FIFA's Elo ranking measures Portugal more fairly than its critics claim

Key takeaways

• The FIFA men's world ranking now works by exchanging points between the two teams in each match, not by averaging results over several years.
• Under the current system, beating a much weaker side earns little, while losing to one costs a great deal, which rewards consistency over volume.
• Competitive matches count for more than friendlies, so a European side's ranking is driven mainly by qualifiers, the Nations League and tournaments.
• The strongest objection is that teams from different confederations rarely meet, so comparing ratings across continents rests on relatively few matches.

The ranking now rewards the right things

Interest in the FIFA world ranking comes round with every international window, and Portugal is no exception: search activity tracked through Google in Portugal shows people there looking up both the ranking and the national side. That curiosity usually comes with some suspicion. Supporters everywhere suspect the table is arbitrary, that it can be gamed, or that it reflects reputation more than results.

My argument is that this suspicion is out of date. The method FIFA uses now is a version of the Elo system, which was first developed to rate chess players. It replaced an older formula that averaged weighted results over a multi-year window. The older approach had well-known faults. It rewarded teams for simply not playing at certain times, and it could be pushed around by careful choices of friendly opponents.

The current model starts from a simpler idea. Before every match, each team holds a rating. The match is treated as a test of that rating. If a team does better than its rating predicted, it gains points. If it does worse, it loses points. Whatever one team gains, its opponent loses. Over a long run of matches, a team's rating settles at roughly the level its results support.

For a side like Portugal, which plays almost all of its competitive football against European opposition under UEFA's structures, this gives a fair answer to the question the ranking is meant to address: how strong has this team shown itself to be, relative to everyone else, on recent evidence? The three mechanisms below support that claim, and a serious objection follows.

Every result is now judged against an expectation

The core of the Elo approach is the expected result. Before a match, the gap between the two teams' ratings is turned into a probability-like figure for how each side should do. A strong favourite is expected to win. A match between evenly rated teams is expected to be close to a draw.

After the match, the actual result is compared with that expectation. A win counts as a full result, a draw as half, and a defeat as nothing. FIFA's regulations also give a separate value for matches settled by penalty shoot-out, which sits between a draw and a win for the side that prevails. Points change hands according to how far the actual result departed from the expected one.

The consequences matter for a team that is usually favoured. When a highly rated side beats a much lower-rated opponent, it gains very little, because the system already expected that. When it draws with or loses to such an opponent, it drops noticeably. When it beats another highly rated team, it gains much more.

This is exactly how a fair measure of strength should behave. It stops a team climbing the table by piling up comfortable wins against weak opposition, and it makes a slip against a modest side cost something real. Under the old averaging method, a series of routine victories could inflate a team's standing. Under the current method, those victories mainly confirm a rating the team already holds.

For Portugal, as for any European side with a strong recent history, a high ranking can only be kept by beating the teams it is expected to beat and by competing well against its peers. A good reputation is not enough on its own, which is the point.

Match importance separates friendlies from competition

The second mechanism is a weighting for the importance of each match. Not every international fixture moves the ranking by the same amount. FIFA's rules apply a coefficient by match type. Friendlies carry the least weight. Nations League matches, World Cup and continental qualifiers, and matches at final tournaments carry progressively more, with the later stages of the World Cup finals at the top.

This weighting answers one of the oldest complaints about national-team rankings, which is that friendlies distort them. Friendlies are often used to try out players, give squad members minutes and test systems, and the result is secondary. A ranking that treated them like a qualifier would be measuring something other than competitive strength.

For European nations, the calendar now makes this mechanism especially effective. Since UEFA brought in the Nations League, many dates that would once have been friendlies are now competitive fixtures, organised in leagues of teams at broadly similar levels. A side in the top division plays a steady run of matches against strong opponents. Those matches carry more weight than friendlies, and the Elo calculation treats them as demanding tests.

So the European schedule feeds the ranking a regular supply of meaningful, weighted results between closely matched teams. For Portugal, this means its rating is determined mostly by matches that both sides were trying to win. That is the right input for a strength measure.

