Brazil are the red-hot favourites to win the World Cup, while Australia are ranked as the least likely nation in the entire tournament to make it to the knockout stage. Maths says so.
FiveThirtyEight is run by the recently-famous statistician Nate Silver, who correctly predicted Obama would win the US Election.
In conjunction with ESPN, the site and Silver himself have created an algorithm to predict the outcome of football’s grandest tournament.
The Soccer Power Index (SPI) has four central principles.
- It’s predictive rather than retrospective
- Previous matches are weighted based on importance, however this does not mean a Cup game is necessarily more important than a friendly, with the team’s line-up determining the match’s weight. A team guaranteed to progress in a tournament may rest their stars for the final group match and therefore the game loses significance, whereas certain nations have such strong footballing history, even a friendly will see both sides playing at full strength.
- Teams are given both an attacking and defensive rating based on goals scored and conceded. The idea is to have a high attacking rating and a low defensive one.
- Players for each team are given an individual rating based on their club form. This is the loosest of the four principles but, as they put it, “merely knowing that a player is in the starting lineup for FC Barcelona or Chelsea tells you a fair amount about him.”
The variables are calculated and an SPI is given as a mark of 100. This mark is defined as “what percentage of the possible points a team would accumulate if it played a round-robin against every other national team”. With hours to go before the tournament begins, Brazil are favourites, with an SPI of 94.8 per cent. Australia’s 69.4 per cent is only higher than Algeria’s mark of 63.4 per cent.
But with Algeria in Group H – playing Belgium, Russia and South Korea – they are given an 18.9 per cent chance of advancing, compared to Australia’s 7.8 per cent chance of getting through the ‘group of death’ against Spain, the Netherlands and Chile.
“That’s unfortunate for Australia, which is the odd team out and has less chance than any other squad of advancing to the knockout stage, according to SPI,” wrote Silver.
We are rated a 7 per cent chance of beating Spain, 9 per cent against Chile and 15 per cent against the Netherlands. The likelihood of losing is above 60 per cent in all three matches.
Meanwhile, Brazil are considered all but certainties of progressing to the knockout stage of the tournament, SPI giving them a 99.4 per cent likelihood of progressing from Group A (Mexico, Cameroon and Croatia).
Yup, that’s Australia at the very bottom. (via FiveThirtyEight)
Furthermore, Brazil are a 45.2 per cent chance of winning the whole caboodle. Second favourites Argentina are given a 12.8 per cent chance, Germany are third at 10.9 per cent and Spain are fourth with 7.6.
Australia are given a 0.0 per cent chance of holding the Cup aloft, although it may serve as some comfort that Algeria, Honduras and Iran are given the same likelihood, and 22 of the 32 nations competing are rated a less than 1 per cent chance.
By SPI stats, Brazil are more likely to win the Cup than almost every other nation combined. Their recent form at home suggests the SPI may be spot on.
Brazil haven’t lost a match on home soil since 2002 and that was a friendly against Paraguay to celebrate Brazil’s victory at the 2002 World Cup. The last time they lost a game that actually mattered in Brazil was a 1975 match Copa América match to eventual champions Peru.
To put that in perspective, no member of Brazil’s playing squad was even born the last time they lost at home.
As the tournament progresses so will the SPI. You can stay up-to-date with its wild fluctuations as Australia breeze inflict three punishing victories by clicking here.
This will be the first World Cup where goal-line technology is used. Check out Joe’s story on the amazing seven-camera system at Techly.
Andrew Kennard
Roar Pro
Your comment about home continent weighting is slightly misleading I think. The only teams to ever win outside their home continent are Spain (once - South Africa) and Brazil (twice, Sweden 58 and Japan/South Korea 02). So while Belgium may go farther than Chile, to win the world cup would be the first time a European team has ever won in the Americas. The proof will be in the pudding, I'm excited just to wait and see.
