Probabilities of World Cup victory:
1 New Zealand 65.1966%
2 Australia 16.0248%
3 South Africa 7.9437%
4 France 7.3474%
5 England 2.1818%
6 Wales 0.5421%
7 Ireland 0.3734%
8 Scotland 0.2265%
9 Argentina 0.1302%
10 Samoa 0.0136%
11 Italy 0.0134%
12 Georgia 0.0045%
13 Japan 0.0014%
14 Fiji 0.0006%
15 Canada 0.0002%
16 Tonga <0.0001%
17 Romania <0.0001%
18 Namibia <0.0001%
19 USA <0.0001%
20 Russia <0.0001%
N.B. New Zealand's home advantage is worth 63.15 ratings points.
Assumptions:
1. Probabilitiy of a draw is assumed to be zero.
2. If two teams were tied on points at the end of the pool stage, the
probability of a team being ahead on bonus points is assumed to be the same
as the probability of a head-to-head win.
Comments:
New Zealand are clear favourites due to having a clear lead at the top of
the ratings, combined with home advantage.
Myk
Just a quick follow-up. Looking at Betfair odds of World Cup victory, it
appears that if the above probabilities are accurate, that the only teams
that are good value bets are:
New Zealand @ 1.77
France @ 19
Related question: Why the hell is Russia paying 500, when Italy is paying
910? Not that either of them has much chance, but who's laying those odds?
Myk
> New Zealand are clear favourites due to having a clear lead at the top of
> the ratings,
Weren't they the bookies' pre-tournament favourites/joint-favourites in
1995, 1999, 2003 and 2007, too?
I wonder if similar or greater probabilities were generously forecasting
NZ winning the Tri-Nations?
Probably, because on paper they should have won it. But these stats don't
reflect the rough trend that although NZ can beat everyone in the world in
tests without even trying, they're very capable of losing the big games that
*really* matter.
All teams are capable of losing the big games that *really* matter.
What's your point?
Two Dogs
I think he's saying that just as New Zealand are consistently better
at winning, so they are consistently better at losing. They always
target the biggest prize to lose - no half measures for them!
UD
France 3x as likely as England? Love to see how.
How can they be consistently better at losing? Namibia is
consistently better at losing. Romania is consistently better at
losing. England is consistently better at losing. A team that is
consistently better at winning *cannot* be consistently better at
losing. You can't have it both ways.
Even if you take the small sample that is the RWC you will find that
South Africa has an 86% winning percentage, Australia an 85% rate, and
New Zealand an 83% rate. Not a lot in it, really, for the top three
teams in the world.
England, on the other hand, is a full 10 percentage points lower
than New Zealand, which makes them consistently better at losing at
Rugby World Cups.
Two Dogs
England have a tougher pool, with two other teams with a decent chance.
France's probability of advancing from the pool is 94.0%, while England's is
82.4%. In that, England have a fair chance of coming second (25.7%) in which
case they are almost certain to meet the hosts in the quarter final, whereas
France can't meet New Zealand again until the final. And, France are a
little higher in the ratings than England so if they meet in the QF France
has a higher probability of advancing (55.9%). So due to a combination of
those things, England has a much lower probability of winning overall.
If England beat Argentina on Saturday, their probability will increase a
lot.
Myk
In 1995 pre-tournament bookies favourites were Australia (justly so,
they having dominated internationally since winning the 1991 RWC). Not
counting a cup warm-up match against Canada, NZ had won only two of
their last 6 matches before the 95 cup (losing to France and Australia,
beating only South Africa). Some would say NZ were "joint" favourites in
2003 on the basis that favouritism was too close to call, with different
bookies citing either NZ or England, but I think more money was on
England overall. If so, it would be the only time the bookies favourite
won (Australia were favourites in 1987).
The thing is that if the bookies are good at their job (and they usually
are) then favourites are the most likely to win. But only in the long
run. Not necessarily in a sample as small as 6. And even in the long
run, in a field larger than two to say that the favourite is the most
likely to win does not mean that the favourite will win more often than not.
