Reaction training works better when you measure sets, not single attempts. A best score can be fun, but the average across multiple rounds tells a truer story.
Remove anticipation
Random delays matter. If a test becomes predictable, you stop measuring reaction and start measuring timing guesses.