Finest Hour, Darkest Hour

The FIFA rankings, required to be used in any seeding of national teams for competition, have never exactly had a stellar reputation amongst fans. They were introduced in December 1992, because somehow soccer didn’t have national rankings at all before 1992, and were quickly and repeatedly picked apart through multiple iterations. The version used from 1993-1998, for instance, ranked teams the same way you’d rank clubs on a league table: three points for a win, one for a draw. And then, after giving a timeframe, they just left it like that. That is, though, still the basic seed behind assigning points; it’s just that now there are other factors to it… none of which are margin of victory or the location of the match; these were actually removed in 2006. Continental coefficients are involved, updated only after World Cups and only taking into account the continent’s World Cup performances.

Because of their status as literal garbage, competing (but, to many’s chagrin, unofficial) rankings have popped up. Chief among them is the Elo Ratings, which are quite often used by onlookers over and above the FIFA rankings as a truer statement of relative ability. The Elo Ratings, based off the system used to rate chess players and since adopted to all sorts of other competitions (including the NFL, as per FiveThirtyEight), give every individual nation their own individual strength rating- the rating they’re being ranked with. Location is considered. Margin of victory is considered, with diminishing returns as the goals pile up, and actually, FiveThirtyEight did a pretty good job giving the general gloss-over when they adopted it to the NFL.

But let me give the 99-cent version. The ratings assigned to each team, along with a bonus given to the home team (if any), are essentially a reflection of what Elo expects would happen in a game between the two. The game being played is weighted on the basis of the status of the competition- a World Cup match will get more weight than a World Cup qualifier will get more weight than a friendly. The scoreline of the match is measured against what Elo projected, and points are exchanged accordingly. Winning a match will gain you points; losing will cost you points (except on uncommon occasions where no points end up being exchanged at all). Points are exchanged in a zero-sum fashion: any points you gain come directly out of your vanquished foes’ stash, and vice versa.

What I’ve been doing- and I hope you appreciate this; this was exhausting to put together- is going through the Elo Ratings of each and every squad currently existing in the database- that’s 234 teams, some of which aren’t in FIFA but are scored anyway because you never know if they will be someday, with defunct nations such as East Germany and South Yemen recorded but set aside- and recording the match from which they gained the most Elo points (which I’m referring to as their ‘finest hour’) and the one in which they lost the most (their ‘darkest hour’), as of the latest available results on February 11. If there’s a tie, and there were several, I simply recorded all tied matches.

You can find the table in Google Docs here.

The purpose here is twofold: first, to sate a bout of curiosity that now makes me wish I could trade my brain in for one that asks me to do actually fun things; second, to see how well the calls made by Elo mesh with the actual memories of fans.

As it turns out, it swings fairly wildly in that regard. Very rarely is a friendly, or what was regarded by Elo as a friendly tournament (which I’ve gone ahead and identified when the Elo site didn’t), anyone’s finest or darkest hour, and those for whom it is are among the weakest, and not likely to play very many competitive fixtures anyway. Past that, though, Elo has its own ideas as to what is to be celebrated or shamed more than anything else. Multi-goal margins are abundant on the lists, exploiting the margin-of-victory amplifier. But it’s not always the biggest margins that get there either. Australia’s infamous 22-0 win over Tonga and 31-0 win over American Samoa in the 2002 World Cup qualifiers, the ones that helped catalyze their move from Oceania to Asia, went barely noticed, if at all, by Elo because of the gigantic disparities in talent… which was kind of Australia’s point. The Tonga win only garnered one measly point for the Socceroos. The American Samoa win, even at a 31-goal margin, got them nothing at all.

Beating good teams wins big for you, losing to bad teams costs you big. Merely taking care of business won’t make for large swings.

A match that screams in the newspapers may be far quieter in Elo. It may depend on whether perceptions of the teams involved were properly calibrated. England’s momentous loss of face in the early 1950’s, where it was hammered home to them once and for all that the soccer world revolved around more than just a British axis, is a good example of this. The 1950 World Cup loss, 1-0 to the United States, did in fact cause a big Elo swing, with 56 points transferring to the Americans, easily enough to be many countries’ finest or darkest hours. There was no way that would have happened 9 times out of 10, or even 24 times out of 25. However, the 6-3 loss to Hungary at Wembley in November 1953, and the 7-1 pounding in the rematch in Budapest six months later, registered far less; the latter registering as an 18-point swing and the latter only 10. The English had still thought of themselves as the best in the world, but Elo had begged to differ, ranking them only fourth on the day of Wembley behind third-ranked Brazil, second-ranked Argentina… and top-ranked Hungary. Elo was not overly surprised at what had happened, but then, Elo is an emotionless algorithm, not anyone who had lived, died and bled Three Lions red.

