It is mostly trained skill. Fast case-math candidates are not doing advanced math in their heads. They are using repeated shortcuts, rounding rules, and pattern recognition that reduce the number of steps.

The biggest gains usually come from three things: improving multiplication and division fluency, breaking large calculations into smaller chunks, and speaking your math out loud in a clean sequence. That last point matters because interviewers reward clarity, not just the final answer.

You also need pressure-specific practice. Math that feels easy alone at your desk can feel very different when someone is evaluating you in real time. You may also find it helpful to read AI case-math drills if you want a more structured way to practice this.

I got faster once I stopped trying to do everything exactly. Smart rounding saves so much time.

For me the issue was not the math itself, it was talking and calculating at the same time.

Honestly repetition. I used to freeze on percentages and now it is mostly automatic after enough drills.