Apteryx
I think you have to take this in the context of NZ generally being the
favourites.
Myk,
I'm guessing your probability calculations are solely based on what a
team has achieved in the past. How do you account for a team at the
peak of their performance or starting to wane like England (2003) and
I believe the AB's in this world cup? I would guess the bookies take
this into account as theirs is a judgemental call.
Stex
I certainly didn't. You can't lose home and away and be a joint
favourite, however questionable the wins -- the points were still on the
baord.
> with different
> bookies citing either NZ or England, but I think more money was on
> England overall. If so, it would be the only time the bookies favourite
> won (Australia were favourites in 1987).
Bookies just follow the money. And the money follows the average punter.
Neither have anything to do with reasoned analysis.
>
> The thing is that if the bookies are good at their job (and they usually
> are) then favourites are the most likely to win. But only in the long
> run. Not necessarily in a sample as small as 6. And even in the long
> run, in a field larger than two to say that the favourite is the most
> likely to win does not mean that the favourite will win more often than
> not.
The favourites are more likely to win a series. The random natur eof
world cups means favouritism (and ability) is of limited value.
That is to invent outcomes that aren't really there at all. All rugby
test sides are very capable of losing tests from time to time without
any particular pattern.
It is tempting to cherry pick the "big games" from that selection and
accord them different status, but that's just making things up after the
fact to suit personal prejudice.
So you're saying that New Zealand have a 65% chance of whitewashing a
three match test series, possibly against Argentina, South Africa and
Australia?
I'm wondering how many times in the past they have managed three-test
whitewashes -- less than 65% I'm guessing.
What's the basis of the maths you're using here, without confusing my
simple brain with complicated equations and things?
You do realise that someone has to win the tournament?
--
cheers,
calvin
But before the tournament starts, no team has to have a better than 50%
chance of winning (and in a field with 5 teams having realistic chances
and maybe 4 or 5 others having unrealistic chances of winning but
realistic chances of beating any one of the top 5) it would be unusual
for any team to have a 50% chance). Personally I agree with Myk and the
bookies than NZ are the team most likely to win. But I disagree with
them that their chances of winning are over 50%.
Apteryx
NZ at $1.77 is ridiculous
We are now in the quarter /semis /finals and the opposition will be top
teams ...so for 3 games its hard
So what % chance do you give other teams? And what do they sum?
--
cheers,
calvin
More or less agree with your percentages. I would add about 5% to 10%
to NZ and subtract it pro rata from the others.
Even if Myks numbers are spot on it still means there is more than a
1 in 3 chance that come October 24 there will be much complaints over
refs or chefs .
Myk,
Stex
*****
It's solely based on past performance, as embodied in the AQB ratings. See
my reply to Rick.
Myk
The ratings can be used to generate a probability of victory in any given
game, based on the difference in ratings between the two teams. The higher
the difference in your favour, the higher the probability of victory.
Using a bit of computing power, I can then work out the probability of any
given combination of results occurring. There are 47 games in the World Cup
(excluding the losers final which is a waste of everyone's time, including
the teams involved), so there are 2^47 or 1.4x10^14 possible permutations
(actually its much simpler than that since there are some outcomes that
can't occur, and I really treat each pool as a separate set of calculations
and calculate the probability of each team advancing in first or second).
Then simply add up the probabilities where each team wins to get their
probability of victory.
Myk
So you're telling me there's a chance? Yeah!
Watch out! One of those juvenile whooshes just missed your bonce...
UD
Eeeeeeyeah..... So a team that has a track record of doing well to make
the 3-4 playoff is the dead cert.
I suppose this is why we play the game, to prove stat fans wrong.
I bow to your expert knowledge on the subject.
So? We've beaten the teams that knocked you out, twice, whilst they were
playing at home. It's the RWC, NZ just aren't an issue.
> "Myk Cameron" <m...@image.co.nz> wrote in message
> news:9cmgn8...@mid.individual.net...
