Odraude |

As the title suggests, what if you used 2d10 instead of 1d20? How would it affect the outcome of rolls and such? Here is a chart that shows the statistical difference between the two. In the parenthesis are (average/Standard deviation).

master_marshmallow |

It would make auto-fails impossible, and it would make auto-successes much less common

playing this way, i would say that any time you roll 2 10's (which would be a crit range) you auto-hit and when you roll 2 1's you auto-fail (depending on how you play also damage your equipment) without rolling confirmation rolls

it cripples crit ratios tho, as on a 15-20 threat range you now have a 6% chance of threatening a crit vs. a 30% chance

i always prefer better probability over higher averages, but thats me, you'd definitely have to compensate for the change in mechanics

Odraude |

i did do my math wrong so lets check the work

we need our results to add up to 15 or above

d1 needs a 7 or above meaning 4 numbers

(4/10)

and d2 needs an 8 or above, meaning 3 numbers

(3/10)

or vice versa, do the probablility and we get

12/100 or 12%idk where 21% came from

The 21% comes from the chance of hitting a 15+ on the bell curve of 2d10. Looking at the charts I linked to, Getting a 15 is 6, a 16 5, and so on, adding together to become a 21%.

Odraude |

The real issue is on very small crit ranges. For instance, you have a 1/100 chance of getting a 20.

Agreed. For 2d10 to come close to the 5% chance, you'd have to make 20's actually 18-20 (6% chance). 19-20 would be 17-20 on 2d10 (10%), and 18-20 would be 16-20 on 2d10 (15%). Not a hard progression to remember admittedly. Although when you get to Improved Critical, things would get weird. 17-20 would be 15-20 (21% as opposed to 20% on 1d20), 16-20 would be 14-20 (28% as opposed to 25% on 1d20), and 15-20 would be 13-20 (36% as opposed to 30% on 1d20!). So trying to make Improved Critical work in this system would have to change from a 1d20 system since critical builds would get considerable better if you try and scale them accordingly.

Nipin |

How about increasing the crit range for all weapons by 1 (e.g., 20->19-20 and 19-20->18-20) and then make crits auto succeed. I would also remove any ability or enchantment which doubles crit ranges. Now the typical crit build which used to be 15-20 with 30% chance to potentially crit becomes 17-20 with 10% chance to have a guaranteed crit.

On a side note I would take abilities which used to auto-confirm crits (such as the rage power) and have them increase the crit multiplier by one instead (e.g., x2->x3 and x3->x4).

I always hated that a crit is not always a crit. Additionally, this would simplify the game slightly by getting rid of the extra roll to determine if the hit crits. I can't tell you how confusing that rule is for every single player we teach the game.

Odraude |

So we can definitely see how this affects criticals and the feat Improved Critical and Keen weapons. If one were to use 2d10s, make the expansion of 15-20 into 14-20 instead, which under the 2d10 system would be 28%. Slightly under the 30% of 15-20, but close enough? I'd suggest completely reworking the feat/*keen* property to something else entirely. Maybe it could allow you to roll two d20 to confirm and take the highest?

How does one think skills, saves, and attack rolls would be affected? Especially since they will usually hover around the 11 mark for rolling?

master_marshmallow |

higher averages overall mean making checks is a lot easier

at low levels where you have <5 ranks into something this is a GODSEND

at higher levels it makes rolling skill checks all but unnecessary

with a minimum roll of 2+ Ability score, the skill system would have to be reworked in a way that made up for it by either increasing the DCs or reducing the skill ranks available, unfortunately this complicates the (imo) much better skill system PF has over 3.5

Whale_Cancer |

Unearthed Arcana had a 3d6 'bell curve' variant.

I used to love this idea. I ran it once and realized it sucks.

Bonuses and AC/DCs don't scale well enough for this system to work. You'll often get into a situation where characters and monsters will (nearly) automatically succeed or automatically fail at any given action. PCs who put any effort into AC will be unhittable when facing enemies of an appropriate CR.

2d10 might work better, but the same basic problem would persist. The game would need less sources of bonuses for a bell curve to work well.

Odraude |

The average of 2d10 is only .5 higher than the average on a d20. Of course, you do have a better chance at getting that 11 on 2d10. Looking at this graph I posted, I switched to the "At Least" view to see the probabilities of rolling at least a certain number. From what I gathered, from numbers 1-11, 2d10 have the upper hand. If you need to beat a DC that only requires you to roll below a 12 on the dice, 2d10 have a better chance of landing it. The biggest percentage difference is on rolling at least a 7 on the dice, where the 1d20 has a 70% chance while the 2d10 has an 85% chance. This makes for a 15% difference.

Once you get past 7, the percentage differences begin to lessen until 12, where they are equal. At requiring a 13+, however, the 1d20 has the advantage. As you go from 13 to 17, the d20 will land the numbers more often than the 2d10. 16+ and 17+ have the largest differences, at a 10% difference between 1d20 and 2d10 in favor of 1d20. It converges again from 18 on, but the 1d20 still has an advantage all the way through 20. So that makes rolling saves and such much more different with 2d10. As expected, easy DCs get easier, but harder DCs get harder.

Odraude |

Unearthed Arcana had a 3d6 'bell curve' variant.

I used to love this idea. I ran it once and realized it sucks.

Bonuses and AC/DCs don't scale well enough for this system to work. You'll often get into a situation where characters and monsters will (nearly) automatically succeed or automatically fail at any given action. PCs who put any effort into AC will be unhittable when facing enemies of an appropriate CR.

2d10 might work better, but the same basic problem would persist. The game would need less sources of bonuses for a bell curve to work well.

Interesting. I went ahead and added 3d6 to the table in that link. It has the same average as a 1d20, but a lesser standard deviation and a much smoother curvature (with a higher percentage peak) than the 2d10. On a d20, you have a better chance of rolling a 1-6 and a 15-18 (technically 20, but 3d6 can't roll above an 18 heh). Of course, between that, the 3d6 takes the cake at rolling, with significantly higher percentages for that middle ground. It'd be essentially like taking 10 all the time. In terms of rolling "At Least" a certain number, the d20 and 3d6 converge at 11, wherein the d20 takes the lead. The 2d10 and 3d6 converge between 9 and 10, with 2d10 overtaking 3d6 from 10 on. 2d10 ends up having a better percentage of making at least a certain roll than 3d6 from then on out.

Honestly, it looks like 2d10 is closer to 1d20 than 3d6.