Scholarlink

<p>Forecasting plays a pivotal role in effective operational management, providing critical insights for decision-makers. This paper endeavors to discern the comparative performance of human and algorithmic forecasting, especially within crises, to test the resilience and adaptability of these methodologies. Using data from a Swiss automotive distributor, the researchers show that algorithmic forecasts generally outperform human forecasts, particularly during crisis periods, while differences are less pronounced in stable environments. Importantly, the analysis reveals that crises primarily affect the directional bias of forecasts rather than their absolute accuracy. For both human and algorithmic forecasts, crisis periods are associated with systematic overforecasting, while overall forecast accuracy remains relatively stable across crisis and non-crisis situations. This study contributes valuable insights into how different forecasting methods respond to diverse crisis contexts, enhancing decision-making knowledge and organizational resilience in times of uncertainty. Furthermore, by delineating the conditions under which algorithmic forecasts provide the greatest advantages, this paper aims to reduce decision-makers’ skepticism toward algorithmic approaches and support their effective integration into managerial practice.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Navigating uncertainty: human vs. algorithmic forecasting approaches in times of crisis

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月3日

查看全文
  • Alessandra P. C. Huydts,
  • Klaus Möller

摘要

Forecasting plays a pivotal role in effective operational management, providing critical insights for decision-makers. This paper endeavors to discern the comparative performance of human and algorithmic forecasting, especially within crises, to test the resilience and adaptability of these methodologies. Using data from a Swiss automotive distributor, the researchers show that algorithmic forecasts generally outperform human forecasts, particularly during crisis periods, while differences are less pronounced in stable environments. Importantly, the analysis reveals that crises primarily affect the directional bias of forecasts rather than their absolute accuracy. For both human and algorithmic forecasts, crisis periods are associated with systematic overforecasting, while overall forecast accuracy remains relatively stable across crisis and non-crisis situations. This study contributes valuable insights into how different forecasting methods respond to diverse crisis contexts, enhancing decision-making knowledge and organizational resilience in times of uncertainty. Furthermore, by delineating the conditions under which algorithmic forecasts provide the greatest advantages, this paper aims to reduce decision-makers’ skepticism toward algorithmic approaches and support their effective integration into managerial practice.