>> Using the AQB Ratings 7 September 2011:
>> http://www.image.co.nz/aqb/rugby_ratings.html
>>
>>
>> Probabilities of World Cup victory:
>>
>> 1 New Zealand 65.1966%
>> 2 Australia 16.0248%
>> 3 South Africa 7.9437%
>> 4 France 7.3474%
>> 5 England 2.1818%
>> 6 Wales 0.5421%
>> 7 Ireland 0.3734%
>> 8 Scotland 0.2265%
>> 9 Argentina 0.1302%
>> 10 Samoa 0.0136%
>> 11 Italy 0.0134%
>> 12 Georgia 0.0045%
>> 13 Japan 0.0014%
>> 14 Fiji 0.0006%
>> 15 Canada 0.0002%
>> 16 Tonga <0.0001%
>> 17 Romania <0.0001%
>> 18 Namibia <0.0001%
>> 19 USA <0.0001%
>> 20 Russia <0.0001%
>>
>
> Just a quick follow-up. Looking at Betfair odds of World Cup victory, it
> appears that if the above probabilities are accurate, that the only teams
> that are good value bets are:
>
> New Zealand @ 1.77
> France @ 19
>
> Related question: Why the hell is Russia paying 500, when Italy is paying
> 910? Not that either of them has much chance, but who's laying those odds?
>
*Whoosh* Someone mentioned Betfair?
Betfair don't "lay" odds like a traditional betting house/bookie Myk. It's
punter v punter. From good 'ol Wikipedia.
"A betting exchange allows punters (gamblers) to bet at odds set and
requested by other punters rather than by a bookmaker. Members can make
both 'back' bets (normal bets on a selection to win) and 'lay' bets (bets
on the opposite side of the Back, against the selection), thereby
eliminating the traditional bookmaker.
Betfair claims on average 20 percent better odds than those offered by a
traditional bookmaker.[3] Betfair charges a commission on all winning bets,
which is set at 5 percent of the net winnings for most markets, although
according to how much a client wagers on the site, it is possible to reduce
the amount of commission paid to as low as 2 percent.[4]"
There's so little money on Russia & Italia that the 'odds' are meaningless.
$106 & $1,880 respectively.
as
If there's a million pounds put on Tonga for the opener and only a
thousand put on the ABs, who do you think will be bookies favourites ?
>>
>
> Eeeeeeyeah..... So a team that has a track record of doing well to make
> the 3-4 playoff is the dead cert.
> I suppose this is why we play the game, to prove stat fans wrong.
Really? I thought it was for the money & women.
as
> ABs were favourites in 1987 (when have they not been favourite to win at
> home?).
WTF? That is complete and utter rubbish. The Wallabies had beaten the All Blacks in '86 in New Zealand (2-1 but it should have been 3-0 except for Bevan's APPALLING decision). So the Wallabies were clear favourites based on form over the previous year. Of course, no one at the time knew what a young and truly brilliant team the '87 All Blacks were to become. That only became obvious over the next few years.
And only in retrospect can we now see that Alan Jones' coaching style for the Wallabies was bound to cause eventual burnout. They were going downhill in '87.
>
> Same can be said for 1991, although in hindsight they may have been on the
> wane.
Yes, exactly. It was the turn of the Wallabies to have a young emerging team and for the All Blacks to have a team on the (relative) slide after a period of dominance.
>
> In 1995, perhaps not outright favourites, but certainly up there with Aus
> and SA. Results from the previous year not so good, but they were clearly on
> the up.
Australia was favourite off the back of what had happened over the preceding few years, but anyone with half a brain could see they were a team on the slide and full of dysfunctional older players like bloody Campese (who couldn't even find touch from penalties ... twice in a row in one game, from memory. Infuriating stuff)
>
> Again in 1999, ABs were favourites having won the tri-nations.
A TN win, yes, but that was very misleading and didn't fool many ...
The Wallabies had played the TN without Larkham, Horan & Eales (who were injured). Wasn't Rod Kafer at 10 for part of that TN? The Wallaby team which won the RWC was VASTLY stronger than the TN team. I remember being very, very confident about the Wallabies prospects going into that RWC. Excellent coach, excellent captain, really well balanced team.
