AuraTracer智迹闻
中文

EVENT DOSSIER

A Fairness Audit of the Duckworth-Lewis-Stern Method: Format-Specific and Gender-Differential Bias, with an Interpretable Calibration Layer for Cricket Target Revision

2026-09-07 12:00 Science 🔥 42.2 heat score
1sources
1days unfolding
42.2heat score
5mentions
SummaryAI generated

The researchers conducted a large-scale audit of the Duckworth-Lewis-Stern (DLS) method widely used in international cricket matches and found that the algorithm had format-specific biases and gender differences. Data showed that in women’s first-base matches, the DLS method averaged a 7.63-point overestimation of scores. To address this issue, the research team generated a dataset containing 233,550 synthetic scenarios based on Cricsheet data and compared it with five modern alternative models. Finally, they proposed a lightweight calibration layer called DLS-Cal, which consists of 27K parameters. This module significantly reduced the absolute deviation in women’s first-base matches from +6.19 points to +0.65 points, while maintaining the original calibration effect for men’s matches.

Related eventsRELATED EVENTS
Key entitiesKEY ENTITIES
Bi-LSTMCricsheetDLS-CalDuckworth-Lewis-SternXGBoost

Coverage · reports per dayLANGUAGE SPLIT

Entity relations
Bi-LSTM × Cricsheet1Bi-LSTM × DLS-Cal1Bi-LSTM × Duckworth-Lew…1Bi-LSTM × XGBoost1Cricsheet × DLS-Cal1Cricsheet × Duckworth-L…1

SignalsSIGNALS

Keyword heat
  • Duckworth-Lewis-Stern1
  • DLS-Cal1
  • Cricsheet1
  • Bi-LSTM1
  • XGBoost1

All reports (1)SOURCES

A arXiv cs.LG en 2026-09-07 12:00

A Fairness Audit of the Duckworth-Lewis-Stern Method: Format-Specific and Gender-Differential Bias, with an Interpretable Calibration Layer for Cricket Target Revision

研究人员对国际板球 DLS 方法进行了大规模审计,发现其存在格式特定偏差及性别差异:在女子一垒比赛中,DLS 平均高估比分达 7.63 分。研究基于 Cricsheet 数据生成 233,550 个合成中断场景,对比五种现代替代模型后提出 DLS-Cal 校准层,该轻量级模块(27K 参数)将女子一垒绝对偏差从 +6.19 降至 +0.65 分,同时保持男子校准不变。