The community prediction on Metaculus for a related question ("Will Russia annex any part of any Baltic country by 2035?") is currently 4% (https://www.metaculus.com/questions/8786/russian-annexation-in-the-baltics-before-2035/).

@Liedholm This one resolves also by rocket attack, without land invasion.

@Irigi Yes, the questions aren't identical! I just wanted to highlight the similarities, in case people find it useful for assigning their subjective probabilities.

@Liedholm Also, according to this website (https://calibration.city/), Metaculus is calibrated worse than conventional prediction markets, and questions in the politics category with predicted 0-5% probability resolve true 0.8% of the time, and the ones with predicted 5-10% probability resolve true 1.5% of the time on Metaculus, so the current 5% translates to roughly 1% actual probability.

@StepanBakhmarin Fascinating, thanks for sharing!

The fact that Metaculus users seem to make unusually "humble"/under-confident predictions, paired with what seems to be superior performance on Brier scores when compared to Manifold (see e.g. the plot of "Brier Score from Midpoint Probability" vs. "Market Duration (percent)" here: https://calibration.city/accuracy), makes me slightly more suspicious of the high probability estimate for this question on Manifold (currently at 24%).

What if its Belarus making an attack. Does the proxy relationtiop between russia and belarus affect this market?