Which is not to say the ABs weren't pretty handy too. Just to say that the ABs and Wallabies were equal favourites for me personally, going into that RWC. The TN didn't come into it.
>
> In 2003 ABs are favourites ahead of the Poms.
Very close for favouritism between the two with England slightly in front. Them meeting in the final was an intriguing prospect.
>
> In 2007 ABs are favourites once again.
Yeah probably.
But the Boks were clearly a fantastic side, perfectly adapted for the rules at that time (kicking more than they passed), beautifully coached and fantastically captained.
In retrospect, the Boks should probably have been equal favourites, but, yes, it wasn't that clear going in.
--
Mike
And this is my argument - these are not "just" test matches, they're
huge-pressure, one-off KO matches and that's what I think makes the
difference.
My honest opinion is that the AB players won't have felt this kind of
pressure anywhere except the RWC, and it builds and builds. Lose a game on
tour and you're still on tour & still top of the rankings, lose a 3N game
and you're still in the 3N. Lose a RWC knockout game and it's all over.
There are no second chances after the group stage of the RWC.
RWC history shows that NZ have a 33% probability of reaching the final, and
a 16.5% chance of winning it. England have a 50% probability of reaching the
final and also a 16.5% chance of lifting the trophy.
NZ are more likely to play in the 3rd/4th place playoff than play in the
final.
Slightly flawed logic in that NZ will be at home. How often have they
won 3 in a row @home against top opposition? Quite often I would
guess.
But this is the WC and as you say pressure could tell.
Also the opposition probably want it more than they do in a dead
rubber.
There was no TAB sports betting in 1987. But regardless of what you are
certain of, Australia had the lowest quoted odds with the British
bookies. People are confused about the quoted odds in 1987 and 1995,
because by the time they reached the finals in both those years, NZ were
certainly favourites to win it. But they did not start either tournament
in that position. The shortening of the odds on NZ in 1987 was
understandable given that by the time they reached the final against
France, their closest win had been by 30-3 over Scotland in the QF, a
team who had drawn with France in pool play. And they beat Wales 49-6 in
the SF (in Brisbane, because Australia were co-hosting that RWC), and
Wales went on to win the 3/4 play-off against Australia, who France had
beaten to reach the final.
Apteryx
> >> Much less scientifically than Myk, I'd rate NZ as about a 40% chance,
You seem to have somehow found a third team of the quality of SA/
Australia for us to play in the quarters. Not sure why. Two of the
possible opponents have never beaten us, and the third has lost to us
9 times in a row.
And as far as reliable prediction, I never claimed otherwise - a
winning chance of 66% leaves plenty of room for horrific failure.
Cheers
Brent
That sample is too small and therefore dodgy.
By those stats SA has a 50% chance of making the final and a 100%
chance of winning it if they get there. So 50% chance of winning?
Yeah right
> "alvey" <al...@atattat.com> wrote in message
y.
>>
>
> Thanks for that. Hence the question: Who is laying those odds on Russia and
> Italy?
Drunks.
as
I've included all 6 RWC finals, SA haven't won all of them so their chances
this time can't be 100%.
Boks have appeared in 2 out of 6 finals, so their chances of reaching
another are 1 in 3, or 33%, and their chances of winning the final are half
of that - 16.5%, same as NZ. I think you'd probably have to factor in some
kind of handicap for each team based on how long ago they last reached the
final or won it, so on that basis I'd say the Boks are more likely to reach
and win the final than the ABs.
Could be anyone, you could lay them if you wanted to. Just lay Russia at
slightly more than the current best, and you might get a few punters
sticking a cheeky rouble on them, some people do back the outsiders as well
as the favourites so that they still make money no matter the outcome. In
theory you could lay them at $50,000 because they're never going to win it.
Or at least you'd hope so if you're laying them at stupid odds.
Presumably Italians are more likely to back their own country out of sheer
jingoism, so their price will fall while Russia's goes up. It's the reason
why England football team always seem to be low-priced at bookies, it's
because so many English fans back them even when they know we don't have a
hope.