Making it fun

Once the game loop works, tune four things: the player’s main action, feedback, difficulty, and scope. The bar is not “it runs” — it’s fun in the first 10 seconds. The patterns page provides code for the ideas below.

Start with the main action

Before naming the verb, decide the fantasy — what the player gets to be. “Catch” is a different game as a frantic short-order cook, a goalie, or someone saving falling stars. Pick the fantasy, then name the verb in one word: jump, shoot, steer, match, dodge. Build the loop around it: act → see the result → respond → act again under changed conditions.

A loop this tight should be legible just by watching it for ten seconds, and fun with no art and no sound. Prototype with generated shapes from picogame_shapes before committing to final art — if the action isn’t fun as a rectangle chasing a circle, better art won’t save it. Fix the rules and feedback first.

Give actions immediate feedback

Feedback connects an input or collision to a result the player can see and hear, applied in the same update that handles the event — a delayed response reads as sluggish even at 30 fps. Not all feedback is equally cheap. Roughly in order of fun gained per byte spent:

1. A sound on the main action — the single best return for the least code; a beep confirms what the eye already saw. The simplest beep needs no .wav file:

   beep = picogame_audio.tone(880, 90)   # a short high blip, built once
   audio.sfx(beep)                       # fire it the frame the catch happens

   What carries the meaning is the contour — high reads as good, low as bad. The ready-made picogame_sfx kit gives you both:

   Catch (a light blip):
   Miss (a low, downward tone):

2. A short hit-flash — set a sprite’s opaque pixels to flat white for 1-3 frames, then clear it. Almost free:

   spr.flash = pg.rgb565(255, 255, 255)   # on impact
   spr.flash = 0                          # 2 frames later: back to normal

   

3. A restrained screen shake — picogame_fx.Shake, a few pixels, decaying fast. More reads as noise, not impact:

   shaker.add(0.6)      # on a hit — trauma is squared, so small events barely shake
   shaker.tick(0, 0)    # every frame (feed a camera offset here if you have one)

   

4. A brief hit-stop — freeze the sim for 2-8 frames on a big hit; it makes the impact connect. No engine primitive — you skip your own ticks:

   freeze = 4               # on a big impact
   if freeze > 0:           # at the top of the loop, while frozen:
       freeze -= 1; clock.tick(); continue

5. A small particle burst — a pop or ring on catch or score:

   ps.emit(x, y, 16, 4, 24, pg.rgb565(255, 210, 120))   # burst on the event
   ps.tick()                                            # every frame

   

Spend the first item before reaching for the fifth — one good sound beats five weak particle effects. Match strength to the event and keep the play area readable; picogame_fx and the native blit effects cover most of this, see effects and the patterns page.

Flashing safety

Never flash the whole screen faster than about 3 Hz (≥10 frames apart at 30 fps), and watch for many hit-flashes clustering past that — it’s a seizure risk. Prefer a localized sprite.flash to a full-screen white frame, and offer a reduced-effects toggle.

Difficulty that breathes

Difficulty should rise in a sawtooth, not a straight climb: build tension for 20-40 seconds, release it at a milestone (a cleared wave, a checkpoint), then re-engage about 10% harder. A flat ramp goes boring by minute two; a monotonic climb never lets the player exhale.

• Ramp speed, density, or variety before adding health. This changes the play pattern instead of padding the same encounter.
• Telegraph every threat. A readable wind-up, and enough time to react — the human floor is about 250 ms (roughly 8 frames at 30 fps). Below that, a hit stops feeling like the player’s fault.
• Be generous — this is what makes a game feel fair, not easy: coyote time and a jump buffer (picogame_input.Timer), i-frames after a hit so one mistake can’t chain-kill, a hitbox smaller than the player’s sprite.
• Restart in under half a second. Re-initialise the run in place; never send the player through a menu to try again.

Scope is a feature

Build the smallest complete version of the main loop, play it, and add from there. Cut a feature if it doesn’t change the player’s decisions or the strength of the feedback — a fourth enemy type that plays exactly like the first three is not depth. The device’s RAM and timing limits are a useful boundary, not just a constraint to fight.

What players remember

A run is remembered by its peak and its ending, not its average frame. Give each run a clear high point, and land the ending deliberately: on failure, show the result, add one last visual or audio beat, and offer a restart with no delay. A flourish on death costs little and is what makes the player reach for the button again.

Before → after: a catch game

A one-button game — move left/right, catch falling shapes, three misses and you’re out. Bland version: silent, shapes fall at one constant speed forever, “GAME OVER” text, no restart hint. Applying the ideas above, cheapest first:

• Sound — a blip on every catch, a distinct low thud on every miss. One picogame_sfx.Kit call; changes how every catch feels immediately.
• Hit-flash — the caught shape flashes white for 2 frames before it leaves the pool.
• Shake — a 3-4 px picogame_fx.Shake on a miss only, decaying over half a second — not on every catch, or it drowns the good feedback in noise.
• Fairness — the catcher’s hitbox is a few pixels narrower than its sprite, and a shape that grazes the edge still counts. Without this, misses feel like the game’s fault, not a slow reaction.
• Difficulty sawtooth — fall speed and spawn rate step up every 5 catches, each step ~10% harder, with the two seconds after a step-up left untouched to let the new speed register.
• Peak-end — the run’s best streak is shown, large, at the top of the game-over screen, above the score; A instantly starts a new run.

None of this changed the rules of catch. It changed how catching feels — and that’s the whole page.

---

Next: the patterns turn each of these into code; the tutorials build three games applying them.