General Betting
There is currently 1 person viewing this thread.
12:57 at The Moors, the 8 horse was forecast at 70.0
As an example, suppose 100 simulations were run of a race happening later today. Of those 100 simulations, the horse named Apple Tree won 20 times and the horse named Bay Wold won 8 times. Apple Tree's forecast probability is 20/100 = 20%. Therefore Apple Tree's forecast price is 1/20% = 5.0. Similarly, Bay Wolf's forecast probability is 8/100 = 8% and its forecast price is 1/8% = 12.5. If a horse does not win any of the simulated races, a default price of 110 will be set.
How can a horse be 70s?
As an example, suppose 100 simulations were run of a race happening later today. Of those 100 simulations, the horse named Apple Tree won 20 times and the horse named Bay Wold won 8 times. Apple Tree's forecast probability is 20/100 = 20%. Therefore Apple Tree's forecast price is 1/20% = 5.0. Similarly, Bay Wolf's forecast probability is 8/100 = 8% and its forecast price is 1/8% = 12.5. If a horse does not win any of the simulated races, a default price of 110 will be set.
How can a horse be 70s?
sort by:
Show
per page
Replies: 25
|
By:
Apparently, in one of the simulations, this horse came second, but the horse in front was found to have virtually taken steroids. The virtual trainer contested these allegations in front of the VBHA, and the case is on-going, thereby confusing the forecast pricer, who doesn't know whether to award a win or not.
|
|
By:
Cheers, that works.
|
|
By:
are there rules for how this works
|
|
By:
key part is where is says "as an example", in fact the betfair simulator only runs about 70 simulations
|
|
By:
I guess they have to run the simulations until each horse in the race has won at least once, otherwise the forecast price for a horse which didn't win in the first 100 simulations would be infinity.
|
|
By:
I wish they would introduce this for real horse racing - running the same race over and over so we can all get an accurate forecast for "true" prices.
|
|
By:
They do betfairy. It's called all weather.
|
|
By:
£23 matched on the 16:39 at Piccadilly Park, looks like this has really taken off
|
|
By:
It won't be long, before our good friend, zipper, and his million pound boys get this sorted!
Just need AK to put up some naps and off we jolly well go! |
|
By:
How are forecast prices for the races derived? Forecast prices are derived from the observed results of a limited number of simulations. The creator of Winners Circle runs 100 simulations of each race and records how many times each horse wins a simulated race. This gives a forecast probability, which is then displayed as forecast prices. These are not true odds. As an example, suppose 100 simulations were run of a race happening later today. Of those 100 simulations, the horse named Apple Tree won 20 times and the horse named Bay Wold won 8 times. Apple Tree's forecast probability is 20/100 = 20%. Therefore Apple Tree's forecast price is 1/20% = 5.0. Similarly, Bay Wolf's forecast probability is 8/100 = 8% and its forecast price is 1/8% = 12.5. If a horse does not win any of the simulated races, a default price of 110 will be set. No more simulations than those needed to derive the forecast prices are run, so true odds are never derived. Any changes to the number of simulations being used to derive forecast prices will be announced on the Announcements page. Are there dead heats? No. Dead heats are not possible in Winners Circle. Although the finish of races may be extremely close, Winners Circle is always able to determine a winner in every race. The absence of dead heats also reduces the complexity of betting on these markets for customers. I'm bothered now :( |
|
By:
Don No1 08 Mar 16:38
£23 matched on the 16:39 at Piccadilly Park, looks like this has really taken off :D |
|
By:
Does seem very strange must admit.
|
|
By:
There are definitely races where they run more than 100 sims. If you look at the daily spreadsheet the win percentages occasionally run to many decimal places, wouldn't be the case for 100 races. I'm sure my earlier explanation is correct -- they run the simulations until each horse has at least one win, otherwise horses which didn't win in the first 100 sims wouldn't have a forecast price.
|
|
By:
SM If a horse does not win any of the simulated races, a default price of 110 will be set.
. Don't think they run til everyones won, plenty of 110 shots quoted, looks like a blip for them to return 70's . You'd need to look at the other odds to try and figure out why they forecast 70 |
|
By:
Ah, I should have read the whole thing instead of scanning through, Baby Jesus. Damn my short attention span.
Still that doesn't explain races like the 13th at Mad Dingo today. Take a look at the upcoming races excel sheet. From the percentages you can see that they haven't run exactly 100 simulations. |
|
By:
SM,
Is there a way to get backdated forecast prices? I can get all the backdated results, and the forecast prices for today, but the column with the backdated forecast prices as per the documentation appears to have been deleted |
|
By:
Column X in the 'upcoming races' excel file show the forecast prices for each horse's previous ten runs - is that what you're after Lori?
Tidbit of info -- I've done correlation analysis on the forecast prices and I can tell you they are quite accurate. I wish they'd do away with the forecast prices altogether, would make things much more interesting. |
|
By:
I'm looking for that, but for historical races.
I had a feeling from some messing about that they were rather too accurate to make the game worthwhile, would prefer 25 trials for some nice errors. |
|
By:
*Thanks for any reply by the way, forgetting my manners here :)
|
|
By:
Also suspect there's some porkies being told. I cant for the life of me find any difference in the skill level of the trainers!
|
|
By:
The whole thing is poorly done imo Lori. The trainers and jockeys are purely cosmetic, each horse flips between trainers and jockeys for each run, a horse can run in two races twenty minutes apart at different tracks -- there's no level of engagement.
If they could create a virtual world which at least in some way mirrored real horse racing they'd be on to something. Until then I suspect liquidity will stay at Norwegian handball correct score levels. |
|
By:
Forgot to add -- even the speed ratings are largely irrelevant.
|
|
By:
Yeah, I've managed to create my own track and stall independent ratings for each horse and found that jockeys do make a difference, albeit an irrelevant one, and I'm sure trainers do nothing at all.
Not only that, but we're working on a sample size of 50 races when it runs a pre-game trial of 100 races, which puts us at a silly disadvantage. Finally I'm pretty sure that "conditioning" is something that has a giant factor in the game and we have no way of finding out what condition a horse is in. I have a few ideas regarding that, but I think it's just a way of saying "We've randomised this enough to screw you over". I could be wrong, but I'm very skeptical. |
|
By:
A good point about 'conditioning', Lori. There are certainly factors built into the races that aren't apparent from the variables we are given. Sometimes I'll see a 1.85 shot and try and figure out why it is so short but there is rarely sufficient explanation in the lines of form.
|
|
By:
The Calculator 08 Mar 13:04 Apparently, in one of the simulations, this horse came second, but the horse in front was found to have virtually taken steroids. The virtual trainer contested these allegations in front of the VBHA, and the case is on-going, thereby confusing the forecast pricer, who doesn't know whether to award a win or not. PMSL. Brilliant. Do you think it will be long before Betfair signs a MoU with the VBHA & we start seeing the virtual jockeys in court for race fixing? |