Well maybe Warzone isn't as complex or strategic as we make it out to be. If you look at predicted vs. actual win rates on the 1v1 ladder, you'll find that a rating advantage of 125 when your rating is 1600 is actually about as meaningful as a rating advantage of 200 when your rating is 1800 (I might have the exact numbers off, but in either case the closer your rating is to average the more meaningful a given rating advantage is).
This is the opposite of what you'd see in chess, where players closer to average (1500-1700) tend to underperform their Elo rating predictions while players higher up (2300-2500) tend to be pretty close to matching them. In other words, in chess gaps near the average are exaggerated and gaps close to the edges are underestimated- and this suggests that there's actually higher kurtosis in chess; the distribution of skill is taller/wider than the Normal distribution. E.g., the kurtosis of this blue line distribution (the logistic distribution) is higher than the kurtosis of the big red one (the Normal distribution):
If we see the opposite trend in Warzone, then perhaps that suggests the opposite- the Normal distribution overestimates kurtosis, and the tails are smaller than Elo ratings assume. Maybe players at the edges aren't
really all that far from the average player- at least not as far as their rating suggests. Maybe Warzone isn't a skill-driven as we assume and after a certain point of avoiding relatively fundamental flaws + doing some advanced optimizations that exploit the choices given to you by game settings, it becomes more of a rock-paper-scissors-style crapshoot where the biggest advantage is in figuring out your opponent's mind instead of taking advantage of a strategic feature of the game.