Roll under or Roll over, which one is better?

Roll under or Roll over, which one is better?

It depends

But the way GURPS handles roll under is pretty solid. 3d6 bell curves and the assumption of 10 as the average.

Looks like a wall to me.

Generally speaking the more dice you're rolling, the more 'average' the result will be. Of course, it'll vary, but the bell curve will be more accurate with more dice. 3d6 is pretty good for this, but other piles-o-dice systems can also work.

Over, if you don't leave a mark on the table you're doing it wrong.

I guess in general, bigger numbers being better is more intuitive, so roll under?

but rolling bigger is better,no?

Literally doesn't have an effect on probability

You raise a valid point.
I guess if you're mostly rolling against a static number, roll over is better.
If you're rolling against skills/stats, roll under is better.

I prefer roll over because it's what I started with, and because to me it feels more natural that generating the larger number should have the better result.

Say that to d20 shit where your die doesn't actually matter. Wouldn't happen in a roll under system.

True, I just feel that 3d6 is that comfortable part of the curve where the results do tend to have a nice average, but arent ALWAYS average You still get crits in a margin thats accessible to play. They matter. Theyre rare, but they happen. Savvy?

Rolling over has always felt more natural to me since generally in board games etc. bigger numbers on the die are better.

This goes double if multiple sizes of dice are involved, as bigger dice have bigger numbers and are thus able to go higher.

Then again, having high numbers in skills also feels intuitive which in conjunction with roll over means arbitrary target numbers or THAC0 levels of fuckery

Linear roll-under, d20 granularity. It is officially The Best™.

Roll Through.

Show the tavern's furnishings who is boss.

Speaking of THAC0, what's the thread opinion on the old D&D way of switching back and forth between roll-over and roll-under? A good method of defeating biased dice, or pointless convolution?

Roll over > roll under.

So they both win.

That's a problem with system math rather than roll over.

If you presume the game design is very good, I'd probably say rolling over is better as it's more intuitive. If not, I'd say under just because I think it's more apt to restrain the writers. You can only have so much power creep when the worst number you can roll is still under 20 and the best is a 1.

I like you.

I actually agree with this, though mostly "changing target number" on a 20 scale. I would still go roll over (target number goes down with XP) because rolling 1 or 20 is pretty iconic in the hobby.

Roll over, because it's easier for casuals and newbies to comprehend.

Mainly pointless.

You can do something interesting with the combined system, Rolling under works for base ability checks where the DM probably hasn't even thought of the difficulty number and just wants to see how will you do something.

Doing that and have *and* stuff like your saves being a static numbers you have to roll over is just stupid.

Roll under makes players feel better by dent that the bigger number they have the better they are. imho

dint*

>dice pool
roll over
>D20
roll under

lol that pic fuckin got me at the end there

>you want to roll less than this number
>I don't understand

Rolling a Big Number = Bad just seems weird to me, but both ways are equally valid.

I just prefer roll + stat > target number, because that way both can be big and good.

>Generally speaking the more dice you're rolling, the more 'average' the result will be
That's not how it works at all.

It is true, it's just not the way to say it. If you're rolling 3d6, it will most often end up at 10 and 11

I like curves, but they make calculating the impact of bonuses/penalties a bitch. Being more drawn to the average is good for simulationist games tho.

I prefer roll under because I never roll anything higher than a 10 on a d20 or a 40 on a d100

HA

>what is variance?
that's exactly how it works

Please explain? Is there some issue with statistical distribution?