[teaspoon]

You're claiming that you can infer, just from a triple that I guess, whether I am attempting to confirm or refute the "ascending even numbers" hypothesis. But if I guess (2,4,6) and learn that it doesn't adhere, that's as much a refutation as if I guess (1,2,3) and learn that it does adhere. (2,4,6) looks like an attempted confirmation to you only because you have prior knowledge that the rule is not more specific than "ascending even numbers".

[jeremysalwen]

So I suppose it would be more accurate to say he is testing whether you adequately attempt to falsify your theory. If you only attempt to falsify it by finding false positives, but never false negatives, you are not doing it with sufficient rigor.

[teaspoon]

Is there an objective metric according to which guessing only positives is a less rigorous strategy than guessing both positives and negatives? Given no prior knowledge about the space of possible rules?