Week 1 Tutorial: Three predictions

Read the specimen's rules, then turn three opinions about its balance into predictions that a measurement could prove wrong.

25 February 2027

Before the tutorial#

Read rules.ts from top to bottom. It is about 800 lines and it is the whole game. Do not run it and do not look for a playable build.

In the tutorial#

The first 20 minutes are in pairs. One person states an opinion about the game from the rules alone (“the shockwave looks weak”, “the second boss comes too soon”). The other refuses to accept it until it has a number, a unit and a run count in it. Then swap.

The rest of the time you write your own three and put them on the wall. The room votes on which prediction is most likely to be wrong. Keep that one; it is the first thing you test in week three.

Where it goes wrong#

Three ways, and the room will do all three in the first ten minutes.

• No run count. “A player who skips upgrades survives past 66 s” cannot be checked. Once? Always? “In at least 1 run out of 15” can.
• A prediction that cannot lose. “Sometimes the second boss is too hard” is true of every game ever made. If no result would embarrass you, it is still an opinion.
• A unit that hides the claim. “Survives longer” is a direction. “Median survival rises by more than 5 s over 51 runs” is a size.

What a finished answer looks like#

Banning the shockwave costs an ordinary player fewer than 3 wins out of 51. Wrong if it costs 3 or more.

One sentence for the claim, one for what would prove it wrong. Week 6 measures this one.

Afterwards#

Your three predictions are your answers on the first quiz. Nothing else is marked this week.