You cannot study your way to returning a 200 km/h serve. Native speech is the serve.
Nobody returns a professional serve from theory. You can know the mechanics cold — toss height, kinetic chain, likely placement off this ball position — and none of it survives contact with the actual ball, because the ball arrives faster than deliberate thought. The return exists only as a conditioned reflex, and the reflex is built one way: standing in the box while aces go by. There is no other apparatus. Club players who spend years rallying never develop it, not because they lack talent, but because rallying is a different game played at a different speed.
Native speech is the serve. Real conversation arrives reduced, blurred at the word boundaries, in bursts that peak at two and three times the speed of classroom audio. Like the serve, it outruns deliberate thought — by the time you have consciously parsed the first phrase, the speaker is two phrases ahead. And like the serve, comprehension at that speed is not knowledge. You may know every word on the page and still watch the sentence go past you for a clean ace. What returns a serve is not knowing; it is a nervous system that has seen thousands of serves.
Now look at how languages are actually studied. Grammar is serve mechanics on a whiteboard. Vocabulary decks are watching slow-motion footage. Graded listening material is rallying — comfortable, sustainable, and capped at a speed no opponent will ever offer you. A patient tutor is a coach feeding balls: genuinely useful, and constitutionally incapable of serving at match pace, because the whole arrangement exists to keep the ball in play. All of it can fill years. None of it puts you in the box.
The uncomfortable conclusion follows directly: the only practice that builds the return is facing real serves at real speed, and losing, repeatedly, on purpose. Nobody speaks B1 tennis — native speakers play the pro game among themselves, and the only way in is to stand on that court and get aced until your brain catches up. The losing is not a phase to be minimized. The losing is the apparatus.
This is exactly what a dictation stream of native recordings is: a serve machine set to pro speed. Random sentences, real speakers, no rally, no mercy — and a scoreboard that records every ace against you. The spaced repetition brings the serves you missed back until you can touch them, then back again until you can return them. A good regime even feeds the occasional easy ball to a struggling player — same sentence, friendlier speaker — but it never pretends the easy ball is the match.
The players who eventually return serve are not the ones who studied longest. They are the ones who stood in the box longest. The method is not complicated. It is just not comfortable.