AI summary · arXiv: 2611.13194v1 · econ.GN

Taste or Strategy? Nationalistic Bias and Competitor Penalties in Figure Skating Judging

Jialin Liu, Shu Shen
Original paper Submitted to arXiv on 09 Nov. 2026 License: CC BY 4.0 Also in: fr

How to read this summary? Illustrated guide

The essentials

Figure-skating judges show strategic nationalistic scoring: they favor compatriots and target close rivals, and panel composition can alter close competition outcomes.

Question

The paper asks whether judges’ compatriot premiums reflect strategic favoritism rather than shared taste or familiarity, and whether they penalize compatriots’ rivals. It also examines whether these scoring patterns and judge-panel composition affect competition results.

Method

The authors analyze official scores and judge assignments from 172 competitions across ten seasons, covering 23,254 performances and 206,983 performance-judge pairs. They use within-performance comparisons, random start-order and panel-draw variation, and regressions with performance and judge fixed effects; they also simulate admissible panel redraws for the 2026 Milan Olympics.

Findings

  • Result Judges give compatriots higher marks: the estimated average premium is 1.727 percent of the scale for total segment score, 2.644 for grade of execution, and 1.816 for program components.
    See the passage in the paper
    On average, a judge marks a skater from her own federation 1.73 percent of the scale higher on TSS than the other judges on the same panel, 2.64 and 1.82 percent higher for GOE and PCS respectively.
  • Authors’ interpretation The compatriot premium declines with start position even in open-draw first segments, which the authors interpret as evidence that at least part of it is strategic rather than attributable to taste or familiarity.
    See the passage in the paper
    At least about a fifth of the average compatriot premium, therefore, needs another explanation.
  • Result Judges penalize close rivals selectively: the penalty is absent on total segment score when a later-starting rival’s compatriot is already out of contention, and it increases as the compatriot’s start position signals better placement chances.
    See the passage in the paper
    For Segment 2, however, both the penalty on a rival who starts before the compatriot and the penalty on a rival who starts after a compatriot still in contention grow, in absolute value, with the compatriot’s start position.
  • Result In the Beijing-cycle analysis, the compatriot premium increases when later close rivals have compatriot judges on the panel, and increases further with those judges’ estimated prior compatriot premiums.
    See the passage in the paper
    The compatriot premium rises with the trained premia of the later rivals.
  • Result Counterfactual panel redraws change the champion in 45–57% of ice-dance cases and 20–27% of women’s cases, depending on how score uncertainty is modeled; the authors say trimming provides little protection against a typical judge’s bias.
    See the passage in the paper
    The counterfactual analysis shows that, in ice dance, the top placement changes in 45-57% of the 90,090 admissible pairs of panels, with the exact number depending on how the shock is handled.

Limitations

  • The start-order-based definition of close rivals is noisy, so estimated competitor penalties are biased toward zero and are lower bounds on the true penalty.
    See the passage in the paper
    The cost is that the start order identifies a compatriot’s competitors with noise and biases the estimated penalty toward zero.
  • The counterfactual results are specific to the 2026 Milan Olympic events and rely on modeled counterfactual scores, including assumptions about baseline scores and random shocks.
    See the passage in the paper
    The baseline $\nu_{ijp}$ is drawn from a normal distribution with the mean calculated from the average judge scores of the realized panel net of the three terms, excluding those from top-three’s federations (i.e., France, the United States and Canada), and the variance $\sigma^{2}$ being either zero or from within-panel estimates.
  • Judge-specific premium estimates are unstable for judges with limited observations, and the authors caution against using them to identify or accuse individual judges.
    See the passage in the paper
    To properly identify the most nationalistic judges in the data, proper statistical methods for ranking and selection such as Armstrong et al. (2022); Gu and Koenker (2023); Mogstad et al. (2024); Andrews et al. (2024), among others, need to be adopted.

Why it matters

The authors argue that institutional variation in start order and panel composition helps distinguish strategic favoritism from non-intentional explanations in subjective judging. They recommend monitoring judges’ scoring patterns and, when feasible, keeping leading skaters’ federations off judging panels or seating contending federations symmetrically, rather than relying on trimmed means or replacing judges with AI.

