Running a book

[jack]

Perhaps running a book on more things would encourage accurate assessment?

Boss: How long do you think this'll take?
Coder: Four to two weeks.
Boss: It should only take one week!
Coder: Two-to-one says you're wrong.
Boss: Ummm... maybe you're right.

Coder: Hey, who do you think's going to break the build next?

OK, probably not. But for today, last week's bug:

1-1: My misunderstanding of the spec.
1-1: Verilog bug.
2-1: Compiler bug.
9-1: Something else.

ETA: Outcome, my error, by the compiler doing something acceptible but not what I thought. So that could have arguably paid off three of the above.

You: But those don't add up to 1!
Me: Congratulations, you've discovered economics :)

Comments

[simont]

Well, if I were the boss in that exchange, I wouldn't take that bet no matter what the odds were, because (a) it's a sucker bet since the coder could deliberately go slowly in order to win the bet, and (b) even if the coder doesn't deliberately fiddle the outcome, it's still not in the corporate interest to give them a financial incentive to go slowly!