Almost everyone can tell when a mix sounds wrong. What separates the person who fixes it is being able to say «there's a resonance around 300 Hz» instead of «I don't know, it sounds off». That translation from sensation to number is the only thing that turns an EQ into a decision rather than moving bands to see what happens.
EQ Surgeon trains that and nothing else. You hear real material — not test tones — with a curve applied, and you have to locate it. The game isn't the point: knowing instantly whether you were right is, and that's exactly what's missing when you practise by EQing your own mixes.
The modes aren't difficulty settings on one exercise, they're chained skills. A vs B trains noticing that something changed. Multiple choice trains identifying which. Drawing the curve trains producing it, which is what you'll actually do in front of an EQ. Skipping the first to jump to the last is the usual mistake, and the reason many people quit convinced they have no ear.
The reasoning — which filter does what, and why Q matters more than decibels — is in the EQ guide. To associate each band with the instrument that occupies it, Frequency Drop.