Number crunchers like Jockey Club odds

PUBLISHED : Tuesday, 03 June, 2008, 12:00am
UPDATED : Tuesday, 03 June, 2008, 12:00am
 

If you were pushing through the turnstiles at Happy Valley one race evening, and a stranger whispered in your ear 'Doxy's Daughter to beat Likely Lad by a nose in the 8:45', you probably wouldn't pay too much attention.

But if Doxy's Daughter duly romped home a nose ahead of Likely Lad in second place, your interest would certainly be piqued. If your mysterious tipster then approached you again the following week and muttered 'Blue Beggar by a length from Dirty Duke' you would probably be tempted to rush in and bet your bonus on a quinella.

In this case, however, the tip isn't for a horse, but a football team. Two years ago, Andreas Hofert, the global chief economist at UBS Wealth Management Research, pulled off a forecasting feat that would make any racecourse tout glow with pride.

On the eve of the 2006 football World Cup, he fed heaps of historical data about the competing teams and past tournaments into the sort of number-crunching model normally employed to determine stock and bond market allocations, and confidently announced that unfancied Italy would lift the trophy.

As fans will remember, Italy did indeed go on to beat France 5-3 on penalties in the final, held in Berlin's Olympic stadium.

That wasn't the only thing Mr Hofert got right. His model also predicted that second favourite England would crash out in the quarter finals. David Beckham's boys duly lost their quarter final match to Portugal, again on penalties.

After a coup like that, Mr Hofert could be forgiven for hanging up his laurels and attempting no more sports forecasts. But he is at it again. This time he has plugged a wealth of data into his model to forecast the most probable winner of this month's Euro 2008 football championship.

He reckons the Czech Republic will meet Italy in the final, with the Czechs having a 51 per cent probability of winning (see the chart below).

To come up with his forecast, Mr Hofert has combined several historical data sets, including the number of times teams have reached the semifinals, their Elo ratings, which gauge teams' recent performances, and the number of outstanding players in each team, according to Pele's 2004 rankings. The model is then weighted for factors which have a statistical significance in past results, like geographic advantage.

Unfortunately for Mr Hofert's reputation as a tipster, the European Championship is harder to forecast than the World Cup. In 18 tournaments, only seven teams have won the World Cup. But 12 European Championships have produced nine winners.

That partly explains the narrowness of the Czech Republic's expected advantage. Mr Hofert suggests the final may be a draw, with the Czechs winning on penalties.

Such predictions, of course, are absurd. At best Mr Hofert's model is only a loose probabilistic guide.

His forecast that the Czechs have a 51 per cent chance of beating Italy in the final should be taken with a hefty pinch of salt. If the model is run cumulatively, rather than started afresh for each round of the competition, the Czech Republic is still favourite but with only a 7 per cent probability of emerging as the victor. That means there is a 93 per cent chance the winner will be someone else.

Even so, Mr Hofert has been right before. And with the Czechs currently a 16-1 eighth favourite according to the Jockey Club, the team might well be worth a flutter. Just don't bet your bonus.

Share

 

Send to a friend

To forward this article using your default email client (e.g. Outlook), click here.

Enter multiple addresses separated by commas(,)

For unlimited access to:

SCMP.com SCMP Tablet Edition SCMP Mobile Edition 10-year news archive