Depth beats breadth when you are trying to hear a style.
When I decided to get better at this game, my instinct was to broaden. I would play a pop round, then a hip hop round, then a rock round, rotating through the catalogue like someone sampling a buffet. After a month I had learned a little about everything and nothing about anything.
The change that worked was the opposite. One genre a week, every day, until the style stopped being a list of songs and started being a sound. The first week I did it with electronic music, which is my weakest area. By Friday I could tell a house track from a techno track from a trance track without thinking about it, not because I had memorised songs but because I had heard the same production choices enough times to recognise them.
Depth works because recognition is pattern-based. A single example teaches you a song; twenty examples teach you a category, and categories are what the game is actually testing. The archive makes this practical because it lets you find a run of songs in one lane without reading the answers.
The week does not have to be consecutive. If I miss a day, I pick up where I left off. The point is to stay in the lane long enough for the pattern to appear, which takes roughly a week of daily exposure and almost never happens in a mixed session.
On Sunday I write one sentence about what the week sounded like. That sentence is the real product of the exercise, and I have kept them all.
Pick your weakest lane, then use the song list to find twenty songs in it and stay there for seven days. The improvement is not dramatic day to day, but the seventh day sounds nothing like the first one.