Random Point-Buy?

[LGodamus]

its one of the 3 ways I allow PCs to roll in my game

 

[Mustrum_Ridcully]

Our group prefers the 25 point buy method, and I do also prefer it for its more balanced results, but I also would like to have some more variance (especially since it usually ended in Warrior-Types with a Charisma of 8 and Wizards with a Strength of 8). So in my new Arcana Unearthed campaign, I tested a new method that incorporates point-buy and rolling.
I let the players start creating their character with 25 point buy, and before the first game, I ordered them to spend 3 additional points on a single, randomly (roll a d6) generated stat. (For balance reason I allowed remaining points - as in the case when you have to add the points to a 15, which costs only 2 to get it to 16 but 5 to get it to 17 - to be distributed as they saw fit).

Mustrum Ridcully

 

[Dirigible]

I just fill in the blanks for ability score values with the numbers that look right for my character concept.

I'm not alone, then!

I reckon you should assign attributes based on who wins a bikini mudwrestling match.

 

[Psion]

Should there really be a possibility that some players have better starting stats than others?

I would have to say that depends entirely on the group of players. I've played in groups with a 12th level and a bunch of 6th level characters together; ability scores are a small difference by way of comparison.

 

[ichabod]

COST FIXED SCORE ROLL TYPE
0 - 8 or 4d6 Remove Highest
1 - 9 or 5d6 Reroll 1?s, Remove Highest
2 - 10 or 3d6
3 - 11 or 5d6 Reroll 1?s, 2?s Remove Highest
4 - 12 or 4d6 Remove Lowest
5 - 13 or 5d6 Remove Lowest
6 - 14 or 6d6 Remove Lowest
8 - 15 or 5d6 Reroll 1?s, 2?s Remove Lowest
10 - 16 or 6d6 Reroll 1?s, 2?s Remove Lowest
13 - 17 -
16 - 18 -

My group is currently using something very like this, mainly because the above inspired my system:

PSEUDO POINT BUY
0 - 8 or 2d4+1d6
1 - 9 or 2d8+1
3 - 10 or 3d6
4 - 11 or 2d10, reroll 2?s, 19?s and 20?s are 18?s
6 - 12 or 4d6 keep highest three
7 - 13 or 5d6 keep highest three
9 - 14 or 5d6 reroll all 1?s, keep highest three
10 - 15 or 6d6 reroll all 1?s, keep highest three
12 - 16 or 19-2d6, keep lowest one
13 - 17 or 19-2d4, keep lowest one
15 - 18 or 2d6+1d8

I scaled the costs because I don't like the curve on the standard point buy, and while an odd number is worth something, it's not worth as much as an even number.

 

[Conaill]

I was thinking along the lines of giving the players 24+2d4 points, providing a range of 26 to 32, with an expected value of 29. Does this seem reasonable?

AARGH!!! Nooooooo......... I'm all in favor of using point-buy, but point-buy with a randomized point value combines the WORST aspects of point-buy and rolling.

Point-buy:
pro: creates fair and balanced characters
con: allows and encourages min-maxing, resulting in very non-organic looking stats

Rolling:
pro: creates interesting, organic stats
con: created characters that aren't balanced with each other

Randomized point-buy creates (1) unbalanced characters, that (2) can be fully min-maxed. Much as I dislike rolling characters, I wouldn't want to touch a randomized point buy with a 10' pole, and I would probablu refuse to play with any DM who required it.

There are othere alternatives however. For example, it is possible to try combine the BEST aspects or rolling and point-buy. Ideally, you want a stat generation method that (1) creates characters which are balanced with each other, and (2) yields interesting organic stats without being fully min-maxable. Here are some possibilities:

1) Take a deck of cards and throw out all but the 1-6 cards. Deal 6 hands of 4 cards and remove the lowest card from each hand. Gives almost exactly the same stat distribution as 4d6-drop-lowest, but the set of stats as a whole will be much better balanced.

2) The DM draws up a series of 6 or 20 sets of stats which he considers balanced. Players roll a d6 or d20 to choose one of them. (Randomized "default array")

3) Roll 3 stats using your favorite method. the remaining three are equal to 25 minus the previous three (max 18). Point-buy score isn't fully balanced, but at least the total stat values are. For every high stat you alos get a low one, and for every odd one you get an even one.

4) Roll stats using you favorite method and calculate their point value. If the score is above a desired target value, decrease stats at will until it fits (but you are only allowed to *decrease* stat). Vice versa if the point score is below target (only increase stats).

 

[eris404]

I just fill in the blanks for ability score values with the numbers that look right for my character concept.

We did something close to this in one of our games. The DM gave us a list of numbers to make our characters (the same list for all characters, that is) and told us to assign them as ability scores as we saw fit, adding any racial or level modifiers (we were starting at 12th level).

This was actually the best system I ever used - even though the characters had similar stats in some ways, it didn't matter because the combinations of classes, races, etc. made them all very different. I personally felt that it was very fair, since I am notorious for rolling poorly. And the DM was able to set the tone for his game right from the beginning. Gene (the DM) wanted to have a heroic, near epic, game and chose the numbers accordingly. If he wanted a low-powered game, he could have done that just as easily.

 

[AeroDm]

We tried 27 + d6 - d4 with an average of 28 but we found too much variance. As has been said, use point buy to be fair, not random.

My favorite method of stat generation is to use a stat-draft. Have everyone roll up their stats however you want to set the tone of the game (4d6 reroll 1s for higher power, 3d6 for lower power, etc) and then record all the rolls from highest to lowest on a board.

Then roll d20's and allow people to choose what # they want, and of course each # can only be chosen once. This equalizes the randomness of the field a bit. Likewise, we once tried with #s tied to stats. So that 17 wasn't a 17, it was a 17 cha. This led to some excellent situations where people compromised and *gasp* didn't try to max every # but rather gave them away to people who needed them.

Good times.

 

[bubbalin]

My current campaign works fine with 28+2d4...
Anyway, just my two cents.

Just a tip, more dice fo smaller value = less randomness but still some variability.
But I gues most of you know that already.

 

[Psion]

I'd just like to tip my hat to Conaill for both identifying the problem and providing workable solutions.