Figure Skating Scores Reveal Signs of Strategic National Favoritism

Looking beyond generous marks

In an arXiv preprint, University of California, Davis economists Jialin Liu and Shu Shen investigate whether national ties influence figure skating scores strategically, rather than simply through familiarity or shared artistic preferences. Their evidence concerns patterns across competitions, not proof of any individual judge’s intentions. The supplied source is a preprint, not a peer-reviewed publication.

The researchers assembled official scoring records covering 172 international competitions from 2016/17 through 2025/26. Their sample includes 23,254 performances and 620 judges. Published scoring protocols, panel lists and skater biographies let them connect individual marks with federation ties, performance order and competitive circumstances.

Figure skating provides an unusually revealing setting: judges evaluate the same performance, scores become public after each skate, and some start orders and judging panels are randomly drawn.

Comparing judges watching the same skate

The authors constructed a synthetic score for each judge: the total a performance would receive if every judge awarded that person’s marks. They normalized scores to compare different disciplines, segments and scoring eras.

Their main regressions compare judges evaluating the same performance while accounting for each judge’s general generosity. This helps separate national ties from differences in the performance itself.

Across the sample, a judge awarded a skater from the same federation a premium averaging about 1.73 percentage points of the study’s reference score scale. That is not a 1.73-point increase in the official result: an individual judge’s marks enter a panel calculation. The premium was particularly large in ice dance and in more interpretive scoring categories.

A favorable gap alone cannot establish strategic intent. A judge could sincerely appreciate a compatriot’s style or know their abilities better. The stronger evidence comes from when the gaps change.

What random start orders reveal

In first segments with an open draw, starting position is randomly assigned. Here, the estimated compatriot premium declined from 1.99 percentage points of the reference scale for the first starter to 1.29 for the last.

The authors interpret this as consistent with strategic scoring under uncertainty. Early in a competition, a judge has seen less of the field and cannot know how much extra credit might affect placement. Later, the competitive situation is clearer.

Random assignment strengthens the evidence that starting position changes the premium. The explanation involving strategic calculation remains an interpretation, however. Under the authors’ assumption that taste and familiarity do not depend on random starting position, at least about a fifth of the average premium requires another explanation.

Penalties that follow competitive stakes

The study also finds lower marks for compatriots’ close rivals. Instead of defining rivals using final results—which judging itself can influence—the authors use skaters from different federations starting within three positions of one another.

This is most informative in second segments, where starting order follows first-segment results and therefore groups similarly placed competitors. There, judges marked down close rivals skating before their compatriot, and those skating afterward while the compatriot remained in contention. Once that compatriot was out of contention, the estimated total-score penalty on later rivals was close to zero.

Penalties also grew as the compatriot’s second-segment starting position moved later, indicating stronger placement prospects. These patterns support a strategic explanation more directly than a simple average scoring gap.

Panel composition mattered too, but not uniformly over time. In the 2018/19–2021/22 cycle, compatriot premiums rose when later rivals had their own federation’s judge on the panel. The corresponding escalation evidence was absent in the 2022/23–2025/26 cycle. Separating responses to judges’ reputations from responses to their mere presence proved difficult.

Close results, conditional simulations

To explore consequences, the authors simulated admissible judging-panel redraws for the 2026 Milan Olympics. With discipline-specific estimates of scoring biases, the ice dance champion changed in roughly 45–57% of redraws and the women’s champion in 20–27%, depending on assumptions about honest disagreement.

Those figures are not probabilities that the official winners were incorrect. Honest disagreement alone also changed simulated winners in close events. The exercise depends on estimated scoring behavior, assumptions about incoming judges and an approximation that trims judges’ totals rather than each element and component separately.

Other limitations matter. Starting proximity is an imperfect proxy for rivalry. Cross-panel comparisons are more vulnerable to unmeasured performance differences, and several placement estimates are imprecise. Individual judge estimates can also be unstable.

The authors recommend excluding leading contenders’ federations from judging panels where feasible and monitoring rival penalties alongside compatriot premiums. Their central finding is narrower than an accusation: scoring differences track information and competitive stakes in ways that shared taste alone struggles to explain.