Help needed
 

I was looking at my pricing structure for ratings and can't work out why i have it set up a particular way, more than like my old age has started to kick in.

I think it was a member here that put me in to this method, so am hoping that member may remember better than me, if only i could remember who, haha.

The normal way to price up a market is to add up all the ratings to get a total then divide each rating in to the total to get a percentage then convert that to a price.

But what i have is 2 groups (Group A) only uses the top 5 selections and uses the above method (Group B) uses all the runners then apply's the prices to the other runners outside the top 5

What i think this does is tighten up the prices of the top 5 because these would be the most accurate rated runners seeing as 70 to 80 percent of winners come from this group.

Dont know if this will help but I found the 'confidence' element useful & you may be able to apply one figure to group A & another to B

http://betting.betfair.com/horse-ra...222-040810.html

I think you got it from here Shaun if I recall correctly.

Yes mate that was the pricing system, but the extra element of seperating the 2 groups was different.

I guess in the end the reason for doing it the way i have i explained to myself in the first post.


I just hate not knowing where i got the info i use.

I have no idea exactly how your Ratings are devised , even though you freely offer your methods to all members.

I myself use the tally up all the ratings and divide that figure into the individual Ratings.
But the key is that the ratings must be spread with the appropriate gap between them.

Like - an animal that has twice the rating of another animal . MUST , in fact , have twice the actual chance of saluting as the other animal.

And if an animal has a rating of 100 , and another animal has a rating of 10 , then the better animal must have 10 times the chance of winning as the lesser animal - Not just because your formula says so , but it actually has.

You can sort of work out if this is in fact happening by Rating a bunch of animals where the Form is well revealed.
Often a fairly accurate Market can be mentally figured in this case just with your own basic knowledge.

Or fool about with Races and eliminate most of the runners bar the top 3 or 4 chances , and see if the Market spat out , in fact looks legitimate.

Was gonna say something else , but being only fresh back from being banned , had second thoughts :(

Talking about bans - is Barny due back soon into the fold?

There is a free Pricing to Rating available on Ozracingtools.

It can create a price from just about anything.
e.g. place percentages. win percentages, UniTAB ratings, Sportsman ratings.
Av. prize money, Neural rating, Don Scott ratings Worksheet.

One also adjust the market percentage to say 70-80%

Saves heaps of time.

Do you have to input the data 1 by 1 or can you paste like 10 ratings or the whole field in one hit?

That sounded ideal.

So went and had a look.
I went to ozracetools , is that the place ? - they had what sounded like what you say is available.

I never saw any help attached with the program.
How did you advise the Program whether you were inputting Unitab Ratings or Don Scott Ratings or Neural Ratings?
Also , how did you advise the Program to adjust the Program to Frame at say 80%?

I think the Ozracingtools Racing Ratings to Price spreadsheet has its limitations.

I entered a fictious 10 horse field using Unitab ratings (80,82,87,88,89,90,97,98,99,100) and the odds spread was $9.50 for the top 4 rated, $10.50 for the next 4 rated and $11.50 for the last 2 rated horses. This is obviously meaningless if you are trying to set a price for each horse using these ratings. The same effect is shown for Zipform Rating figures e.g 52 through to 62.

Not sure what set of rating figures would work.

I had similar useless meaningless results as Try Try Again.

Hopefully Bhagwan can demonstrate where our lack of knowledge lies.

This is often the case where one member seems to do just fine and dandy , yet cannot pass on anything useful to share with other members because it is not so much the tools that members use , but how they are used.

I've tried on many occasions to Profit on other members ideas , but failed miserably pretty much every time.
And whenever I try to pass on useful knowledge to other members , I end up either failing miserably and embarrasingly , or getting banned for a period of time.

Tools are just that.
Used incorrectly , a botched up result will occur every time.

The link I provided earlier in this thread has been the most useful pricing model I have come across & all that is required is little bit of setting up in excel & then tinker away until it suits your particular circumstances.

I had another look at that link you posted NoRisk , and I found it was exactly how I compile My Prices from My Ratings.

The Key is to insert the correct figure in Cell A2 , the Power factor.
This Number is like the writer suggested , determined experimentally to suit particular Ratings Methods.

In his example , he uses 0.867 because his ratings are based on Weight differences.
Had he used ratings based on Length Differences , then the Power Factor would be 0.656

DON SCOTT - YOU LEGEND!!!!
R.I.P.

Agree Norisk and moeee. Not a ratings bloke but that does look like a good method. That ozeracetools one is compete pants.

The key is converting Kg distance from the winner into a decimal.

Winner on ratings = 1.0
.5Kg from winner = .90
1.0 Kg from winner = .80
1.5 = .70
2.0 = .60
2.5 = .50
3.0 = .40
3.5 = .33
4.0 = .25
4.5 = .20

So...if you have 3 horses within 5kg of the winner and they are rated 60, 58 and 56 you add up 1.0+.60+.25 = 1.85. The top horse is thus 1.0/1.85 or 54%, the next horse is .60/1.85 or 32% and the last horse .4/1.85 or 21%

Does this make sense?

Looks ok to me Vortech

Thats all fine Vortech.

But then you may have someone who prepared their Ratings via a non Form Based Method.
Like the Neurals for example.
Points are allocated for various reasons , and I believe that Shauns Method of Ratings could be one of those type of Programs , but is sort of a educated guess as I don;t play horses and didn't disect Shauns freely available and generously donated hard work.

And the thing about the Link that NoRisk provided , allows for all sorts of Ratings Methods , although I'm not quite sure about Sportz Greyhound Ratings.
They appear similar to the old tipsters poll rather than Ratings as accepted in the usual sense of the word.
Fortunately Sportz has discovered his own method of converting his numbers into Prices.
And he must be doing just fine and dandy because he is now in his 3rd year or more.