Win and place pool ratio

[cadman]

Happy New Year all.

I keep toying with the idea that the information displayed on the tote has something to say! The only problem is that it’s encrypted to the average Joe Blow. I see the work of a genius is very much required. You would think that with the advent of the computer age one would be able perform statistical analysis on the figures or prices that are continuously updated on the tote and from this untangle the puzzle.

I wonder if others share this idea? For ages I’ve wondered if there is any ratio or relationship between the Win and Place pools? The result from various web searches pops up the Dr Z system (refer http://www.thetrac.com/books/drz.html). Has anyone tried or had any experience with this kind of thinking?

Unfortunately my knowledge on how the TAB pools operate is extremely minimal. Can anyone offer some pointers or direction to for a better understanding of the mathematics behind the Win and Place pools? A dummies guide to how the TAB pools operate would be fantastic.

Kind regards
Cadman

[Every Topic]

hi Cadman,
I will leave the lesson on the TAB to the other more knowledge members of the group - as for the Dr Z method and related issues...
1. I believe place betting is far more logical than win betting, and more secure.
2. I believe the tote prices are the most accurate guide to form and therefore should be considered as a serious guide. I believe this because the tote price is a reflection of every piece of knowledge about the race. Its a reflection of the owners, trainers and race track workers knowledge, its a reflection of the experts with their massive computer programmes and a reflection of the rest of us donkeys :smile:
The tote price is telling a lot.
3. I believe that in using the tote price, one should use the win price to ascertain who are the top contenders and then use this to reflect upon the place price.
4. As for Dr Z's method, what I can understand of it... I'd probably go for a variation of it, as the author has done.

its an interesting subject, and controversial

see ya
Every Topic

[becareful]

Hello from smouldering Canberra. Here is an attempt at a "Dummies guide to TAB pools" - hope this helps.

Lets start with the simplest - the win pool. An example is probably the easiest way to demonstrate so I am going to assume an 8 horse race. Each horse has the following amount bet on it:
1. $100
2. $70
3. $60
4. $55
5. $50
6. $30
7. $25
8. $10
The "pool size" as reported by the TAB is simply the sum of the amounts bet on each horse - in this case the pool size is $400 ($100 + $70 + ... etc). To work out the dividend for each horse the first thing the TAB does is take out their cut - for the win pool this is around 14.5% (this is for Supertab - I think the others are around the same). The TAB therefore takes out $58 (14.5% of $400) and the remaining $342 is available to be paid to the lucky winners. Each dividend can then be worked out by dividing that $342 by the amount bet on each horse and then rounding down to the nearest 10c. So the dividend for horse 1 would be $342/$100 = $3.42 = $3.40 (after rounding). So the dividends for our 8 runners would be:
1. $3.40
2. $4.80
3. $5.70
4. $6.20
5. $6.80
6. $11.40
7. $13.60
8. $34.20

One thing to note is that because the TAB rounds down the effective take-out percentage is more than the 14.5% - in the case of the win pool it averages about 15% - with place pools it can be a little higher (this is because the lower the dividend the greater the rounding effect is).

Knowing the above information we can also work backwards to work out how much of the win pool is invested on each horse if we know the dividends. To do this we divide 1 by the dividend and then multiply by (1-TABTake) - in this case you need to use the adjusted TAB take of 15% so 1-0.15 = 0.85

If we look at horse 1 in the above example we have:
(1/3.4) * (1-0.15) = 0.294 * 0.85 = 0.25
So horse 1 has 25% of the pool invested on it (25% of $400 = $100)

For horse 2 we have:
(1/4.8) * (1-0.15) = .208 * 0.85 = .177
So horse 2 has 17.7% of the pool (17.7% of $400 = $70.80)

Due to the rounding you will not get exact investment figures when you do this reverse calculation but it is close enough.


The place pool works basically the same as the win pool except there are, of course, 3 dividends instead of 1 (assuming 8 or more runners).

Lets assume that the same amount was bet on each runner in the place pool as the above win example (so again $400 total with $100 on runner 1, etc). Again the TAB takes their cut first (same percentage as win pool) so we have $342 to be split between the 3 place getters. To work out the dividends we first divide the pool by 3 to get the amount to be paid to each winner - $114.

Each dividend can then be worked out by dividing that $114 by the amount bet on each horse and then rounding down to the nearest 10c. So the dividend for horse 1 would be $114/$100 = $1.14 = $1.10 (after rounding). So the dividends for our 8 runners would be:
1. $1.10
2. $1.60
3. $1.90
4. $2.00
5. $2.20
6. $3.80
7. $4.50
8. $11.40

Again we can also work backward to find out how much is bet on each horse based on the dividend. We use the same formula as the win pool except we divide the final answer by 3 (or 2 if there are only 2 places being paid).

