worldcup Soccer

2015 Women’s World Cup: Group Play Has Almost No Impact On Betting Markets

They say the Masters doesn’t really start until the back nine on Sunday*, but that’s not really true at all. Between the opening round Thursday to midday Sunday, the betting odds wildly shift, with many pre-Augusta favorites falling entirely out of contention and longshots becoming much shorter.

*Ironically, this may be the least appropriate application of the “the beginning of the event/season doesn’t really matter much” axiom. The regular seasons of almost every American sport are far less significant in determining the eventual champion than the opening rounds of a golf tournament are.

A more apt phrase may be that the Women’s World Cup doesn’t really start until the knockout round. That stage of the tournament starts this weekend, bringing significantly higher leverage to each game in terms of determining the eventual champion.

Sure, there were many exciting moments in the World Cup group play round, but they were merely among teams with almost no chance to win the tournament. This is in large part due to the new format of this year’s event. In previous Women’s World Cups, there had been 16 teams divided into four pools of four. Because only 50 percent of teams passed to the eight-team knockout round, there was a much higher bar to clear during the group stage for serious contenders, resulting in elite teams facing elimination and substantial shifts in the betting odds between the start and conclusion of the group stage.

This year’s expansion of the tournament to 24 teams shook things up, though. With the teams still organized into groups of four (this time six groups), advancing two teams from each group would produce a 12-team bracket, which does not neatly divide by twos to produce one champion. This meant that, assuming the tournament was set with 24 teams and groups of four, FIFA had three options for arranging the knockout round:

  • Have the top two teams from each group advance and create a a 12-team bracket in which the four top teams receive byes
  • Have the top one team from each group advance as well as the two best second-place teams to produce an 8-team bracket
  • Have the top two teams from each group advance as well as the four best third-place teams to produce a 16-team bracket

The latter would produce the least exciting, fair structure for the pool play round, and unfortunately, that’s exactly the format FIFA decided to go with.

With the new structure, the tournament will spend 36 group stage games determining how to eliminate a mere eight teams (and the eight worst ones at that) and then only 15 knock-out round games determining how to eliminate the 15 best. That’s a games-to-eliminations ratio of 4.5-to-1 for the group round vs. 1-to-1 for the knockout round. (By comparison, the men’s tournament has a group stage games-to-eliminations ratio of 3-to-1, the same ratio in previous year’s Women’s World Cups.)

Because only the lowest of teams are eliminated in the group stage, and there was never really any threat for the true contenders to not clear such a low bar to reach the knockout round (some teams were over 99% likely to advance according to FiveThirtyEight’s model), the group play games seemed rather insignificant, and that’s exactly how it played out. Only one team that was expected to advance to the knockout round by FiveThirtyEight’s metrics did not, and by only a tenth of a WSPI point: Columbia (80.3 WSPI) over France (80.4 WSPI).

The betting markets also suggested the group stage—which represent 70 percent of the tournament’s total games—had extremely little impact on the championship picture (via OddsChecker):


Note: The win probabilities are derived from the implied percentages (i.e. 4-1 would convert to 20%) and then dividing them by the sum of the implied percentages on every team so that they sum to 100 percent.

Sweden and Germany are the only two teams who had greater than 2.5 percent changes in championship probability after the conclusion of the group stage. Sweden’s odds shifted largely due to its group stage struggles, which pitted it against top-ranked Germany in the first round of knockout play, while Germany’s boost came in large part due to its dominant performance in the group stage (which included a 10-0 thrashing of Ivory Coast). The Germans unsurprisingly advanced Saturday, 4-1.

What would have been a far better solution for structuring the new 24-team World Cup to increase market fluctuation during the group stage would have been to arrange either an 8-team knockout round or a 12-team one with byes for the four top group winners. Again, the group stage consists of over 70 percent of all the games played in the tournament, so putting more weight on those games should be of great importance for both the entertainment and fairness of the event. Such alternative structures would, of course, cut down on the number of games compared to the current, new format, but given that the tournament is already expanding this year, it would still be a growth compared to year’s past.