I think my suggestion handles this situation though. If the cheaters are playing well enough that you can identify them as cheaters, then they would be getting flagged by my hypothetical server-side anticheat as "good" and so need to run the client-side software too.
Also I don't think "twitch streams" are a great sample of "normal level gameplay", most people there would be ranked comparatively highly in MMR I would have thought? That or I'm leaving tons of money on the table by not streaming my wood-tier CS:GO :D
I agree that matchmaking would be a factor if it were purely designed to create fair matches with an infinite number of available players would match cheaters against very good players, and the # of players vs skill level was fairly linear.
But limitations in available players in the queue (especially as skill goes up), "experience optimized" matchmaking (which psychologically manipulates you into playing more by feeding you wins and losses at the right time)...I'm not so sure.
I used to spend a significant amount of time playing a tactical shooter type game, and what I figured out was that the game would purposefully throw you soft games and hard games, especially when first partying up with people on your friend list.
One of the developers of the game publicly disclosed that their matchmaking algorithm took several dozen games to figure out how good you were. I tried creating a new account, and it took a very long time before I stopped being given opponents who could barely move, whereas I'm dodge-rolling, using my character's ult, using grenades, etc. It was so bad, I felt terrible and stopped using the secondary account. So cheaters (who keep creating new accounts) probably enjoy plenty of playtime. There were multiple cases of obvious cheaters running around achieving relatively high account levels, showing that anti-cheat measures took a long time to kick in.
Lastly: the anti-cheat software available for this relatively small game was incredibly sophisticated. Sure, you could set the hitbox detection and zero fuzz and go for all headshots, but that would make you jump off the page stats-wise, you're right. As long as you weren't super greedy, you could probably "on paper" look like just a fairly skilled player.
The other problem? The game had verystrong controller assist. So much so that many pro players abandoned the game fairly early on. It was very difficult to tell the difference between someone cheating and someone who was just good at controller play, because controller aim assist is intended to help casual players compete against MNK players. It's not intended to help very skilled controller users, who will absolutely decimate MNK players in a lot of types of gunfights (particularly close range stuff.)
For warzone, I don't believe that game has any sort of MMR, so the situation is a little different there compared to most other games people stream on twitch. Warzone seems to be especially bad, even when compared with other games in it's genre.
There's also another issue where some people play with devices like Cronus which are borderline cheats and essentially hardware scripts (better/no recoil control, scripted bunny hopping, that sort of thing).
Damn near all online games have some form of matchmaking algorithm, but the point of them is to maximize the time you spend in-game by learning your patterns. There's also matchmaking that intentionally puts you on a team with someone better tan you who has cosmetics you don't, so you hopefully associate being kickass with having that skin.
I feel like showing things like replays of the kill cam can help with this sort of thing. That should make things like wallhacks more obvious as they get flagged by other players.
Also I don't think "twitch streams" are a great sample of "normal level gameplay", most people there would be ranked comparatively highly in MMR I would have thought? That or I'm leaving tons of money on the table by not streaming my wood-tier CS:GO :D