In addition to the biases, there is another flaw in this analysis- they assume that quality is a linear function of critic scores. So, jumping from 30 metacriti

In addition to the biases, there is another flaw in this analysis- they assume that quality is a linear function of critic scores. So, jumping from 30 metacritic score to 50 metacritic score is considered the same as jumping from 60 to 80. This is quite unlikely to be the case. Thus, they should at least also incorporate absolute Metacritic score into the 'disappointment metric'. As an example, Kanye's MBDTF is great, but his two albums prior to it were 'very good', with an average of 77. While, Lenny Kravitz went from 58 to 70, i.e. 'ok' to 'pretty good'. You should penalize higher absolute scores on previous albums for 'surprisingly successful'. Likewise, higher absolute scores on current albums must be penalized in the criterion for 'most disappointing'. One certain way of addressing these biases and the linearity assumption is by plotting the distribution of metascores, and using that as a prior to improve fairness.