Behavioural Data Science Seminar Series: Chengwei Liu on Skill versus Luck
3013 , 5.00 , #Behavioural #Data #Science #Seminar #Series #Chengwei #Liu #Skill #Luck
On January 22, 2021 Dr Chengwei Liu (Associate Professor at the ESMT Berlin) led a discussion on #Skill versus #Luck in a variety of contexts at the #Turing #BehaviouralDataScience Seminar Series.
Performance metrics are all around us and usually higher #performance is assumed to indicate superior future performance. If the correlation between past and future performance is positive such a #heuristic may work well even if past performance is an imperfect #predictor of future performance. However, recent theoretical work has demonstrated that statistical regression effects can also lead to a “less-is-more” effect where expected future performance is a decreasing function of past performance for some performance ranges, which would violate the higher-is-better heuristic. We empirically test whether #lessismore effects occur in 4 domains, using 8 datasets and 3.4 million actor-performances altogether. All datasets show strong and non-linear regression effects. The presence of less-is-more effects presents a more nuanced role of the impact of luck on performances and suggests an alternative source of #strategic opportunities.
Connect to Chengwei via:
His webpage:
LinkedIn:
Twitter: @ChengweiLiu
Google website:
Chengwei’s recent book on Luck:
Nguồn: https://game5s.net/