A good result for the Moldovan saxophone brides and the Azerbaijani blackboard goth/horse combo, but sadly, that’s the last we’ll be seeing of Montenegro’s Dothraki strip-o-gram:
The model predictions were 7/10 in the first semi-final. The top five were right, but Georgia, Iceland and Albania missed out, with Portugal, Poland and Moldova qualifying instead. Not a perfect result, but well within what should be expected from a probabilistic forecast. In simulations, the average number of correct predictions was 6.8, so nothing to worry about there.
Updated final predictions
I ragged on Portugal a little bit in the previous post, but they’re now even more favoured by gamblers (28% probability on Betfair). On the other hand, the model has upgraded them to a full 1.2% chance. This is probably underrating them, but the act is exactly the sort of thing which betting markets have overrated in the past (see France’s 2011 entry, which was favoured at around the same level).
Italy are still the bookies’ favourite, although slightly less so after Portugal’s rise. The model opinion hasn’t changed - they’re strong, but not as structurally advantaged as some of the other big names.
In general though, the first semifinal played out as predicted, and there aren’t any big shifts in probability.
This is definitely the weaker of the two semifinals. Serbia, Romania, Denmark and Norway are all traditionally fairly strong, but not in the first rank of Eurovision nations. Originally, Russia had been slated to perform here, but their withdrawal has shifted things a bit. Somewhat predictably, the only country which is hurt by Russia’s withdrawal is Belarus. They’re usually a recipient of votes from their larger neighbour, but rarely strong enough to do well elsewhere. They’ll suffer a bit, but it’s cancelled out somewhat by not having to compete with Russia for a qualification spot. For everyone else, the lack of Russia just means one more spot to go around.
Serbia are strongest here (76% chance of qualifying), followed by Norway (75%), Denmark (73%) and Romania (66%). Beyond that, it’s really anybody’s game. Austria (62%) and Hungary (62%) are marginally stronger than the rest of the field, and if we have to make a prediction, then it’s Lithuania (60%), Malta (59%), Netherlands (59%) and Israel(57%). It really is quite open though, and the only countries which would be even remotely surprising are San Marino (28%) and maybe Switzerland (37%).