#### Xeviat

##### Dungeon Mistress, she/her

I've been fiddling with numbers and probability all day for a project, but as ADHD do I had a brain tangent and followed an idea.

I've been using point buy ever since 2001 when one of my players rolled considerably lower stats than everyone else (but good enough to pass under the auto reroll). I ended up letting him reroll, and he proceeded to roll two 18s right in front of me. LOL. My next game just used point buy from the get go, and I haven't gone back to rolling.

But point buy has made me miss something about rolled characters; having unusual high scores. A fighter with a high int (who isn't an arcane trickster), a wizard with a high strength, or really almost anyone having high int that isn't a caster. Sure, there are some subclasses that reward having a different secondary stat, but having an unusual strength is fun.

Also, rolled stats can help you get a feel for "regular folk". But the 3d6 or 4d6 drop the lowest don't generate "normal distribution" curves. The more dice you roll, though, the closer you get to normal distribution. I poked the numbers around and found that 5d4-2 gets you 3-18, average 10.5, with a standard deviation of 2.5.

3d6 is also 10.5 average, but with a standard deviation of 2.96.

3d6 drop lowest is 12.24 average with standard deviation of 2.85

and 7d4 drop 2 lowest, minus 2 is 12.85 average with a deviation of 2.43

While I think about using rolled stats again, I rolled up some random stats in order to see if they produced playable characters. Back in 3E, rerolls were forced if sum of modifiers was 0 or lower, or if highest score is 13 or lower. Here's what I got:

Dexterity 9 -1

Constitution 9 -1

Intelligence 9 -1

Wisdom 13 +1

Charisma 9 -1

Total Mod -2

Strength 13 +1

Dexterity 15 +2

Constitution 12 +1

Intelligence 14 +2

Wisdom 12 +1

Charisma 11 +0

Total Mod +7

Strength 15 +2

Dexterity 14 +2

Constitution 11 0

Intelligence 16 +3

Wisdom 16 +3

Charisma 10 0

Total Mod +8

Strength 13 +1

Dexterity 12 +1

Constitution 8 -1

Intelligence 14 +2

Wisdom 13 +1

Charisma 16 +3

Total Mod +7

Strength 14 +2

Dexterity 13 +1

Constitution 10 0

Intelligence 10 0

Wisdom 17 +3

Charisma 14 +2

Total Mod +8

Now these stats have some interesting spreads.

#2 is an all arounder that could work for a few different builds (especially when background bonuses are factored in)

#3 has high int and wis that won't synergize and a low con, but could make a different wizard with oddly high str and wis

#4 could be interesting, even with that low con; background could be used to bump up con, or the con could encourage a very sneaky/chicken rogue

#5's also good, could be a good cleric even with the average Con.

Would you try out 7d4, drop lowest 2, minus 2?

Alternatively, what if it was 5d4 with 2 rerolls; that makes it feel like a game (any 1s would be obvious rerolls, but rerolling a 2 would be a calculated risk). I don't know how to check the statistics on that unless I forced rerolling the 2 lowest.

------------

Another idea I had on the side was having players build their characters with point buy as normal, but then have them roll a 1d6. Based on what they roll, in order down the ability scores, I could give them a random boost worth 4 point buy points; Since the point buy costs seem to be a point per modifier of the new stat (13 is 5, 14 is 7, 15 is 9, so 16 is 12, 17 is 15, 18 is 19). If that falls on their 15, it only boosts it to 16, but if it falls on a 13 it would jump to 15, 11 or 12 to 14, 9 or 10 to 13, and 8 to 12.

Also, the 5d4-2 distribution looks like this:

76.36% are between 8 and 13

95.9% are between 6 and 15

99.8% are between 4 and 17

Only 0.10% of stats are a 3 or an 18, so really really rare. Oh, 6d4-4 would get a 2-20 spread with an average of 11... okay I'll put down the dice.