Points are exchanged, not accumulated

The third mechanism is less often discussed but may matter most. Because the system is built on exchange, the total number of points in circulation stays broadly stable. One team's gain is another's loss. Nobody can create ranking points out of nothing by playing more matches.

This closes off the gaming that damaged the older formula. Under averaging, the number and timing of matches could change a team's figure in ways unrelated to how good it was. Associations were known to think tactically about whether a friendly would help or hurt their average. Under the exchange model, playing an extra match is just another test. A team that is as good as its rating suggests should, on average, neither gain nor lose from it.

Fairness also needs the system to be transparent, and the exchange model is open to inspection. FIFA publishes the formula and the coefficients in its ranking procedures, so anyone with a team's results can rebuild the calculation. When Portugal's position moves after a window, the cause can be traced to particular results against particular opponents. A ranking whose movements can be explained match by match is much harder to dismiss as arbitrary.

FIFA's regulations also include a protection for final tournaments: a team eliminated in a knockout round does not lose points for that elimination. The aim is not to punish teams for reaching stages where they must face the strongest opposition. This is a deliberate design choice. One can argue about it, but its purpose and effect are clear.

The strongest case against: confederations rarely meet

The best argument against my position deserves a full hearing. It is not that the formula is wrong within a group of teams that play each other often. It is that the ranking claims to compare all national teams worldwide, while the matches that would make such comparisons reliable are rare.

Nearly all of Portugal's competitive football is played against European opponents. The same is broadly true of South American, African, Asian, North and Central American and Oceanian teams within their own confederations. Inter-confederation matches happen mainly at World Cup finals and in friendlies, and friendlies carry the least weight. An Elo system can only measure the gap between two pools of teams through the matches that link them. If those links are few and lightly weighted, how Europe as a whole compares with another confederation rests on thin evidence.

Critics add several design choices that weaken the measure further. The formula does not account for home advantage, so a win at home and a win away count the same despite the very different difficulty. It ignores margin of victory, which throws away information about how dominant a result was. And because seeding for qualifying draws and final tournaments is often based on the ranking, a high position can make a team's path easier, which in turn helps protect that position.

Each of these points is fair. Taken together, they suggest the ranking is more reliable for ordering teams within a confederation than across confederations. That is a real limit on what it can claim to show.

My answer is that these objections mainly affect global comparisons, not the measurement of a European side against its regular opponents. Within Europe, where Portugal plays almost all its competitive matches, the links between teams are dense and the weighting favours meaningful fixtures. The criticism narrows the ranking's authority but does not make it misleading about the teams it observes most closely.

Evidence that would change this conclusion

Several kinds of evidence could reasonably overturn this view.

The first is a systematic comparison of predictions. If independent analysis showed that the FIFA ranking predicts the results of competitive international matches clearly worse than alternative rating systems that include home advantage and goal margin, that would suggest FIFA's simplifications cost real accuracy. The public World Football Elo ratings, run independently of FIFA, offer an obvious point of comparison.

The second is evidence of a persistent bias between confederations. If teams from one confederation regularly did much better or worse in inter-confederation tournament matches than their ratings predicted, the cross-continental calibration would be shown to be faulty in a way that matters.

The third is evidence that the seeding feedback loop is strong enough to lock in positions. If teams that just made the higher pots kept their ratings mainly because of easier draws, and not through results against strong opponents, the ranking would be partly measuring its own effects.

Finally, changes to the international calendar could shift the balance. If competitive fixtures between strong European sides became less frequent, the main strength of the current model for a team like Portugal would weaken.

None of these is established by anything in the material behind this article. Until such evidence appears, the reasonable view is that the ranking does a creditable job of what it sets out to do.

Sources and further reading

• FIFA, published regulations and procedures for the men's world ranking, setting out the formula and match-importance coefficients
• UEFA, competition regulations for the Nations League and European qualifying, explaining how the competitive calendar is structured
• Google search trend data for Portugal, the signal that prompted this article
• World Football Elo Ratings, an independent rating system useful for comparison with FIFA's method