Written by AI from the paper. Passages are checked word for word against the text we analysed; the rest may contain errors.

AI-generated keywords: figure skating; judging bias; compatriot favoritism; competition outcomes; panel composition

Read further

  • Comprehensive Summary
  • Layman's Summary
  • Blog article
  • Original abstract
The paper examines whether figure-skating judges give higher marks to athletes from their own federation because of shared taste or familiarity, or whether they strategically favor compatriots and disadvantage close rivals. It also investigates whether judging-panel composition affects scores and competition outcomes. The authors compile official protocols and judge-panel records from major international competitions held between 2016/17 and 2025/26. The dataset covers 172 competitions, 23,254 performances, and more than 206,000 judge-performance observations. Marks are expressed as percentages of comparable scoring scales. To estimate differences between judges assessing the same performance, the authors use performance and judge fixed effects and exploit variation in start order and randomly drawn judging panels. The authors define close rivals as skaters from another federation who start within three positions of a compatriot. This measure does not rely on final results, but it is imperfect. Across the full sample, judges score compatriots 1.73 percentage points higher in synthetic total segment score than other judges on the same panel. The estimated premiums are 2.64 points for grades of execution (GOE) and 1.82 points for program components (PCS). These associations alone do not establish intent. Evidence based on start order is more suggestive of strategic behavior. In first segments with a randomly drawn open start order, the compatriot premium falls by about 0.70 points from the first starter to the last. The authors argue that a preference based only on taste or familiarity should not systematically vary with random start position. In this subsample, they estimate a premium of 1.99 points for the first starter and 1.29 points for the last. Under the assumption that taste and familiarity do not depend on start order, at least about one fifth of the average premium requires another explanation. The authors also find penalties against close rivals. In the full sample, judges score a rival 0.12 points lower in total score, 0.26 points lower in GOE, and 0.08 points lower in PCS. In the second segment, where start order reflects earlier standings, the total-score penalty is 0.20 points. The pattern is targeted: judges penalize rivals who skate before their compatriot and those who skate afterward while the compatriot remains in contention. The total-score penalty is effectively absent when the compatriot has already fallen out of contention. The penalty also increases when the compatriot starts later and has a better chance of placing well. Because a rival may face several judges from competing federations, the combined panel effect could be much larger than the estimated effect of a single judge. Panel composition also matters, although the evidence varies by period. In the Beijing-cycle analysis, the compatriot premium increases when more of a compatriot’s later-starting rivals have a judge from their own federation on the panel. It also increases with those judges’ estimated historical compatriot premiums. The authors do not find this pattern in the later Milan-cycle analysis. Estimates of a colleague-compatriot premium weaken substantially after controlling for athlete and program characteristics and are not consistently significant. The paper also finds a premium associated with a compatriot technical controller, but not a consistent premium associated with technical specialists. Counterfactual redraws of the 2026 Milan Olympic panels suggest that judging composition could affect close results. Depending on how honest judging disagreement is modeled, the ice-dance champion changes in 45–57% of redraws and the women’s champion in 20–27%. However, honest disagreement alone can also change champions, and these counterfactual results depend on modeling assumptions. The authors further find that trimming nine judges’ scores removes little of a typical bias: about 97% of a median estimated premium remains.

Was this synthesis useful?

Score: 0

Why do we need votes?

Votes are used to determine whether we need to re-run our summarizing tools. If the count reaches -10, our tools can be restarted.

Figures and plots

Ask about a figure by number or use its question button to request visual analysis. Up to two figures and four image panels per question. Values read from plots are approximate; check the original data.

AI assistant instructions?