If we look at horse 1 in the above example we have:
(1/1.1)* (1-0.15) / 3 = 0.909 * 0.85 / 3 = 0.258
So horse 1 has 25.8% of the pool invested on it (25.8% of $400 = $103)
Again the rounding effect gives us the error in the pool amount.

I hope this is helpful - I don't have time at the moment to go any further with how you might use this info but if anyone has any questions I will do my best to answer them when I get a chance.



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

<font size=-1>[ This Message was edited by: becareful on 2003-01-20 15:36 ]</font>

[Every Topic]

BC, Im sure all of us are thinking of you in Canberra and surrounding areas - it was shocking watching it on TV let alone be there!

I have a question about the TAB...

When the race jumps the NSW TAB odds page shows a (closed - final pool) message. I have often noted my runners dividend and the pool total - I am talking about the place pool here.
Then a minute or so later, still during the race the numbers change again with a few thousand dollars added to the place pool and the odds changing.

Why is this so?
If they give a closed final pool message why then isnt that the pool "final". How can more money be entering the pool after the race has jumped???

am I smelling a rat or do I just need a shower :smile:

all the best
Every Topic


<font size=-1>[ This Message was edited by: Every Topic on 2003-01-20 16:29 ]</font>

[GeneralGym]

I would beg to differ here EVERY TOPIC
"Quote"
2. I believe the tote prices are the most accurate guide to form and therefore should be considered as a serious guide. I believe this because the tote price is a reflection of every piece of knowledge about the race. Its a reflection of the owners, trainers and race track workers knowledge, its a reflection of the experts with their massive computer programmes and a reflection of the rest of us donkeys
"Unquote"
I am of the opinion that a leading bookmakers market is a fairer reflection of a race than the tote.
The Bookie lives and dies by his market where the tote is an accumulation of investments that sets its own market.
A bookie may "take on" a horse where the tote never does.

CADMAM The place dividend should be 1/3 to 1/4 of the Win dividend. Anything outside this start smelling a rat or a huge Allup bet (or bookies lay off) that is lopsiding either the win or place pool dependant on the pool sizes.(Another reason the tote is not the best guide)
The only true Tote market on any given day is the first race of any meeting that has no Allup beeting on it.

[becareful]

Every Topic,

The only thing I can smell at the moment is smoke so cant help you on the rat front :smile:

Seriously though I suspect that the issue with NSW is probably just to do with the delay in updating the web site details. They probably put the Closed message up as soon as the race jumps but dont update the divs/pool details until the final calculations are complete. The final pool is generally not finalised until just after the jump as the TAB outlets do have the ability to cancel bets placed just before the jump - eg. punter puts ticket through automatic reader just before jump and the machine reads the $1 bet as $100 - the operator can override the bet amount or cancel the bet for a certain time after the bet has been logged even if race has jumped so final pool is not available until all transactions have either been confirmed or cancelled. These "in process" transactions probably explain the change in prices. For Supertab the problem is even worse as they have 4 different TABs processing bets that must then be added together to get the final totals.


GeneralGym,
Not sure I agree with your point on bookmakers prices - the TAB price is amazingly accurate in the long term as an indicator of winning chances. There is no doubt that in some races bookmakers prices will be more accurate but in others the TAB will be better. I had a quick look at some figures for last year that may be of interest on this subject. What I did was looked at all the horses that started at a particular price, calculated the expected strike rate and compared that to the actual strike rate. For example in a 12 month period there were 1418 horses that had a TAB price of $1.90, $2.00 or $2.10 (I grouped three prices together to smooth out any irregularities). With an average price of $2.00 you would expect a strike rate of around 42.5% (0.5 * 0.85 tab margin) - the actual strike rate was 45.3

For the prices from $2 to $10 the results were:

PRICE ExpSR ActSR
$2.00 42.5% 45.3%
$3.00 28.3% 28.4%
$4.00 21.3% 20.0%
$5.00 17.0% 16.7%
$6.00 14.2% 12.9%
$7.00 12.2% 11.2%
$8.00 10.2% 11.3%
$9.00 9.4% 9.5%
$10.00 8.5% 8.1%
$20.00 4.3% 3.7%

As you can see the actual strike rate is remarkably close to the expected rate up to $10.00

With regards to your point on allup betting it is generally fairly irrelevant on most races. The proportion of money in any pool from allups is almost always insignificant as very few people place allup bets and the ones that are placed are generally on favourites which means they have little impact on the tote prices. The other factor that minimises the impact of allups is that they generally go into the pool as soon as the previous race is declared (so around 30 minutes before the jump time) so there is plenty of time for the prices to equalise to correct prices before jump (eg. if big allup goes in on horse 1 and drops the tote price to $2 when it should be $4 then punters will avoid that runner or bet on the other "value" runners until the prices equalise).

[becareful]

Getting back to the original point as to is there any way to use this information as per the Dr Z. approach?

