Another approach is having an "explore and test" or "Training Wheels" mode. Correct notes played by the user behave exactly as they do right now. But for a wrong note played by the user, instead of the round ending, you could just gives some audio-visual feedback that this is not the right answer but continue to expect the user to find the right answer. This way we can hunt and peck our way to the end of the round.
This creates a problem in that it's easy to muddle your way through without learning anything. To prevent this:
Once it gets to the end of the round in this mode, if the user had even 1 wrong note selected in this round, the game will then expect them to play it perfectly once again (like it does now).
This way you get both the hunt and peck exploration and the final "now that you've had your time to get your ears and fingers in order, play it correctly."
1. Doesn't work on Firefox because of compatibility. (Web MIDI could not be enabled: DOMException: WebMIDI requires a site permission add-on to activate index-C82-1cwq.js:10:56869
doInitialize https://lend-me-your-ears.specr.net/assets/index-C82-1cwq.js...
) . Might be worth detecting browser and telling users straight up before they waste any time.
2. How did you come up with the melodies? Hand crafted? AI? Asking because I plan on building this for Gitori(www.gitori.com) at one point and my original idea was taking snippets from MIDI files of famous classical pieces/jazz solos
Also, TIL about the Simon game. Might buy this for my 6 year old :)