This page answers short questions about the fixed nudge in this book. Each answer is a few lines long. Each one links to the section that explains it in full.
Contents#
- Why does a solution that learns nothing matter here?
- Why are so few of its answers pushes?
- Why does it hold back glasses that already have room?
- Is calling a crowded table "done" a mistake?
- Why does its score have no spread?
1. Why does a solution that learns nothing matter here?#
It is the floor. It costs no data and no training. Every other solution pays for something. A solution that scores below this one bought nothing with what it paid. The introduction says more.
The next question is about what the arm spends its time on.
2. Why are so few of its answers pushes?#
The camera is on the wrist, so the arm must move to see. In the film, 3 of 26 answers are pushes. The rest move the camera. A film of five tables shows it.
The next question is about which glasses the chooser is shown.
3. Why does it hold back glasses that already have room?#
Nothing is lifted off the table in this book. So a glass with room stays where it is. The borrowed chooser never met such a glass before. It would push it again and again until the clock ran out. Glasses that already have room explains the change.
The next question is about how a table ends.
4. Is calling a crowded table "done" a mistake?#
No. The solution looks round the whole table before it says "done". So it has seen every glass. It stops because it will not push the glasses that are left. Tables it called done while still crowded gives the reasons.
The last question is about the three walks.
5. Why does its score have no spread?#
Nothing in it is random. Given the same table, it gives the same answers. So all three walks gave the same numbers. Why the three walks give one answer explains why no "± 0" is written.
The next chapter changes one step of this loop, and asks what that one change is worth.
← What is different · Geometry generates, a model ranks — what is different →