Euro 2012: England is semi-final probability

Ahead of the first ball being kicked at the Euro 2012 tournament, England is the third-ranked team and has a 68 per cent chance of getting out of the group stage according to the UK's top academic soccer statistician. You read that right. With one of its best players unavailable for the first two matches, a manager who's had …

COMMENTS

This topic is closed for new posts.

University of where?

"UK bookies are well known for relying on home sentiment, and a misplaced sense of soccer entitlement"

No, bookies are not stupid. They don't just do a probability analysis they also do a risk analysis - that is, to put it coarsely: RISK = PROBABILITY x COST. So even if the probability of something happening is relatively low the risk can be high if the cost is high.

In England people are going to bet on England in large numbers. They do this with absolutely no regard for the fact that England have little chance of winning. But if England do win the bookies liability will be huge, hence the risk is high, hence their odds are lowered. But this has nothing to do with the probaility of England winning.

Re: University of where?

Spot on. Even more accurately, they'll be adjusting the odds along the way dependent on the entire betting pool. Effectively they're spread-betting across the punters with a balanced book so that whoever "wins" incurs a cost of less than the entire pot.

If anything, the England bookie odds are more predicated on how much is bet on the other teams.

Re: University of where?

The bookies balance their book so they have to pay out pretty much the same amount no matter who wins.

Because people in England will bet on their team winning regardless of any objective assessment of their team's potential performance, the bookies will be overweight on England, and have to adjust the odds accordingly.

Ask an octopus

Statistics are all well and good - but you can't really predict something as complex as an entire football tournament accurately. I seem to recall these types of reports before the last world cup and they predicted to do rather well. As it turned out, we were poor (well below our standard) and we're torn apart in the 2nd round.

If you ask any half knowledgeable football fan, I recon most of them would tip Spain to win the tournament.... this is without looking at any hard statistics - they just know that Spain are good.

In the end - no one knows who's really going to win the tournament - except maybe Paul (RIP).

ah, bless

Not taking the change of manager into account is a serious flaw.

You can take your stats and ...

Any discipline that has a test called "Tukey's honestly significant difference test" has something to hide.

In 2004 I drew Greece in the office sweep, and thought "that's £2 I won't see again..." A few weeks later I was hiding from a disgruntled Portugese footy fan in a pub...

If it comes down to penalties, bye bye England :) I remember Euro '96 - my appendix ruptured as I watched Southgate put the ball into the keepers hands...true story

"Dr Ian McHale, senior lecturer in statistics in the University of Salford"

So, not cleaned up on the fixed-odds footy betting and retired yet then Ian?

Keep working on those models........

Re: "Dr Ian McHale, senior lecturer in statistics in the University of Salford"

Perhaps he, you know, love his job?

I am seriously underpaid for the job I do but I don't plan on retiring even if a huge pile of dosh falls in my lap.

No hope...

... I have no illusions about England's performance, all I'm hoping for right now is that they score a few goals so I can get some money back from Currys for a laptop purchase.

Does it take into account the uselessness of referees and linesmen when spotting that the ball has in fact crossed the goal line by a significant margin?

I think that's derived from the FIFA technological progress factor. Just divide the result by zero.

then again...

if you look at the other four of the top 5 i don't think many people would disagree with them - so either his england rating is a ridiculous anomaly, or maybe there's some science in there - sometimes science does produce weird but, it turns out, accurate predictions

mine's the coat with straws I've clutched earlier...

Re: then again...

Forget about England, the way to spot that this Prof's stats are a bit dodgy is to note that the Republic of Ireland is being put down at joint-6th favourites with a 7% chance of winning, more than Italy, France and Portugal, while their correct ranking should be right about last, together with Czech, Poland and Ukraine (in fact, marginally less, I would say).

England aren't among the favourites (That's Spain, Germany and , until Saturday, Holland), but I certainly would put them among the outsiders to win (together with France, Italy and Russia). I think Italy's and France's ratings are artificially depressed by the shambles of a World Cup they had 2 years ago, otherwise the prof's stats aren't too far off.

One aspect of the model that seems pretty sound is the balance between favourites and outsiders. Bookies tend to give very short odds on the favourites and very long odds on outsiders, as the first poster noted this is more to do with more people betting on the favourites than this being the real probability. In bookie-world, the chances of the top 50% most-favoured teams would be something like 95%, with the combined total of the others being about 5%. In the Prof's model the top half of the contenders are given 71% combined, with the bottom half given 29% combined. In real life, the last few tournaments (16-team) were won by 3 teams who were among the favourites (Spain, France, Germany) = 75%, and once by a rank outsider (Greece) = 25%. Extending that back through the 8-team format, again there's 3 'favourite' wins (Netherlands, France, Germany) and one outsider (Denmark), again 75%-25%.

Begorrah!