I actually did some research on this some time ago and came to the conclusion that the opportunites in Australia were quite limited. Firstly the system developed by Dr Z uses the US tote system which pays on Win, Place and Show where we only have Win and Place (with Place equivalent to the US Show). From the analysis I did it seems that our place pools are more in line with the win pool than seems to be implied by the Dr Z stuff I read as there seemed to be relatively few opportunites where the place dividend was enough of an "over" to overcome the TAB take. The other big problem (and this was the real killer as far as I was concerned) was that the place dividends can change significantly from just before the jump when you would have to place the bet to when the final div is declared. If you are betting on say a 10% calculated overlay but then the price drops by 15% after the jump then you have just gone from an overlay to an underlay.

Of course there is no doubt that there are opportunities where the place divs are definitely out of whack - eg. there was a trots race last year where the winner paid $1.30 for win and $1.50 place - the question is can you find them often enough to avoid dying of boredom while you wait?

[Fryingpan]

Hello everyone

A question to Becareful. Could you tell me a bit more about how these percentages work. They are very interesting. I understand the coincidence between the expected and the actual. It's just the percentages add up to over 100 if you sum them together. So just wondering if you could explain a little bit about the proportions. Thanks.

"For the prices from $2 to $10 the results were:

PRICE ExpSR ActSR
$2.00 42.5% 45.3%
$3.00 28.3% 28.4%
$4.00 21.3% 20.0%
$5.00 17.0% 16.7%
$6.00 14.2% 12.9%
$7.00 12.2% 11.2%
$8.00 10.2% 11.3%
$9.00 9.4% 9.5%
$10.00 8.5% 8.1%
$20.00 4.3% 3.7% "

[becareful]

Fryingpan,

The percentages are the strike rate for all horses that started at the specified price over a 1 year period. For example with the $2.00 starters there were 1418 horses that were priced at $1.90, $2.00 or $2.10 during the year. Of these 643 won their race for a strike rate of 45.3%. The expected strike rate (removing the TAB margin from the price) was 42.5% so these horse are slightly more successful than you would expect from the price (ie. they are slight "overs" on average) but unfortunately not enough to overcome the TAB cut. The reason the percentages add up to more than 100 is that I was just looking at the strike rate for each price point - obviously you generally dont have a race with a $2, $3, $4 and $5 runner all in the same race. If you took a race with say a $2 favourite, a couple of $5-$6 runners, couple of $9-$10 runners and a few $20+ runners and added up the strike rates for each of those prices then that should add up to 100%

Hope this clarifies it for you.

[woof43]

Now imagine if you applied those same expected s/rates for every horse by any trainer, you would have a far more significant number to work with, then do that by jockey, then do that by barrier, then do that by distance and that is only the start...that gives the Null hypo..those figures can be very powerful if applied correctly.

[GeneralGym]

Becareful

We may have to agree to disagree on this one.
I am a firm beleiver that the Bookie sets the market while the tote is market driven. It is all very well to come up with the figures that you have but at the end of the day horses win on ability not on how much money is on them although it is always a good feeling to see a plunge on a horse after you have yours on.
I do my form, narrow a field down to 3/4 or less genuine chances and place the money on one or two of them. I try to clear a certain profit margin but I only take odds that I consider are "value to me"
I don't wait till the market is set and put my money on $2 favs because statistics show that 45% of them will win.
As a matter of interest from your figures what percentage of winners are priced at $10 or less?

[Fryingpan]

Thanks Becareful,

Much appreciated. I was doing rubic cubes in the head previously.

[becareful]

Fryingpan - you're welcome :smile:

GeneralGym,
I am not saying that the market price is some magic predictor of each race - obviously sometimes the bookies and the punting public get the prices wrong and it is only by identifying these "mistakes" that you can make a long-term profit (obviously by betting on the runners where the price is higher than it should be). As to whether the general opinion of thousands of punters is more accurate than the opinion of a much smaller number of bookies - I am not sure but I think the TAB prices are slightly more accurate.

With regards to your question on the percentage of winners priced less than $10 my figures give the following:

48% of winners are $4.90 or less
78% of winners are $9.90 or less
89% of winners are $14.90 or less
93% of winners are $19.90 or less
99% of winners are $49.90 or less

[GeneralGym]

Becareful

I'll beg to differ again with you on the word "accurate" but will concede that the tote prices tend to give less chances at shorter odds where the Bookie tends to keep more runners at safer odds especially earlier on in the betting.

thanks for you figures which confirm my line of thinking over the last few years that any odds over $15 are risky odds.
I always think twice when my selections come up at odds over $15.
I am dubious about under $3 but that is a personal thing as short odds mean a higher investment outlay to get a decent return.

[GeneralGym]

I haven't read the Don Scott book but would certainly like to. I understand that they are hard to get and expensive now.
I have read the following site online and was quite taken by it.

http://www.horseracingaustralia.hl.com.au/ultimate/index.html