Statisticians' corner - Bookmaker's implied probabilities for the RWC

Now as we all know, the Bookies aren't always right (or no-one would bother betting). But they generally make a profit, by gauging (a) what they think the probabillities of an event happening are and (b) how they expect punters to bet. So with the RWC only 8 or 9 weeks away I thought I'd have a look at how the bookmakers think things will turn out. So I've taken averaged odds from a number of bookmakers and converted them to probabilities for suggesting where the bookies expect teams to finish.

Probabilities are for the stage at which a team exits - eg "Group Stage" is the probability of a team not reaching the quarter finals, and so on.

Country	Group Stage	Quarter-Final	Semi-Final	Runner-Up	Winner
Argentina	44.8%	47.1%	5.5%	1.8%	0.7%
Australia	0.5%	11.0%	33.1%	35.5%	19.9%
England	6.3%	37.8%	31.5%	15.8%	8.6%
France	2.0%	45.1%	31.3%	15.6%	6.0%
Ireland	8.0%	62.6%	19.2%	7.4%	2.8%
Italy	86.1%	10.8%	2.5%	0.5%	0.1%
New Zealand	0.4%	5.2%	21.9%	17.9%	54.7%
Scotland	46.6%	46.6%	4.4%	1.8%	0.5%
South Africa	0.7%	17.1%	59.2%	10.2%	12.8%
Wales	16.3%	60.6%	17.8%	3.8%	1.5%

Some interesting outcomes there - South Africa rated more likely to win than finish second, presumeably because (if results go with seeding) they're likely to meet NZ in the semi-final then Australia in the final. The draw also makes Australia strong second favourite and England 4th favourite. Argentina are marginally favoured over Scotland to get out of their pool, which is interesting - possibly bookies fishing for bets on Argentina there. And the bookies think there's only 1 chance in 6 that Wales won't get out of their pool. Ireland need to beat Australia in their pool to avoid being on the SA/NZ side of the draw which pushes down their favouritism.

Comparing those odds with what 606V2 posters think here, https://www.606v2.com/t4284p50-who-is-going-to-win-the-rugby-world-cup#277227 (though a first-past-the-post vote result and a probability aren't really directly comparable), it look like 606V2ers are awfully bullish on Ireland's prospects, and have rather forgotten South Africa compared to the bookmakers. Though while researching I did find 1 bookmaker offering 16/1 on them going out in the quarters and just 8/11 to win - I suspect someone's pasted numbers in the wrong boxes.

Senior SAS programming.

I like it. Like what it shows. Bookies don't mess about. The major % shift, as teams progress, as compared to their odds in the group. That's where the bookies make their real money. On blind national faith. And praying NZ lose, to really clean up.

Anyone who thinks SA wont make a SF is mad. And if they made the Final - they'd win it.

But, Oz will. Vs the AB's.

In my humble and righteous opinion.

Yeah, I don't think us Irish really think we're contenders though? Despite the 606v2 poll!

It is gratifying that we're sixth favourite to win though. Take that, 14 other teams.

I think it's more interesting to look at the cumulative totals. For England, we have:

England
Chance of reaching the quarter-finals: 93.7%
Chance of reaching the semi-finals: 55.9%
Chance of reaching the final: 24.4%
Chance of winning: 8.6%

The same can be done for any nation, but it suggests that England are hot favourites to make the quarter-finals, a better than evens chance of making the semi-finals, however we would be underdogs from there. About in line with what most rational England fans expect.

"but it suggests that England are hot favourites to make the quarter-finals"

huh?

At 0.4% exit in pool play NZ are faves to get 1/4's followed by Oz, SA and France.

Youve seen something I havnt. Why cumulative?

Taylorman wrote:"but it suggests that England are hot favourites to make the quarter-finals"

huh?

At 0.4% exit in pool play NZ are faves to get 1/4's followed by Oz, SA and France.

Youve seen something I havnt. Why cumulative?

I think he means hot favourites out of their pool.

Well the bookies make our chances 1 in 4 of beating the All Blacks, I'll take that.

I would have thought our chances would be 1 in 10 in New Zealand.

So NZ punters there is a guarantee of money for you to make.

Just remember that the bookies odds are not just an indicator of where they think the results will go; they can also be a reflection of where punters are putting their money, therefore England will be shorter priced than other nations simply becuase a lot of punters in the English market will put some money on Egnland, possibly more out of sentiment than anything, whereas a serious punt would be on NZ rather than SA reflecting the probable outcome, so SA are slightly further out in the betting.

This is why at every soccer World Cup England are among the favourites with the bookies (as trumpeted in the tabloids/Talksport/Sky Sports) however the odds actually reflect the bookies protectiing their position rather than the statistical odds.

IL

Kiwi , liking the analysis there

Some interesting figures there Kiwi.

Like Ireland are the most likely to go out at the QF's despite not playing NZ....

Mickado, defo some anomalies tho - Argentina less likely to go out at the group stage than Scotland, but more likely to lose their quarter-final (if they progress) than Scotland (likely against the All Blacks!)

I think it's fairly safe to say that if Scotland or Argentina have to face the ABs in the QF, the likelihood of them getting knocked out should be more like 90%.

So Scotland are just as likely to go out at the group stages as they are to go out in the QF? I’m reading that as 50/50 that they make it out, if the odds are longer than ½ for them to qualify then that’s a good bet right? or is the tail wagging the dog in my calculations?

Nah, think you've got it right, Micakdo

Mickado wrote:So Scotland are just as likely to go out at the group stages as they are to go out in the QF? I’m reading that as 50/50 that they make it out, if the odds are longer than ½ for them to qualify then that’s a good bet right? or is the tail wagging the dog in my calculations?

It's slightly better than 50:50 that they make it out - the quarter final odds are just the probabillity that they'll go out in the quarters (which is the same chance as they'll go out in the group). There's a 53.4% chance thay'll go further than the group stages once you add on the chance of going out in the semis, as runner up or as winners.

Taylorman wrote:"but it suggests that England are hot favourites to make the quarter-finals"

huh?

At 0.4% exit in pool play NZ are faves to get 1/4's followed by Oz, SA and France.

Youve seen something I havnt. Why cumulative?

As biltongbek suggested, I meant England were hot favourites to make the quarter-finals from Pool B. Sorry for the confusion, I should have been more clear.

Kiwireddevil wrote:

Mickado wrote:So Scotland are just as likely to go out at the group stages as they are to go out in the QF? I’m reading that as 50/50 that they make it out, if the odds are longer than ½ for them to qualify then that’s a good bet right? or is the tail wagging the dog in my calculations?

It's slightly better than 50:50 that they make it out - the quarter final odds are just the probabillity that they'll go out in the quarters (which is the same chance as they'll go out in the group). There's a 53.4% chance thay'll go further than the group stages once you add on the chance of going out in the semis, as runner up or as winners.

That's why cumulative. So you can see the chances of making each stage, rather than the "point of exit". If you wanted to see the four "most likely" to make the semi-finals, you should look at the cumulative table, rather than the point of exit table.

Eg, France, England, Australia and South Africa are more likely to go out in the semi-final than New Zealand, does that mean they are more likely to make the semi-finals than New Zealand? No, it doesn't, because New Zealand have a 94.5% of making the semi-finals or better, so you can almost pencil them in to the semi-finals right now.

You missed an extra cloumn or bookies odds on choking in each round which throws it all into confusion. sorry kiwi could not help a wee joke