L29 TRACE ANNOTATION WORKSHEET
Label: No AI

Run you are annotating: L29_runs/L29_run_scripted_hasty.txt
(the run L29_run.py produces with no options)

For each step, fill in five lines. Write in plain words.
  DECIDED   what the decider chose, and the reason it gave
  EXECUTED  what the loop actually did in the sandbox
  CAME BACK what the result said (the next observation)
  JUSTIFIED was the decision supported by what had been observed so far? yes / no / partly, and why
  CORRECT   was the step correct in fact? (it can be justified and still wrong, or the reverse)

Step 1
  DECIDED:
  EXECUTED:
  CAME BACK:
  JUSTIFIED:
  CORRECT:

Step 2
  DECIDED:
  EXECUTED:
  CAME BACK:
  JUSTIFIED:
  CORRECT:

Step 3
  DECIDED:
  EXECUTED:
  CAME BACK:
  JUSTIFIED:
  CORRECT:

Step 4
  DECIDED:
  EXECUTED:
  CAME BACK:
  JUSTIFIED:
  CORRECT:

Step 5
  DECIDED:
  EXECUTED:
  CAME BACK:
  JUSTIFIED:
  CORRECT:

Step 6
  DECIDED:
  EXECUTED:
  CAME BACK:
  JUSTIFIED:
  CORRECT:

FIRST INCORRECT STEP: ____
Which later steps inherited the error, and how?


HUMAN DECISION POINTS
Propose at most three places to stop the loop and ask a person.
For each: the step number, what the person is shown, what they decide,
and why this point is worth the delay.

1.

2.

3.
