Making Perception Checks Less Random

[Captain Joy]

I don't like the excessive ranges in the margin of success you get in GURPS when you make a perception check using the usual 3d6-roll-under-target-value. I think from now on I'm going to have my players roll two Fate dice and add the result to 10. This will give results of: 8,9,9,10,10,10,11,11,12.

This way, having a +1 or +2 better perception than somebody else actually means something. Should I make Perception and Acute Senses more expensive because of this?

Any other unforeseen consequences I should be aware of?

I thought about a mechanic like rolling a d6 or d4 and adding the result to 10, but the two Fate dice give a nice distribution.

[corwyn]

I just have passive perception equal to 10 for cases where someone is using stealth, etc. against them. Applies to PCs and NPCs alike. Fast, simple, done. While I've been doing that for awhile, there is a pyramid article, Turn it up to 11, that talks about the same idea and expands it quite a bit.

[Brandy]

I've posted before that borrowing the rules for take 10 and take 20 from The Other Game (tm) as Take 11 and Take 6 would work pretty well.

I definitely need to start digging in to the pyramid vaults, though. I know there's lots of good stuff there.

[evileeyore]

Originally Posted by corwyn

I just have passive perception equal to 10 for cases where someone is using stealth, etc. against them. Applies to PCs and NPCs alike. Fast, simple, done. While I've been doing that for awhile, there is a pyramid article, Turn it up to 11, that talks about the same idea and expands it quite a bit.

Yup, I adopted the Pyramid article's tips as well.

[Anthony]

The thing about perception checks (and a lot of other checks, such as shooting people) is that a high degree of randomness makes sense as long as you assume that there are a whole bunch of unmentioned situational modifiers out there, and that much of the randomness of your roll isn't actually for variation in skill, it's for clarifying those unstated situational modifiers. As such, the more you nail down exactly the situation under which the roll occurs, the less random it should be.

[Captain Joy]

Originally Posted by Anthony

The thing about perception checks (and a lot of other checks, such as shooting people) is that a high degree of randomness makes sense as long as you assume that there are a whole bunch of unmentioned situational modifiers out there, and that much of the randomness of your roll isn't actually for variation in skill, it's for clarifying those unstated situational modifiers. As such, the more you nail down exactly the situation under which the roll occurs, the less random it should be.

Well said. If it were something like reading an eye chart at the a doctor's office, I'd probably go with "turn it up to 11". If it was to notice that the eye chart was covering a hole in the wall–creating subtle color/hue variations in the eye chart–I'd want to introduce some randomness.

[Say, it isn't that bad!]

I have that problem with strength checks in most RPGs. You try to lift something, and either succeed, succeed spectacularly, fail, or fail spectacularly.

Real-world experience and observation says that actual chance only comes into it when the weight you're lifting is within about 10% of your max, or the object is awkward and causes leverage difficulties. The latter case could maybe be better handled by a "dexterity" check.

[Skarg]

Your percentage breakdown is very convincing to me that in those cases, I quite like the way a straight 3d6 roll against 10 vs. 12 (or 5 vs 7) works out. I think the average dice are weird because they focus importance on the middle range, leaving no outliers, and make for forced limited outcomes even with typical values. There aren't many circumstances under which I think I'd want that.

On the other hand, I have so many decades as GM relating to things as rolling 3d6 against TFT/GURPS scores with modifiers, that I know what values to assign to get what chances I think I want. If you have different habits for assigning modifiers, maybe you could get what you think you want from average dice.

[dfinlay]

Here's the thing: if you actually want character 1 to almost always beat character 2 in perception checks, realistically, that means either C1 has some truly amazing ability in the sense and/or (more likely) C2 has some real deficit in it. Thus, the attributes should be more 10 and 14 (someone with an amazing ability) or 6 (someone with some real issues in the sense - partially deaf or very poor vision, for example) and 10 or even 6 and 14. In this context, C1 will very often beat C2, but still not always (maybe (s)he was completely zoned out when it happened). This seems reasonable to me. Scores of 10 and 12 are more within the normal range of ability, where there is some variance in capability but you usually don't see people almost always beating other people. Maybe you just need more spread out scores to get the effects you want.

But then, you might say, having high PER becomes expensive. Well, if you want it to be cheap, you can always change the price, but I have changed the cost of PER to [8] in my games and my players still think it's worth it. PER is a very useful attribute in most games.

[ArchonShiva]

Seriously, [20] for "Notices stuff pretty much all the time" is nearly an amazing deal.

Splitting Per from IQ probably serves your purposes better than Fate Dice, actually. And IQ is underpriced with free Per and Will anyway.