Tools for this paper

Explain it differently 1 free rewrite left today
Save to a collection Group papers in collections to compare them, chat with them and export them. Continue with Google Create a free account Log in

Advanced tools Premium

For researchers who follow a field closely: citation alerts, journal-club slides, version comparison and Zotero export. Discover Premium

Zotero Zotero guide
Citation alerts Citation alerts guide
Weekly letter Presented in "Economics and finance this week" of 16 November 2026. Economics and finance this week: When incentives change the score · Receive one by email for free

Papers around this one

See them on a map

Ranked by how close their title and abstract are to this paper’s. Select some to read them together.

Closest papers

  1. Related Broader context
    Biased forecasts and voting
    Jacopo Bizzotto, Davide Cipullo, André Reslow · 2026 · The Scandinavian Journal of Economics · Not summarized yet

    “Biased forecasts and voting” models strategic distortion by forecasters seeking to influence voters, while this paper studies judges’ strategic favoritism in skating scores.

    Abstract

    We introduce macroeconomic forecasters as strategic agents who may bias their reports to influence their audience. We test the motives of forecasters using monthly forecaster‐level data around the Brexit referendum. To guide our analysis, we model the way in which a referendum voter forms expectations based on the reports of a politically motivated forecaster. The forecaster trades off its reputation for honesty against the opportunity to influence the vote. The model predicts that the forecaster lies selectively and continues to do so even after the referendum. Our analysis finds that forecasters exposed to financial losses from Brexit issued significantly less accurate GDP growth forecasts…

  2. Related Broader context
    The Price of Diversity
    Hari Bandi, Dimitris Bertsimas · 2021 · Computers and Society · preprint, not peer reviewed · AI summary available

    “The Price of Diversity” develops methods to reduce selection bias while preserving merit, whereas this paper examines nationalistic bias in expert judging.

    Abstract

    Systemic bias with respect to gender, race and ethnicity, often unconscious, is prevalent in datasets involving choices among individuals. Consequently, society has found it challenging to alleviate bias and achieve diversity in a way that maintains meritocracy in such settings. We propose (a) a novel optimization approach based on optimally flipping outcome labels and training classification models simultaneously to discover changes to be made in the selection process so as to achieve diversity without significantly affecting meritocracy, and (b) a novel implementation tool employing optimal classification trees to provide insights on which attributes of individuals lead to flipping of thei…

  3. Related Broader context
    Inherent Trade-Offs in the Fair Determination of Risk Scores
    Jon Kleinberg, Sendhil Mullainathan, Manish Raghavan · 2016 · Machine Learning · cited 2,257 times · AI summary available

    “Inherent Trade-Offs in the Fair Determination of Risk Scores” proves conflicts among algorithmic fairness conditions; this paper studies favoritism and rival penalties in human judging.

    Abstract

    Recent discussion in the public sphere about algorithmic classification has involved tension between competing notions of what it means for a probabilistic classification to be fair to different groups. We formalize three fairness conditions that lie at the heart of these debates, and we prove that except in highly constrained special cases, there is no method that can satisfy these three conditions simultaneously. Moreover, even satisfying all three conditions approximately requires that the data lie in an approximate version of one of the constrained special cases identified by our theorem. These results suggest some of the ways in which key notions of fairness are incompatible with each o…

  4. Related Broader context
    A Statistical Model of Serve Return Impact Patterns in Professional Tennis
    Stephanie A. Kovalchik, Jim Albert · 2022 · Methodology · cited 4 times · AI summary available

    “A Statistical Model of Serve Return Impact Patterns in Professional Tennis” models tennis tracking data, while this paper uses skating scores to test for judging bias.

    Abstract

    The spread in the use of tracking systems in sport has made fine-grained spatiotemporal analysis a primary focus of an emerging sports analytics industry. Recently publicized tracking data for men's professional tennis allows for the first detailed spatial analysis of return impact. Mixture models are an appealing model-based framework for spatial analysis in sport, where latent variable discovery is often of primary interest. Although finite mixture models have the advantages of interpretability and scalability, most implementations assume standard parametric distributions for outcomes conditioned on latent variables. In this paper, we present a more flexible alternative that allows the lat…

  5. Related Broader context
    How to Find a Needle in a Haystack
    Theodora Helimäki, Åsa von Schoultz · 2024 · Political Behaviour in Contemporary Finland · Not summarized yet

    “How to Find a Needle in a Haystack” examines voter evaluations and candidate-choice heuristics, while this paper studies judges’ nationalistic scoring behavior.

