You are the worst possible judge of your own game's difficulty. You've played every version, you know every mechanic, and you built the thing. Everything feels easy to you and brutal to everyone else.

The first sixty seconds decide

A new player needs to feel competent almost immediately. Not challenged — competent. Let them do the core action successfully within a few seconds, before anything can punish them.

Teach one thing at a time, through play rather than text. If your game needs an instructions screen, it usually needs a better first level instead.

Ramp on the right axis

Increasing difficulty by making numbers bigger — more enemy health, faster spawns — works for a while and then just feels like padding. Better ramps introduce genuinely new demands: a second thing to track, a mechanic that combines with an old one, a constraint that invalidates the easy strategy.

difficulty-curve.txt
My game: <core mechanic in one sentence>
Currently it gets harder by: <what you're doing now>

Design a progression for the first ten minutes. For each stage:
- What the player is learning or being tested on.
- What's introduced, and what's deliberately absent so far.
- Roughly how long before the next change.

Flag any point where a new player would likely quit, and what I'd
change there.

Failure needs to be fast and cheap

The single biggest determinant of whether people retry is how quickly they can. Instant restart, no menu, no cutscene, no loading. "One more go" only happens if the go starts immediately.

Make the failure legible too — the player must understand what killed them. Deaths that feel arbitrary end sessions.

Watch someone play

Nothing substitutes for this. Sit a person down, say nothing at all, and watch. Every place they hesitate, misread a control, or die confused is a design problem. Resist explaining — if you have to explain it, that is the finding.

Note where they smile and where they put the phone down. Those two timestamps tell you more than any amount of analysis.