Attribute Score Generation - How Do You Do It?

[Zhnov]

plain Jane...

1. 4d6 twice, drop low die each set of 4, take highest of 2 sets.
2. Repeat #1 five more times, assign ability scores as wished.
3. Allow player to reroll if they wish.

Most settle for what they get the first time (not electing to reroll)--we have been rp'ing together since dirt was young, and while we all have some aspect of powergamer in us, it's usually satisfied by at least one attribute at 15+

We also use two additional 'flavor' attributes:

Comeliness -- yes, we are still using that from Dragon #2, or whatever. IOHR, it may be used to modify Diplomacy and a few other checks against the opposite sex, instead of Charisma.

Perception -- average of Int, Wis, Cha rounded down--ability score bonus is used on Spot, Search, Gather Info instead of standard attributes. ...yes, the "law of average attributes" dumbs down the bonuses for these skills... but that is a legacy rule under review.....

Orlic

 

[danzig138]

Why do you penalise higher level starting characters like that? WOuld you make him keep his scores if he could have made it 0 or better, but instead chose to raise diffeerent abilities which didn't reduce the penalty?

If I understand your question correctly, then by his distribution of points, his total mod would still be in the negatives, I'd probably allow a reroll of all scores. But if he wanted to keep them, with a total negative mod, I'd let him. I won't make anyone reroll. If a player likes his scores, even if they are low, I'll let him play them. I don't consider it a penalty for high level characters. That high level character was low-level at some point, and had to deal with lower scores, and choices have to be made. Now, my method doesn't work for everyone, especially those in the Bigger-Than-Life camp, but it's worked for me for years.

 

[ichabod]

Okay, here is your system, with means and medians for the
die rolls:

[color=silver]
Cost	#    Mean   Median
0    8     8.76   9
1    9     9.58   9
2    10   10.5   10.5
3    11   11.61  11
4    12   12.24  12
5    13   13.43  14
6    14   14.27  15
8    15   15.39  16
10   16   15.94  16
13   17        
16   18
[/color]

The thing that strikes me is that it's not consistent with respect to these statistics. The median is sometimes equal, sometimes higher. This is not as big a deal, as in all cases you have a 60-70% chance of getting the flat value or greater. However, the lower mean for 16 I see as a definite problem. It is the only case where on average you roll lower than the flat value. However, at that point it is rather hard to find a distribution with a mean of 16 and an average above 16.

Which is the heart of the problem. Ideally, you want a median of the value, and a mean higher than the flat value. However, that requires a right skew distribution. As you go higher, that means a tighter distribution. The tighter the distribution gets, the less you are risking by taking a die roll.

In short, I think it's a really cool idea that has practical problems, and you've probably implemented it about as well as can be done.

 

[totoro]

I favor point buy over all else, but almost everyone I play with prefers the random. Then, when the random roll is bad, most people want to reroll. It struck me as unfair that the worst roll you could make was just enough that a reroll was not meritted. So, if you roll really bad, you get another shot at it, but not if you roll slightly below average.

The system that seems to make the most players happy is point buy (28 points) followed by 4d6, drop lowest. Then you can choose to use either the point buy or what you roll.

 

[totoro]

The only problem I see is that rolling randomly is so much better than point buy. I guess that encourages players not to play it safe. Perhaps that is the goal?

Have you considered using different-sized dice to more accurately match the point buy costs?

Here are the averages for Herremann's method. I've included the set value (left) and the rolled value (right):

0 8 8.755401235
1 9 9.584
2 10 10.5
3 11 11.60546875
4 12 12.24459877
5 13 13.43016975
6 14 14.27379115
8 15 15.39453125
10 16 15.94628906
13 17 -
16 18 -

I've included a more in-depth Excel file for those interested.