Brazil’s national meltdowns over the 2-1 loss to Uruguay in the 1950 World Cup final, and the 7-1 loss to Germany in the 2014 semifinals, though? Both fully Elo-approved. Uruguay swiped 52 points and Germany 82. Both matches were Brazil’s newest darkest hours, both in Elo and in real life.

Another thing to note is that it is, in fact, possible for a country’s finest or darkest hour to also be their first. Teams are granted a starting amount of points that are then used to score their first match and send them on their way, and sometimes that first match ends up being a major calibrating moment. Most competitions using an Elo variant give everyone the same starting amount, but that would present some problems if, for example, a new nation declared independence from one sitting below the designated starting value. It’s not like the new country instantly became a better soccer nation by slicing itself away from a big chunk of the pre-existing talent pool. So every nation is given its own starting amount upon inauguration, always a multiple of 100.

For example, when England and Scotland played the first-ever international match in 1872, they were both granted an opening rating of 1800. Fast forward to July 10, 2012, when Sudan had a rating of 1474 as of the day newly-established breakoff South Sudan played its inaugural Elo-rated match against Uganda. It wouldn’t make any sense for South Sudan to be anywhere above 1474 to start out, let alone the 1800 England and Scotland had to work with (an amount that as I write this would put them on par with the likes of Croatia, Ukraine and the United States). Instead, South Sudan was given a starting value of 1300, which they’ve proceeded to bleed away with repeated early losses. Nations here were assigned what seems to be whatever opening rating appeared to make sense at the time. Mexico started at 1700. The United States and Canada opened against each other, starting at 1600 apiece. Lebanon started at 1500. Malaysia started at 1400. Yemen started at 1200. Belize started at 1000. Comoros started at 900. Mongolia started at 800. Tibet started at 600.

Elo is also capable of determining that a country’s finest and darkest hours come one right after the other. Bulgaria saw this happen in the 1994 World Cup. On June 21, they had their darkest hour, losing 3-0 to Nigeria in the opening match of the group stage and shedding 53 points. Nine days later, though, after having gained 47 of them right back beating Greece 4-0, the Bulgarians defeated Argentina 2-0 for 68 points and their finest hour… though actual Bulgarians would probably dispute that, figuring that the 2-1 victory over Germany in the quarterfinals (worth 45 points) was their actual finest hour. Hristo Stoichkov aficionados probably aren’t putting the Greece win over the Germany win.

The United States sees a similar scenario. Though actual Americans would probably put the 1950 win over England as their finest hour (56 points), Elo disagrees. Instead, it opted for the 2-0 win over then-top-ranked Spain in the 2009 Confederations Cup (64 points). Which when brought up would strike Americans as not exactly a bad choice either, as looking back it was certainly regarded as a big win, having snapped a 35-game unbeaten streak by the Spaniards. Barely a month later in the CONCACAF Gold Cup final, though, came America’s Elo-approved darkest hour, surrendering 76 points to Mexico in a 5-0 loss at Giants Stadium. Embarrassing, sure, but hardly America’s actual darkest hour, as anyone who suffered through the long decades in the pre-90’s wilderness will immediately tell you. Even after the match, American captain Brian Ching put the result in perspective, noting, “We get another crack at them in two weeks.” Ching was referring to the World Cup qualifier in Estadio Azteca just around the corner… which Mexico also won. But still.

So no. Elo’s not perfect at this either. It’ll make some mistakes, it’ll toss up some strange results sometimes. But what it does appear able to do is highlight points in a country’s history that at least merit a look beyond that which they were originally given. If you see a particularly large swing, it might be worth looking into (again, the chart of largest swings per country is here).

At the very least, it’ll be more worth looking into than FIFA’s three points for a win.


Tags: , , , , , , , ,

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: