How was the theory proven correct?
Has someone actually looked into the code of the game, or are you just talking about average spawn rates?
The coeff 6 items don't have a rarity of 1 in 22,000. Effect coeff 6 items have a 1 in 15,000 chance, and point coeff 6 items have a 1 in 40,000 chance.I am too tired to do any maths right now, but wouldn't a 1/15000 rate on effect C6 items and a 1/40000 rate on point C6s average to something between 1/20000 and 1/30000 for C6 items in general? I mean it's clear that effect C6 items are more common than point C6 items. So the 1/22000 would sum up all C6 items.
Coeff 5 items have rarities nearer 1 in 1,000 rather than 1 in 700. This one could be right, I've never counted this one on my account. Even though it doesn't influence the numbers too much, you should still consider the free

,

and

that can be got as Coeff 0 items and therefore shouldn't be included in the calculations.
is "coeff 7" in the code, and is 60% as likely as finding a specific coeff 4. For me, it is more common than some other C4s. Also there is pd33 of course, so it is difficult to count. But this also may be right.
To get a specific coeff 6 effect item, you need to on average find 60,000 effect items. 600,000 are needed on average to unblock a specific c6 effect item. If you got 100 effect items per game, that would mean playing 6,000 games to unblock a certain c6 effect item. To unblock certain point c6 item, 16,000 games are needed on averageAlthough this sounds realistic, this one can definitely not be proven by looking at the average account and comparing, since there are way too little C6 items found in total.
But again, it would be good to know how this was proven. If someone really looked in the code and found all this out, then it has to be true, and all my arguments are invalid

But if I just look at how many items people got, I've got my doubts about some of these "theories".
