Fighting games are possibly the only genre I would think where we could demonstrate statistically significant effect sizes because the amount of other variables shrinks substantially. A 1v1 2d fighting game is practically laboratory conditions for measuring input latency. You are fully aware of your and your opponents position at all times, fully focused on reacting to them and executing your strategies, and in some titles missing one frame means a combo fails to execute. In Counter Strike it's a full 3d environment with large levels, different weapons and items, fuzzy positioning, lots of stimulus coming from all different directions, objectives to consider, team tactics to execute and so on.