[naloth]

Originally Posted by Thamior

So? I explain why it works better using your own example. With basic rules it boils down to truth/lie dichotomy with no way of knowing if you have succeeded. And it is too punishing. I have described the scenario in my terms. Could you describe it in yours thoroughly? The same situation: 10 suspects. Character with Sensitive and IQ 10, character with Sensitive and IQ 13.

Using your rules, just ask them each 5 times, in a different way, using Empathy to detect if any of those 5 statements is a lie. Since you'll be unlikely to get more than 1 false positive, 2-3 statements verified, and 2-3 statements of "not sure." If you get a contradiction, you know you've hit your 1/64 chance of a crit failure.

As the rules are right now, you can't rely on your chance of failure to give you "not sure" so you have to base your Empathy on the fact that you're X% sure of any given statement.

Who told you that? This is a total misunderstanding of how the rules of GURPS and RPGs in general work in the first place.

That you can often reattempt the same thing later?

You don't get to reroll that. It's something like a Reaction Roll and Intuition for that matter. Maybe once in a blue moon GM may allow to reroll it but only when something significant happens under very special circumstances. It's the impression, you can't just keep asking yourself for new impression, it sticks.

You apparently need to go back and read the advantage. It has 3 different pieces:
1) You get "first" impressions if you want.
2) You get a skill bonus (always).
3) You can use it determine if someone lied to you (any statement).

Hahaha. I get it now. You see, you totally break the Basic Set rule by rerolling. It's like you know your success chance (say, 20%) and you keep rerolling to make enough examples. And then you see the pattern. And then your success chance becomes irrelevant. Way to go!

Sarcasm or lack of comprehension? Your statement amounts to "whoops, you missed with your first attack, guess you can never attack anyone again!" This is not a re-roll. This is a further attempt to do something similar later. If nothing interferes with you doing it again and again (subject being held for interrogation), there's nothing that prevents you from either benefiting from extra time or repeat attempts.

Failure is indeed misleading in BOTH cases. BUT success in the spot check is immediately evident. You can act on it reliably. Not so with Empathy. You are never sure if you have succeeded. This is THE difference and this is why your example is bad. It's nuances like these that matter and that you seem to either don't notice or don't grasp.

You want to change failures, such that you can get reliable information from them. If you fail to notice, you shouldn't get information that you will act on. Your changes aren't about making making success more likely, they are making failures useful.