-11-05 22:46, crash wrote:
Of course it may be possible to make money just from stats. analysis alone, but greater profit exists with some intuition added [quite a bit actualy].


Afraid I disagree with this view, but I guess we each have our own way of making a profit.
If I want to win, I use stats.
If I want to gamble for fun, I use intuition, never the twain shall meet.

Where's Hay Chee? :wink:

[crash]

Another Stats. junkie has turned up, hi Chrome. Wondering when you would show. You have misunderstood the way 'intuition' has been used. It is the method of sifting the stats. and all other info. into a meaningful bet [or not].

Cheers.

<font size=-1>[ This Message was edited by: crash on 2003-11-05 23:35 ]</font>

[puntz]

crash
it all comes down to simply this;

least amount of effort, for maximum possible profit, in the shortest practical time, and with the least amount of disharmony.




<font size=-1>[ This Message was edited by: puntz on 2003-11-06 00:00 ]</font>

[Chrome Prince]

On 2003-11-05 23:31, crash wrote:
Another Stats. junkie has turned up, hi Chrome. Wondering when you would show. You have misunderstood the way 'intuition' has been used. It is the method of sifting the stats. and all other info. into a meaningful bet [or not].


Apologies crash,

Seems I may have misinterpreted your previous post.

Now back to my database for my Oaks Day fix :eek:

:grin:

[crash]

No disharmony here Puntz. Just a zestful thread bubbling along with my usual dollop of sarcastic [as Umrum once labeled it] humour.

Chrome,

I obviously should have chosen another word in my posts here to describe the process of turning info into bets. 'Intuition' was not a good choice in hindsight as everbody here has interpreted the word literaly [fair enough] : 'the direct perception of truths independent of any reasoning process'. The meaning has easily perceived conitations [that it dosn't deserve] that are incorrect but the general perception inplied was understandable.

Cogitative or even cognisance [with some intuitive wisdom] would have been more descriptive of the meaning ment to be conveyed.

Cheers.



<font size=-1>[ This Message was edited by: crash on 2003-11-08 02:15 ]</font>