Worldbuilding & Navigation · Week 6 of 9
Test against reality
Test it · Judge it
FORGE Test · Evaluate
Today's goal
Run the model several times and compare with real data.
By the end
Test log.
What we'll do (60 minutes)
- Predict the model's behaviour10 min
- Run 5 times; record outputs30 min
- Compare with real data15 min
- Record failures and suspected causes5 min
Need a hand?
Compare one output with one number.
Want a challenge?
Explain the mismatch.
Try this week
AI this week
Kōkako, your AI mentor: Asks you one question at a time, such as “What does your model leave out?”. It never tells you the answer or whether you are right. You: Decide which rule to change, one at a time, and retest.