Nice one boys?
 

Pinjarra 06 last night, worst 3 price ranked get up 1st, 2nd & 3rd. Where's that leave the 'form students' now, or for that matter, any 'student of the punt'? Oh, I forgot, this is the 'wild west' after all.

What percent formula do they use to obtain 27% ?

It's on the near top left, where it shows
1.75 to 1.55 = 27%
----------------------------------------------
According to percent formula A
= ((1.55 - 1.75) / |1.75|) * 100

= 11.4286% decrease

According to percent formula B
= ( | 1.75 - 1.55 | / ((1.75 + 1.55)/2) ) * 100
= 12.1212% difference

Source reference
http://www.calculatorsoup.com/calcu...-calculator.php

--------------------------------------------

It's harness racing after all. :) Does anyone know where to find the replay?

Not there yet but should show up this page soon.

http://www.harness.org.au/video-replays.cfm

Correct me if I'm wrong any of you stats genuise's out there, but the odds of getting the trifecta (using Betfair back price) was 100 x 70 x 46 = 321,999:1 The trifecta paid about $5,000 from memory.

I collect Betfair data each day and have over 6 years worth. I have tables based on 1st favourites price, number of starters and price rank. I know the historical max back price winner for that period, of any combination. How often is a new record broken for one of those combinations? .....About every 2nd day!

That's why basing a system on historical data is flawed to an extent and you have to ensure that you have many more wins, than flaws, in your system argument.

Thanks, it's there now. The favourite was way behind the rest of the field, but apart from that I didn't see anything unusual. Not that I would know anyway.

Assuming that the thing I read on the internet regarding the Discounted Harville method was correct, and that I didn't stuff up my implementation for 12 runners from the example that was given with 7 runners, a "fair" price for the winning trifecta (given the win odds for the field - and using the presumption that the odds are a somewhat accurate representation of the true probabilities) would be $72,844.14.

As a general rule, I suppose you should avoid betting all combinations that are of greater value than the pool size. (Though that's quite a lax rule as it's likely you would share the pool with other winning bettors).


...just for interest's sake, the fair trifecta price for the three shortest runners would have been $34.37.

You ought to know Walkermac (Excel 2014 Cup winner) :)

I was using the Don Scott method lol. Harville must've come after him? No matter what the odds, a freaky outcome, hey?

Apparently Harville published 5 years earlier than Scott ;) I understand he was looking for opportunities in the place market, where a runner's win odds indicated that its place odds should be shorter than what was on offer. (Aside: Harville's trifecta odds for your race were $354,649.49 - which is so far different from the other result that it makes me suspect that either one - or both! - of my calculations is wrong)

*but* The "Discounted" part of it came about because his results didn't fit the actual data well. It was noted that horses were coming 2nd only 81% as often as their win odds indicated they should (and only 65% of the time for 3rd). I haven't seen it explained very well why this is the case, but the next mob of statistician/mathematicians took these "discounts" into account for their formulation

The figures seem dependent on the track though (or rather, I suppose, the pool of bettors/bookmakers working a track/s). The figures for Japan (vs the Hong Kong ones stated previous) were 88%/80%. Was it 'cause the fave/longshot bias wasn't demonstrated in the Hong Kong data, was it due to different styles of racing between the two countries? It wasn't really explained.

A final tidbit: odds of exotics better estimated the finishing order. It was surmised that that was where the smart money was. Not necessarily that more canny gamblers were betting there, but that it wasn't such an obvious source of information as the win odds are to an ordinary punter (who would presumably overbet the horses already highlighted by the market as the likelier chances). Plus it's more difficult information to get.


Went off tangent a bit, but I happened to be reading about all that when you presented your odds estimate.

The VIC pools (and dividends) for the Exacta (especially) were markedly higher for that particular race and compared to the other TABs. I guess if you can manipulate the price on one TAB you can choose those TAB's odds at any number of online operators, load up and still keep the TAB's odds long. Seems the right race for it: 2 runners at low odds, pretty much the remainder of the field 20s and up. Would anyone else bother betting the longshots? If you were certain those 2 at low odds would finish out of the placings you'd make a killing.