The simulator has two distinct calculations. The voting engine chooses a ranking for each voter and allocates points. The Entry Strength page compares the ratings you entered. The second calculation is a convenient summary; it is not an estimate of the probability that the first calculation will produce a particular winner.
A vote starts with eligible entries
For each voter, the engine excludes that voter's own country. A custom named jury with no country association can rank every participating entry. Jury and public rankings are generated separately when both vote types are active. A voting group's enabled round and vote types determine whether it participates in that stage.
In the basic model, the engine repeatedly selects one candidate from a weighted pool, removes it, and continues until it has up to ten recipients. Removal prevents the same entry receiving two awards from one voter in the same round. If fewer than ten entries are eligible, the ranking is shorter.
Larger ratings mean more weight
The basic weight begins as the rating raised to the power of 1.6. For the first five ranking positions, ratings of at least 300, 500, or 900 receive additional boosts of 1.3, 1.8, or 2.5 respectively. Mid-range ratings receive a smaller boost in later positions. A random multiplier between 0.85 and 1.15 varies the weight, and a weighted random draw then selects the recipient.
This makes high-rated entries more likely to appear high in each voter's ranking without fixing the winner in advance. The process repeats for every voter and for each enabled vote type. It models an entertaining fictional competition, not measured real-world voter behavior.
Rankings become points
Standard scoring awards 12, 10, 8, 7, 6, 5, 4, 3, 2, 1 to the first ten recipients. When only three entries are eligible, those recipients receive 12, 10, and 8; the unused awards are not redistributed. Jury-only or public-only mode omits the other round.
The app's UMK-style classic scoring multiplies public awards by three. With equal jury and public voter coverage, that supplies a 25/75 balance. Custom groups enabled for only one vote type can change the effective balance. The app's separate Melodifestivalen-style workflow has its own qualification and round handling; treat it as a selectable simulation format rather than a guarantee of exact current broadcaster rules.
Totals are the sum of jury and public points. The simulation, live scoreboard, and full results use the same tie order: public points, then combined jury/public 12-point awards, then the country name alphabetically. In single-round voting, the inactive vote type has zero points. The alphabetical fallback is an app convention and does not implement every official tie-break criterion.
What complex mode adds
Complex mode uses a separate ranking model incorporating genre preferences and regional relationships. Entries have genres, and voters have jury and public taste profiles. These influence the relative appeal of the candidates. As a result, two entries with equal headline ratings can perform differently against a particular voting field.
The preferences and regional relationships are modeling assumptions. They should not be interpreted as factual statements about an individual voter or a prediction of a country's official points.
Reading Entry Strength correctly
The comparison page raises jury and public ratings to 2.2, normalizes each set across the full field, and combines the resulting shares. It uses a nominal 50/50 balance for standard combined voting, 25/75 for the UMK-style combined setting, or the active round alone for single-round voting.
For example, a displayed 20% relative strength means one fifth of that normalized comparison, not a one-in-five chance of winning. The calculation does not run the vote engine, model semi-final qualification, or account for all custom groups and complex-mode preferences. Normalized shares add up to 100% across the full field, while only the top ten are displayed; displayed rounding can introduce small differences.
For a points-level example, see reading a four-entry scoreboard. For the full sequence, use the first-contest guide.