    Abstract

    Candidate choices are characterized by a more personal and intricate decision-making process compared to party choices. In this more complex choice, not only substantive evaluations but also perceived competences and the personal characteristics of candidates can affect voters’ choice of candidates. In this chapter, we present voters subjective evaluation of what matters for their choice of candidate and how these have developed over the last five latest parliamentary elections. We then move on to explore which type of voters are prone to use certain types of heuristics in their decision-making process. Here, we use register data on candidates in combination with stated candidate choice in t…

  6. Related Broader context
    A Spectral Approach for the Dynamic Bradley-Terry Model
    Xin-Yu Tian, Jian Shi, Xiaotong Shen et al. · 2023 · Methodology · cited 8 times · Not summarized yet

    “A Spectral Approach for the Dynamic Bradley-Terry Model” estimates time-varying sports rankings from pairwise comparisons, while this paper tests bias in figure-skating scores.

    Abstract

    The dynamic ranking, due to its increasing importance in many applications, is becoming crucial, especially with the collection of voluminous time-dependent data. One such application is sports statistics, where dynamic ranking aids in forecasting the performance of competitive teams, drawing on historical and current data. Despite its usefulness, predicting and inferring rankings pose challenges in environments necessitating time-dependent modeling. This paper introduces a spectral ranker called Kernel Rank Centrality, designed to rank items based on pairwise comparisons over time. The ranker operates via kernel smoothing in the Bradley-Terry model, utilizing a Markov chain model. Unlike th…

  7. Related Broader context
    AI Oversight and Human Mistakes: Evidence from Centre Court
    David Almog, Romain Gauriot, Lionel Page et al. · 2024 · Machine Learning · cited 16 times · Not summarized yet

    “AI Oversight and Human Mistakes: Evidence from Centre Court” studies how Hawk-Eye affects tennis umpires’ decisions, while this paper examines nationalistic bias in skating judges’ scores.

    Abstract

    Powered by the increasing predictive capabilities of machine learning algorithms, artificial intelligence (AI) systems have begun to be used to overrule human mistakes in many settings. We provide the first field evidence this AI oversight carries psychological costs that can impact human decision-making. We investigate one of the highest visibility settings in which AI oversight has occurred: the Hawk-Eye review of umpires in top tennis tournaments. We find that umpires lowered their overall mistake rate after the introduction of Hawk-Eye review, in line with rational inattention given psychological costs of being overruled by AI. We also find that umpires increased the rate at which they c…

  8. Related Broader context
    Goodhart in the Planner: Sampling-Based Control Exploits Rewards Learned from Human Preferences
    Soovadeep Bakshi, Apurba Bose · 2026 · Robotics · AI summary available

    “Goodhart in the Planner” studies reward exploitation in robot control, whereas this paper investigates strategic favoritism and competitor penalties in figure-skating judging.

    In short

    Preference-trained reward models can be exploited by harder planning, while calibrated regularization reduces—but does not eliminate—the resulting loss in true return.

How are these papers found?

We compare the title and abstract of this paper with those of every paper summarized on this site, and with the papers Semantic Scholar recommends, those it cites and those citing it (for a much-cited paper, its most recent citations and, from OpenAlex, the most cited ones). Among the prior work and the follow-ups, the papers that use each other in depth, a method it uses and the much-cited follow-ups come a little higher. The levels (very close, close, related) come from the similarity of the texts, not from a judgment of quality; * marks a paper without an abstract, compared by its title. The sentence under a paper is written by an AI from the two abstracts: check it in the papers. A paper marked not relevant by three readers leaves the list.

Updated on 16 November 2026.

The illustrated guide to the papers around a paper

Disclaimer: The AI-based summarization tool and virtual assistant provided on this website may not always provide accurate and complete summaries or responses. We encourage you to carefully review and evaluate the generated content to ensure its quality and relevance to your needs.