The home of intelligent horse racing discussion

# Calculating Place Odds from Win Odds

Home Forums Tipping and Research Trends, Research And Notebooks Calculating Place Odds from Win Odds

Viewing 15 posts - 1 through 15 (of 51 total)
• Author
Posts
• #1881
alan1
Member
• Total Posts 167

I know this has been raised on here before, and from what I remember without a satisfactory conclusion, but has anybody got any ideas on how to calculate place odds based on win odds.

I tend to just divide the win odds by the number of places available e.g. A horse at 9/1 in an 8 runner hcap would have place odds of 3/1?

But this is a bit crude and doesnÃƒÂ¢Ã¢â€šÂ¬Ã¢â€žÂ¢t take into account the shape of the book such as when a very shortpriced fav is distorting the odds.

So has anybody got any better suggestions?<br>

#63692
Aragorn
Member
• Total Posts 2208

Alan,

Do you mean forecasting your own or working out what they are? Because if its the latter you would just multiply it by the fraction that the bookies are offering for places i.e. 1/5th or 1/4 on 16 runner handicaps..

There are other people on here much better informed on the former.

#63693
alan1
Member
• Total Posts 167

I’m trying to work out the REAL odds. So in a 100% book what are the real odds of a horse being placed.

I’m not trying to calculate what a bookmaker would pay out.

#63694
Anonymous
Inactive
• Total Posts 17718

Quote: from alan1 on 2:38 pm on June 4, 2007[br]I’m trying to work out the REAL odds. So in a 100% book what are the real odds of a horse being placed.

I’m not trying to calculate what a bookmaker would pay out.

Alan<br>I’m no statistician, but surely for 3 places you would multiply the 2nd part of the sp by 3, for 4 places by 4 etc., so that 4/1 against a win equates to 4/3 for a place, and so on?<br> Ãƒâ€šÃ‚Â  Ãƒâ€šÃ‚Â  Ãƒâ€šÃ‚Â  Ãƒâ€šÃ‚Â  Ãƒâ€šÃ‚Â  Ãƒâ€šÃ‚Â  Ãƒâ€šÃ‚Â

#63695
dave jay
Member
• Total Posts 3386

You can’t calculate real place odds from real win odds.

#63696
Anonymous
Inactive
• Total Posts 17718

Quote: from dave jay on 4:54 pm on June 4, 2007[br]You can’t calculate real place odds from real win odds.

Errmm,………Alan did state ‘a 100% book’.<br>8 rnrs @ say: 7/1 the field = 7/3 the place, unless I’m missing something?<br>

#63697
Drone
Participant
• Total Posts 5112

Try googling "Harville Formula" and the -s -e plurals, and maybe adding "Benter" to narrow it down<br>

(Edited by Drone at 5:12 pm on June 4, 2007)

#63698
dave jay
Member
• Total Posts 3386

You could use the Harville formula or simply multiply your win odds by the number of places, but it doesn’t work because a horse that hasn’t got a chance of winning might have a very good chance of placing, and visa-versa, as I suspect you are finding out Alan. You are better off trying to work out what you think the place odds actually should be rather than trying to work them out from the win odds, IMO.

#63699
non vintage
Member
• Total Posts 1268

Very difficult to do, and almost impossible to do accurately.

Basically, the shape of the market is all important.

A 7/1 shot in one 8-runner race where all the runners have an equal chance of winning doesn’t have the same chance of placing as a 7/1 shot in another 8-runner race where there is an odds-on favourite.

In essence, you have to build a model which uses a 100% book to calculate the chances of each horse coming second to each other horse, and then onwards if there are 3 or 4 places.

I have just thrown together something for 5-runner races which is of limited use but does illustrate the point. Realistically, this could be expanded to 6-runners without becoming unwieldly, but unless you are happy to bung ‘the field’ in as one selection, probably won’t be of much practical use…

#63700
Prufrock
Participant
• Total Posts 2081

I agree with dave jay.

You can use a mathematical model to provide you with a starting point but the individual horses and even the races they are running in count for a great deal.

#63701
Anonymous
Inactive
• Total Posts 17718

<br>Surely, if you use a mathematical model based on win odds a 1/5 shot is 1/15 to place, and anything that can’t be proven mathematically becomes subjective, and therefore as open to individual interpretation as any theoretical book, win or place?

#63702
Smithy
Member
• Total Posts 720

Depending on the shape of an 8-runner race, a 5/1 shot in the win market can range from 1.3 to 2.5 in the place market (if not greater variance).

How does the mathematical model based on win odds help here?

#63703
Prufrock
Participant
• Total Posts 2081

The mathematical model is, I believe, Harville’s, as mentioned before and as used (but not credited) in The Mathematics of Betting by Phil Bull. The mathematically "true" odds are arrived at by a much more complex process than simple division.

However, that can only ever be a starting point. There are limiting factors on how close to "certainty" a horse’s place odds can get. Ãƒâ€šÃ‚Â

Would you back an even-money shot in a 15-runner novice chase at 1/5 to be placed, given that if it does not win it could well have fallen, unseated, been brought down or carried out?

Then there are the horses themselves. I’m sure we can all quote examples of horses who are good at being placed but seldom win; of horses (or should that be trainers?!) who either win or are not sighted; and of horses whose running style lends them to all or nothing or the opposite.

(Edited by Prufrock at 7:54 pm on June 4, 2007)

#63704
Maurice
Participant
• Total Posts 355

I haven’t come across this debate before so I’m thinking as I go, therfore my thinking will probably be flawed…

A 7/1 chance means the horse as one chance in 8. To be placed therefore means 3 chances in 8, or 5/3. It’s a huge difference and probably explains why place-only odds against the first five or six in the market often appear very skinny.

If I can get 1/4 or 1/5 the win odds, I reckon I’m doing very well. If you fancy an 8/1 shot EW at 1/4 the odds, your Ãƒâ€šÃ‚Â£1 EW (outlay Ãƒâ€šÃ‚Â£2) nets you Ãƒâ€šÃ‚Â£12 (ie 5/1 overall). If it only gets placed, your returns are Ãƒâ€šÃ‚Â£3 (ie 1/2 overall). If you are unsure about the horse winning, I reckon you’re a lot better trying to get 2/1 place-only for a Ãƒâ€šÃ‚Â£6 return, but you’ll probably struggle to get it.

#63705
Prufrock
Participant
• Total Posts 2081

Just reading round the subject on t’internet and the fact that the purely mathematical principles of Harville’s Probabilities do not accurately reflect "real life" seems to have been addressed by something called "Discounted Harville’s".

Anyone who has played place markets at all frequently will know just what a complex subject this is to pin down.

Viewing 15 posts - 1 through 15 (of 51 total)
• You must be logged in to reply to this topic.