In cricket, one-day-international (ODI) batsmen have traditionally been compared on the dimensions of batting average (BA) and strike rate (SR). The conventional method of computing BA assumes that runs scored by a batsman follow an exponential or geometric distribution. This results in unreasonably equating batting inconsistency with batting mean. Our study shows that a Weibull distribution model gives a very sensible assessment of a batsman’s inconsistency, independent of his BA. It also provides a superior fit to batting scores of ODI batsmen. We also introduce a measure for ‘quality-runs’ scored by a batsman which takes into account the difficulty level of opposition. Additionally, longevity index and opposition diversity index are defined to make comparisons more holistic. A substantial amount of data engineering effort is made in segregating available data into home, away and neutral matches. The measures proposed in this paper are more comprehensive and granular than those found in the literature. Various combinations of these six criteria are used to rank a select group of great ODI batsmen by assigning objective weights derived from principal component analysis. Finally, multivariate statistical outlier detection procedure produces different lists of outstanding players corresponding to different combinations of criteria. Our proposed methodology may be gainfully used by a team management to select best batsmen in a given situation.

在板球比赛中，传统上，一日国际 (ODI) 击球手在击球率 (BA) 和击球率 (SR) 方面进行了比较。计算 BA 的传统方法假设击球手得分遵循指数或几何分布。这导致不合理地将击球不一致与击球平均值等同起来。我们的研究表明，威布尔分布模型对击球手的不一致性给出了非常明智的评估，与他的 BA 无关。它还非常适合 ODI 击球手的击球得分。我们还引入了一种衡量击球手得分的“质量跑动”的方法，该方法考虑了对手的难度级别。此外，定义了长寿指数和反对派多样性指数，以使比较更加全面。在将可用数据分为主场、客场和中场比赛方面进行了大量的数据工程工作。本文提出的措施比文献中的措施更全面、更细化。这六个标准的各种组合用于通过分配从主成分分析得出的客观权重来对选定的一组优秀的 ODI 击球手进行排名。最后，多元统计异常值检测程序会根据不同的标准组合生成不同的优秀球员列表。我们提出的方法可能会被团队管理人员用来在特定情况下选择最佳击球手。这六个标准的各种组合用于通过分配从主成分分析得出的客观权重来对选定的一组优秀的 ODI 击球手进行排名。最后，多元统计异常值检测程序会根据不同的标准组合生成不同的优秀球员列表。我们提出的方法可能会被团队管理人员用来在特定情况下选择最佳击球手。这六个标准的各种组合用于通过分配从主成分分析得出的客观权重来对选定的一组优秀的 ODI 击球手进行排名。最后，多元统计异常值检测程序会根据不同的标准组合生成不同的优秀球员列表。我们提出的方法可能会被团队管理人员用来在特定情况下选择最佳击球手。