A nice example of this used to be seen in record sales where, despite never appearing in the "hit parade", the best selling records in a year tended to be of classical music or "easy listening".
Which has nothing to do with this ranking, I think.
That paradox was probably generated by fact hit parades were based on weeks or very short periods, and trends about pop do change - as everyone knows - estremly quickly, while a classical musical release was more likely to get lower but more stable sellings during the year.
But, in this case, we precisely have the "long selling" ranking, which looks to totally confirm that, in recent times, XXX released are
globally the most sought of - even if, at time of this ranking, they were not yet at all the majority among 2021 releases.
When this happens, the overall trend might even appear to be the opposite of the trends in each group.
I can get what you point, but I don't find so relevant here. I mean, it can be also said that if - for exemple - "ordinary" card lovers would start to buy less XXX cards, but the sales of such cards would still continue to grow (which might already be the case), company would not be likely to reduce the number of XXX releases anyway - since they will almost certainly tend to take more in account global sales than the preferences of a single group of users. So, the overall trend is quite likely to be one which counts the more here.