It's coming home?

All those oh-so-nears wear you down through the years

The Euro 2024 football tournament kicks off on Friday 14th June, so it’s time for a Logging the World preview from a probability and betting point of view. In particular, I’ve seen a certain amount of optimism about England’s chances so I’d like to figure out if - finally! - football might be coming home.

In brief, I think the answer is “it might be”, but that’s all I’m prepared to say. My contention is that no single team has a huge chance of being crowned as champions in Berlin a month later on 14th July, and so if you’re planning to make a bet then your strategy should reflect that, and you shouldn’t get too optimistic about the favourites.

I’m not going to talk about formations, pick out star players or compare coaching philosophies. But I think there are two things you need to bear in mind from a mathematical point of view: a) Fermi estimation b) 24 is not a power of 2.

It could have been all songs in the street

Of course, England have improved immeasurably from the 2014 World Cup, when our only point was a goalless draw against the mighty Costa Rica. Despite a disappointing couple of warmup games, I’m still optimistic that we’ll do better than we did in the Euros two years later when we fell to the Viking Thunderclap in the last 16. Our subsequent record of a final, a semi-final and a quarter-final, along with the presence of world-class players like Kane, Bellingham, Foden and Rice, makes us a team to be reckoned with, and I think it’s fair to consider us among the favourites.

However, that’s a long way from saying that an England win is certain, or even likely. In fact, you might even say it’s “highly unlikely”, in the probability descriptive framework adopted by SAGE during the pandemic¹, and I’d like to explain why.

Let me illustrate my argument using the last European championships. I don’t think anyone would argue that France have been the strongest team in Europe lately: they reached both of the last two World Cup finals, winning one and losing the other on penalties. Yet, in the meantime, they only won one group game out of three in the delayed Euro 2020 tournament, and lost on penalties in the last 16 to Switzerland.

Meanwhile, you may have a memory that Italy’s Euro 2020 tournament win was somewhat inevitable, reflecting a great side. Yet, they required one extra time win and two wins on penalties to navigate the knockout stage, and in advance of the tournament they could be backed at odds of 11/1. Indeed, they didn’t even qualify for the subsequent 2022 World Cup, having been knocked out in a playoff by North Macedonia.

Similarly in 2016, the winners Portugal drew all three of their group games, and needed two extra time wins and a penalty shootout, and were given odds of 16/1 before the tournament.

It’s all very well looking at star names, but would anyone honestly claim that this England lineup (from the 2004 game against Croatia, currently on iplayer) was a worse set of players than the Greece team which won that tournament?

In fact, unless you are Greek or a big football nerd, I’d honestly be impressed if you can name any of that winning team off the top of your head. Luck (injuries, refereeing decisions, the bounce of a ball) seems to be a hugely underrated factor.

Everyone seems to know the score

Overall I think that, from the quarter finals onwards and perhaps even before, modern games of international football are far closer to a tossup than you might think. Teams don’t get long to prepare together, and competition with lucrative club fixtures means that internationals aren’t always a high priority for top players. Looking back on the tournament, we convince ourselves that the outcome was inevitable when it was anything but, and there’s often a case of survivorship bias in our analysis.

Understanding this allows us to think more dispassionately about the odds, with the favourites listed here and converted into probabilities:

• England 10/3 (23% chance)

• France 17/4 (19%)

• Germany 11/2 (15%)

• Portugal 8/1 (11%)

• Spain 9/1 (10%)

• Italy 18/1 (5%)

• Netherlands 18/1 (5%)

I believe that understanding this is a good example of what is referred to as Fermi estimation, where trying to answer a question in one go leads to a less accurate answer than breaking it down into stages.

If you ask me “will England win?” then my feeling is perhaps “sure, they might well do”, and I might come up with a figure like the 23% quoted above. But if it were really the case that the quarter-finals onwards were all genuine tossups, then no one team can have more than a 12.5% probability of winning (even if they were guaranteed to make it that far).

Even if you believe that the better team will be somewhat more likely to win each game, then (as with my earlier piece about Ronnie O’Sullivan) that is a long way from guaranteeing a tournament win. If a team has a two-thirds chance of getting through each round (and would you honestly be that optimistic about England in a semi-final or final?!) then they would have less than a (2/3)^4 or 20% chance of winning the tournament.

Of course, this doesn’t prove anything. But if you believe these kinds of arguments are plausible, then it might suggest that the teams at the top of the list have been given too high a probability by the bookmakers, so it would not be rational to bet on them at these prices. And if that’s the case, it suggests that (as in previous years) then there might be value further down the list. But where?

I know that was then but it could be again

In order to understand this some more, we need to consider the other fact that I mentioned: 24 is not a power of 2. Why does this matter?

Euro 2012 was a 16 team tournament. There were four groups, the top two teams in each qualified for the quarter-finals, which consisted of a group winner against a runner-up. As a result, the group winners all typically had the same kind of advantage, and couldn’t meet another group winner until the semi-final. This kind of structure tends to lend itself to predictability, because a seeded draw tends to keep good teams apart.

However, since 2016 the European championships have expanded to 24 teams. There are now six groups, and an extra knockout round. But because six isn’t a power of 2 either, there’s not a single stage where the group winners first come together - some groups have a different experience to others. We need to dust down our wallchart and study the fixtures a bit.

A key factor is that the structure of the tournament implies that the best groups to be in are C and F, and there is a disadvantage to being in A,B, D and E. The reason for this is that the winners of group A (containing Germany) will likely play the winners of group B (containing Italy, Spain and Croatia) in the quarter-final. Similarly the winners of group D (containing France and the Netherlands) are likely to have a quarter-final against the winners of group E (containing Belgium). Meanwhile the group C and F winners can’t meet another group winner until the semi-finals. It’s not a decisive advantage, but it matters.

So, if I were looking for value, my suggestion would be to focus on those groups. The good news is that Group C is England’s group, which perhaps contributes to the sense of optimism in the odds mentioned above - though even a potential quarter-final against the runners-up in the Italy-Spain-Croatia group wouldn’t be easy.

On the other hand, the route looks relatively smooth for the winners of Group F. The seeded team there are Portugal, and you’d give them a good chance of winning a group also containing Georgia, Turkey and the Czech Republic. Taking a wild stab at how groups might pan out, there’s a plausible route to them playing the Netherlands in the quarter-finals perhaps.

But it’s also worth bearing in mind that not all groups will pan out the way that we expect, and (again as a result of the 24 team structure) there is a chance for third place teams to qualify in complicated ways that can place them in strange and quickly-changing positions in the draw. Even if groups go as you might expect, this could mean teams like Croatia or Poland showing up as a last 16 game for a group winner. Again this can make it harder to pick clear routes through the tournament, and mean that the right betting strategy may well be adaptive as things develop.

However putting it all together, and looking at the odds above my sense is this: England definitely can win, though the lack of cutting edge in their performance on Friday against Iceland was definitely a worry. However, the chances are very much that they won’t, just because any one team doesn’t have a huge chance. And if I were looking for value in the betting market, then based on past performance and the draw, I’d be looking at teams like Portugal and the Netherlands (and perhaps even a little further down the list), at least as a trading bet.

But the main thing is that whether this is right or wrong, I hope you enjoy watching the games, but that you don’t take the outcomes of these random processes too seriously!

1. See for example here