Last month, I introduced you to my Eurovision prediction model, utilizing fan predictions and historical voting patterns to simulate the contest. Back then, the model predicted a Swedish victory, with Austria and France right behind them. One month, four pre-parties, 37 video snippets and 481 rehearsal pictures later, it’s time to look at the model once again, and see how the pre-season has shaped each entry’s chances.
To do this, I once again ran a survey, to collect fans’ predictions of each entry’s potential across juries and televote. In this regard, it is once again worth underlining, that this model is ultimately a reflection of fan predictions, and not some universal truth. While fans have gotten more information to judge the entries on since last time, it should also be noted that different entries have seen different parts showcased. Some rehearsal clips display the song’s big moment while others save excitement for the live shows. In addition, due to the shorter survey time frame, this model only relies on 146 predictions, instead of the 289 of the last survey. Now let’s take a closer look at who made the biggest moves between the two surveys
The biggest movers
On one hand, we see a set of entries that seem to have massively improved their chances. The clearest example here is Cyprus, seeing the biggest strengthening of both televote and jury potential. Also we see big improvements for Slovenia, Serbia and Latvia, who both look better equipped to challenge for qualification with improved televotes. Also in the green is Switzerland, whose simplistic staging seems to have been well received, perhaps even elevating them to a dark horse for the win.
On the other hand, a set of songs have seen slight declines in expectations. The biggest drop here is San Marino, almost exclusively down to a decline in televote, which could be worrying for qualification. Portugal, Azerbaijan, and Montenegro, who were both already looking like unlikely qualifiers have also seen a drop. At the other end of the scale, one of our dark horses from the last survey, Netherlands, have also seen declining expectations, perhaps even taking them out of win contention.
Let’s now take a closer look at how the changes in expectation have changed the model’s predictions for qualification and win chances.
Semi Final One
Looking at semi final one, we see a big set of probable qualifiers, with Sweden, Albania, Estonia, Poland, Ukraine, Netherlands, and Norway all still in the 90%+ range, and Cyprus almost joining their ranks due to their huge improvements. While Belgium seems to be moving closer to the safe category, San Marino has seen a major drop, being right on the edge of qualification, with Iceland staring them down. Croatia, Slovenia, and Portugal all look like likely non-qualifiers, but each with noticeable paths to the final, while Azerbaijan seems to have an uphill battle ahead of them.
Semi Final Two
Turning to semi final 2, we once again see a much more open qualification race. While the first semi final only had two songs in the 30-70% range, this semi final has seven. Finland, Israel, Austria, Malta, and Australia all seem quite safe, while Montenegro and particularly Georgia seem very challenged. The remaining songs all seem to be in play, with Lithuania, Czechia, Luxembourg and the improved Latvia leaning towards qualification and Ireland and Denmark tilting towards non-qualification. The remaining entries, Greece, Armenia, and Serbia all look close to pure coin flips.
Final/win contenders
Finally, looking across the 565 simulations that featured the exact 20 qualifiers projected as the most likely, these were the average results. Here we see Sweden pulling ahead enough with the televote to overcome the jury advantages of Austria and France. It is worth noting that these numbers represent averages, giving them a more smoothed out nature, than how most simulations actually look, with the winner usually being in the 480 range in terms of points.
Lastly, let’s look at our potential winners. Starting with our dark horses, Finland, Albania and Israel all seem to have defended their position and kept themselves in contention, with Finland’s chances slightly increasing and Israel’s slightly decreasing. Meanwhile, the changes in expectation have dropped Ukraine and Netherlands below the 1% mark, while Switzerland has seen a big rise.
Finally, looking at our main contenders, the contest still looks like a three horse race between Sweden, Austria and France. Once again, due to a unique tele-vote strength, Sweden differentiates itself from the two others, coming out as the models most likely winner, at a 37.7% win chance. In fact, it seems to have increased its lead, with both of its jury-reliant contenders, Austria and France seeing small drops
However, overall the contest still looks open, with no one entry overwhelmingly predicted to win. And again, it is worth noting that these numbers are purely predictions, reflective of fan intuition, and could very well get something wrong. In 2024, this model put Dons’ qualification chance at 8.5%, and he still managed to comfortably qualify.
In the end, there’s only one way we can know the contest’s results – by watching it. Happy Eurovision week!