Rep of Ireland ranked 8!

I'll be off now to have a few quid on that, then..

I have England and Russia in the sweepstakes at work.

However, regardless of statistics, I somehow doubt England will do very well....

Erm, am I allowed to write this?

Has Dr. McHale used any racial factors in his statistical analysis when analysing the English Team?

'senior lecturer in statistics'

There are lies, damned lies and statistics.

Thank you, Sir Winston.

Re: 'senior lecturer in statistics'

Not Sir Winston, nor yet Disraeli, a common mis-attribution we can blame on Mark Twain, who really should have checked his facts. Also it's a terrible cliche.

There's a nice variation from "The Accountant" circa 1886.

"... there were three sorts of liars, the common or garden liar ... the damnable liar..., and lastly the expert,.."

Re: 'senior lecturer in statistics'

Mea maxima culpa.

"Variously attributed to Prime Minister Benjamin Disraeli and later quoted by Mark Twain."

Sloppy on my part.

Just in from the University of Elsewhere

A meta-analysis showing that statisticians over-estimate their home side's chances by an average of 32%...

I'm not going to disagree with his rankings, but his probabilities leave something to be desire. Are we seriously supposed to believe that if this tournament were played a hundred times, spain would only win it 12 times? That england would win 10 for every 12 of spain's? That russia would win twice for every 3 of spain's? Nar.

Yes, probabilities are about correct

They looked dodgy to me as well at first glance.... but looking at the history it makes sense. In 4 16-team cups we've had a Greece win (25% probability of a rank outsider winning), and again in 4 8-team cups we've had a Denmark win (again, 25% probability of a rank outsider winning). Typically there are about a quarter of the teams given as favourites (for me, Spain, Germany, Netherlands, France), another quarter considered 'possibles' (Italy, England, Russia, Portugal), and the remaining half being complete outsiders, so divvying up 75% probability between the top 8 teams, and 25% between the bottom 8 gets quite close to the Prof's mark.

12% might be a bit low for Spain but it's probably no more than 15% for either them or Germany

Plural or Singular.

How come Spain ARE plural at one point in this article and yet later Spain HAS a singular verb? FWIW, in the UK, teams are plural.

"In sport, teams are always plural. England are expected to beat the Balearic Islands; Tranmere Rovers have extended their lead at the top of the Premiership."

http://www2.media.uoa.gr/lectures/linguistic_archives/academic_papers0506/notes/stylesheets_3.pdf

Model fail

Hahaha. France 11th, England 3rd, above the Dutch.

Whatever formula he's using it's produced some absolute bollocks.

What does the octopus say?

Paul the Octopus was far more accurate in his predictions last time round. He's no longer with us, but if there are any other octopi around, I am more interested in their predictions than those of football statisticians.

Re: What does the octopus say?

The real question is which enterprising sea life park will produce Debbie the spiritual medium octopus to channel Paul's predictions from the beyond?

Football?

0 fucks given.

...and yet you made the effort to comment?

It's the interweb and it's friday

just here for the lolz

The beauty of this is that you can basically predict whatever you like. The tournament's only run once, there's only one winner, and they only play at most six of the other fifteen teams. You can't run it a hundred times to see if the percentages are correct. Even if Spain go out in the group stage, it doesn't mean the 12% figure was wrong.

No-one will ever know whether Spain had a 12% chance of winning, a 24% chance of winning or a 6% chance of winning. (How do we know there isn't a 50% chance that the tension between the Barca and Madrid camps will escalate into an all-out punchup, leading to an ignominous exit in nearly all those universes?) It's a theory you can't possibly disprove. And therefore, worthless.

That aside, "The model only uses the results of previous International games, the number of goals scored, and location of game." In other words, the fact that England's only international-class creative player is banned for the first two group matches, and they could be out before he gets a chance to play, hasn't been taken into account in ranking them third? Whoops. I know you can't include everything in a model, but if you can't include something like this, then don't bother.

Based on the World cup

We'll bully some minnows early on and get ahead of ourselves, and then as soon as we come across quality in a knockout phase we'll fold like the prima donnas we inevitably show ourselves to be.

Every time I try to give an ounce of care about football (i.e. internationally) we stuff it up.

I shall be watching the cricket - if it isn't raining...

Statistics

73.6% of statistics are made up on the spot.

Re: Statistics

I thought it was 129% of them?

Re: Statistics

I thought that was how much an Apprentice puts into Everything He Does

Are England fans saving money by not traveling to the tournament so that they can spend it in the Bookies?

http://www.youtube.com/watch?v=qykUdp4qRXE

Flawed study

Any statistical study based around the FIFA rankings data is destined to fail.

England the 6th best team in the world?? One below Brazil and one above Argentina? Really???

Classic example of garbage in, garbage out

This topic is closed for new posts.