Steve
Guest
I think it's just a reflection of Nate's model giving a big weighting to the fact the tournament is in South America. The model seems to be relying on this being a major advantage for not just Brazil but other local teams. It would be interesting to see this model against a 'Power Ranking' type model (as is used for NFL and NBA) that ranks teams on neutral fields.
Beauty of a geek brains of a bimbo(atgm)
Guest
Hahaha chile at 5?
Steve
Guest
Not sure what you're implying here, Silver predicted all 50 States in 2012, and 49 of 50 states in 2008. He'd do pretty well to better that.
Ret
Guest
I hope Nate Silver has had more successes than predicting Obama's victory.(Umm, charismatic incumbent v wooden, silver tailed challenger) If not, I'd be happier with last WC's octopus.
Mantis
Roar Guru
I think Belgium will go well, but im concerned about their players lack of experience on the international stage. All the big name sides have players who have the experience in big international matches. Belgium have big club experience, but whether or not that translates to international will be seen.
Mantis
Roar Guru
Good point. I was tossing up between the Argies and Brazil, and this might have tipped the scales in Argentinas favour for me.
TheBlackCat1859
Guest
0.0% isn't "exactly zero" - it's just zero to once decimal place.
apaway
Roar Guru
Nate Silver gives Colombia and Chile a greater chance than Belgium and Italy. This suggests to me a weighting in favour of the home continent sides. With nothing more than football instinct to draw on, I think Belgium will go further than both Chile and Colombia.
The Bush
Roar Guru
I was at the friendly between Argentina and Trinidad and they're gonna wanna play a hell of a lot better if they plan to win!!! But seriously, obviously he's on the money cause there isn't much difference between those odds and the betting agencies. The only difference is that people will vote on the Italy's of this world, amending their odds... (Plus tradition must play a part - it is Italy...)
Ben of Phnom Penh
Roar Guru
Bookies need to take into account the amount of betting being undertaken as it affects their portfolio of risk and hence they will adjust odds to account for this and to ensure their betting portfolio reflects the risk exposure that they have calculated.
Steve
Guest
Gotcha. So they likely hire their own geniuses, who just don't have Nate Silver's profile I assume.
clayts
Guest
They do both. Depending on the type of bookie, they will set a market, then depending on what others are laying backing, and the money that comes in for their own market, they will take their own positions to try and play percentages. That's the art of bookmaking
Shouts Chen
Guest
Australia winning the World Cup with a chance of 0.1%? You gotta be kidding.
Steve
Guest
Do bookies not just take a spread these days, or do they take positions also?
Tom
Guest
How does Nate's method compare with those used by bookies? Does he consult to any of the major betting houses?
Steve
Guest
I'm glad someone wrote about this, I was considering submitting an article. I have the utmost respect for Nate Silver and his predictions methods. That said, I do think he is placing too much emphasis on home advantage, while missing a key variable - a players age. Look at the great names, and basically with the exception of Pele and Ronaldinho (who was not the main man when he won), all of them won their World Cups between the ages of 25 and 30. Maradona, Zidane, Romario, Ronaldo, Rivaldo, Pirlo, Xavi, Iniesta, Bobby Moore, Bobby Charlton, Garrincha, Beckenbeuer. All of them. So with that, can Neymar (the outstanding player for Brazil) really carry his team to a title at age 22, even if it is a home Cup? I think Brazil can, but their odds should be down weighted - 45.2% is a bit high. I think the Betfair numbers are fairer Meanwhile, Spain's best players are either at the upper end or over this age bracket (Iniesta 30, Xavi and Xabi Alonso 34 and 32). The team who have most of their best players in this age bracket are Argentina (Messi, Aguero Di Maria, Zabaleta, Higuain, Mascherano, Garay Lavezzi). Germany are another side who have many key players in this age bracket, although some of their stars are younger - Gotze, Schurrle, Muller, Draxler, Kroos - and their big time forward in Klose is 36.