With Team Wales’ Road to TBC now complete, and France taking a victory over Sweden for 1st place, we’ve now seen 13 of the potential competitors at the Roller Derby World Cup on track in Europe, between this tournament and the earlier “European Roller Derby Tournament“.
What’s quite interesting about comparing ERDT and Road to TBC is just how much the balance of power in some of the teams common between them has changed in a relatively short space of time; and how different some of the newer teams may be to their 2014 versions.
Looking at returning teams first: at ERDT, Wales and Belgium took 3rd and 2nd, but looked very close in strength indeed. We predicted that home advantage would make Wales likely to win against the Belgian team at Road to TBC… but didn’t predict by just how much! In fact, Wales did take our expected 3rd place, with an impressive performance; but Belgium was likely disappointed by their 7th place position by comparison.
France and Sweden, as predicted, duked it out for 1st and 2nd, and France emerged victorious. Sweden’s performance, whilst easily good enough to win against everyone else [despite some opening issues with both Wales and Ireland], wasn’t quite as commanding as their 2014 iteration managed at the World Cup – however, this was also their first tournament, after a relatively recent selection, so this may just be a reflection of the stage of their training so far, versus a more experienced France.
We tipped Ireland for a strong placement, purely on the basis of their 2014 performance, and it seems we were right – despite their team being as new as Sweden’s, Ireland managed to perform creditably and take a convincing 5th place position from Scotland…
…whilst Spain took the 4th place to Wales that we had thought destined for Belgium.
The case of Spain and Scotland is an unusual one, with their bout the closest in the tournament, but having the result of essentially forcing one to take 4th and one 6th place (as Ireland and Wales would likely have taken their eventual positions against either team). This is an issue with this kind of group selection process, where the seeding is designed for “exagony” (the desire for teams to play teams they’ve not played before), rather than technical ranking.
Ironically, Scotland actually performed better than Spain against France, and the two performed almost identically well against West Indies…
On principle, then, we performed the same kind of Bayesian inference on the Road to TBC that we did on ERDT, to determine what the most likely underlying strengths of the teams were.
(As an aside, one key difference from this is that the same method tells us that the performances at Road to TBC were much less consistent than they were at ERDT – the inference needs to assume around twice as much variation in performance for a given team in order to make predictions. This may be a reflection of the relative newness of some of the competing teams.)
That said, the inferred strength distributions (relative to France) are below (this is a slideshow, with different teams highlighted for ease of identification).
As can be seen, it’s fairly obvious that Sweden was most likely the 2nd strongest team in the tournament, and Wales the 3rd. What’s more strikingly different from the tournament result is that the most consistent interpretation of the scores is that Ireland were the 4th strongest team, with Scotland, Belgium and Spain all quite close together in 5/6/7th positions (but with enough overlaps that upsets – like Spain/Scotland – were not unlikely).
Given the way in which the teams have changed in strength even since ERDT, however (and how some of the rosters were selected for Road to TBC), we would be hard pressed to predict performance at the next Roller Derby World Cup across these two tournaments – other than to say that Sweden will probably get stronger, and none of matchups below 3rd place might happen the same way in just a few months time…