Yes, but that doesn’t invalidate his point. Basic math says that in a system with four options if everyone just chooses what they like best and what they like best is completely random, we’d expect to see 25%/25%/25%/25%.
If what they like best isn’t completely random, i.e because Necrolord is clearly the DK covenant and Night Fae are even more clearly the druid covenant, we expect to see the numbers distorted in that direction: people choosing what covenant they like best are likely to pick Night Fae even if they’re not min-maxing because it’s thematically appropriate.
The point is that we can’t calculate “everyone is min-maxing” based on the absolute percentage of people who are in the “best” covenant (or even the best three covenants). We have to consider that up to 25% of each class and quite possibly more in some cases are expected to be in each covenant just because they like it. So:
Okay, so 60% necro minus 25% who we expect to pick it randomly leaves 35% who picked because it’s the best choice (or less —necro is obviously thematically “the” DK covenant). 26% night fae basically fits into the amount we’d expect to pick randomly, leaving 1% who’ve picked for min-max reasons (although to be fair it’s probably more than that—no DK can really think night fae are thematically appropriate so non-min-maxers would look elsewhere, probably evening out the necrolord bulge a bit) 8% venthyr, 5% kyrian—necrolord success had to come from somewhere, they’ve clearly cannibalised venth and kyr. So that leaves us with 25% of necro, 8% venth and 5% kyr choosing based on what they like: 38% of total players (plus a few night fae, probably), which tells us in fact only about 62% of DKs are min-maxing. A majority for sure, but not an overwhelming majority.
That’s about the same breakdown, with the caveat that Night Fae are even better aligned thematically with druids. So we have at least 25% night fae picking what they like, maybe 15% of necro (we could call it the full 25% and observe they’re right in line with random distribution but I feel like few druid players would feel especially aligned with necro), 3% venth and 2% kyr, giving us 45%ish of people picking what they like and something like 55-65% min-maxing.
As we’ve discussed, DK and druid (and Paladin) have the problem of distorting the statistics because they have strong themes that would distort the random distribution even if there was a random distribution. Your third example is incredibly useful in that respect because I don’t think mage aligns perfectly with any covenant and we really can expect non min-maxers picking whatever covenant they like to achieve a fairly random distribution.
So we have 52% kyr, 26% night fae and 21% necro and venth. That’s actually pretty much what we’d anticipate: the covenants trend towards 25% each and the best one dominates. In total though that only gives us 27% min-maxing with Kyrian (the other 25% we expect to pick Kyrian anyway if they were picking what they liked) and maybe 1% with night fae. That leaves us with 28% of mage players min-maxing and a whopping 72% picking covenants they like.
That’s probably not quite right either, but all of this math is pretty sketchy. It’s clear though that we’re talking about far less than 90% of players taking cookie cutter choices; even with the massive distorting effect of druids’ cookie cutter choice also being “the druid covenant” they didn’t get more than 65% of people making the cookie cutter choice. 65% is a majority, but it’s not 90%. In cases like mage where we expect more randomness in choice because any covenant can thematically fit, we actually see quite an even distribution with relatively little distortion toward the cookie cutter choice.
TL;DR: even if everyone was picking covenants at random you would have to expect each covenant to get 1/4 of all players (25%), so when calculating how many people have chosen to min-max you need to take out 25% we would expect to pick it whether it was best or not. This is amplified by classes like DK and druid who are thematically closely aligned with a specific covenant that’s also their best, so both RPers and min-maxers are disproportionately choosing the same covenant. If you account for that, we swiftly see that actually more like 60-65% of people are picking based on what’s best, which is borne out by looking at classes such as mage which has a very even split with relatively little weighting toward the best covenant.