I would call this type of argument a case of regression to the mean rather than "autocorrelation". That, of course, in principle requires independence between p

I would call this type of argument a case of regression to the mean rather than "autocorrelation". That, of course, in principle requires independence between performance and assessment of performance. In many cases, it would make little sense to assume that the performance and assessment of performance are independent. But even then, one can simulate random data with some correlation, and still get a DK effect merely as statistical artifact. An overview of similar critiques, and a similar argument in https://www.frontiersin.org/articles/10.3389/fpsyg.2022.840180/full https://www.frontiersin.org/articles/10.3389/fpsyg.2022.8401... .