a new method for character creation

F

frisbeet

First Post
This went over like a lead balloon on the Wotc boards, so I'm modifying it. Maybe it's still too avant-garde/avant-crap.

Assume that everyone in your 3e D&D world have abilities which follow the 3d6 generation method distribution. If this is the case, then doing a litttle math, you'll find that on average, in 36 encounters with NPCs, one of them will have at least one 18 for an attribute, randomly assigned. That's different from saying that one in 36 will have an 18 Str. If you do a little more math, you find:

at least ?? 18(s), randomly assigned___probability___1 in

1___2.75E-02___36
2___3.18E-04___3149
3___1.96E-06___509169
4___6.84E-09___146199694
5___1.27E-11___78667665893
6___9.85E-15___101559956668416

So if you bumped into 101.6 trillion NPCs, one of them is probably going to have an 18 for every attribute. Somehow I doubt your world is that big.

Now, what about those NPCs with an 18 Str?

?? 18(s), exactly assigned___probability___1 in
1___2.75E-02___216
2___3.18E-04___46656
3___1.96E-06___10077696
4___6.84E-09___2176782336
5___1.27E-11___470184984576
6___9.85E-15___101559956668416

So one in 216 will have an 18 Str.

Here's my idea, given that characters are supposed to be chosen from a cream-of-the-crop distribution (ie 4d6, point buy, etc.):

Want an 18 Str?

Great! Now role 3d6 for your other attributes.

Your character has been selected from a special pool of people, better on average than 215 comrades in one attribute. That's not such a stretch, it it? I mean, if you're reading this forum, and especially if you play D&D, you're already in an elite subset of people, right?

Criticisim welcome.
 

log in or register to remove this ad

I let the PCs roll 4D6, adding up all 4 of the dice to make sure they don't have a lot of weaknesses. It makes for a more rounded character with a player that doesn't feel like his character is inadequate in some respect. Granted, I want the party to feel and act like a team to complete adventures, but sometimes it doesn't happen so simply.

Can you work up the probability of a character with an 18 for one stat and at least 9 for the other 5? Would it be around 1 in 6,912?
 


CRGreathouse said:


I get much higher -- 1 in 969, or 0.103%.
Click to expand...

Yup, that's right.

P(18)*P(>8)^5

=(1/216)*(160/216)^5

160 is the # of ways of rolling 3 dice for a sum greater than or equal to 8. 216 is the # of ways of rolling 3 dice.

Every frequency suggests a new method. Are you comfortable with playing a best in 1000 character? Then assign an 18 to one score and roll 3d6 for the other scores, rerolling any results <8. Or, if you really want to get crazy, play a best in 50000 (46656) character: assign two 18s and roll 3d6 for the other scores, keeping low results.

This is where I think most people start rolling their eyes. But hey, it's your fantasy world. Put it another way: if you're going to invest over $100 in this game (and then WotC have the frickin audacity to shove 3.5 down our throats, after we've bought all the supplemental stuff), play the game the way you (and your DM) see fit. If you wanna drive the Ferrari in fantasy (two 18s), then you should--even as you stipulate that all the NPCs in your world drive something far less cool (3d6 for all scores). Don't worry about others here and elsewhere who scream about how unbalanced, unfair, blah blah blah it is--that's only true in their fantasy world. It was your money that bought you the game, and your time invested. mmkay, Rant.
 

frisbeet said:

Want an 18 Str?

Great! Now role 3d6 for your other attributes.

Your character has been selected from a special pool of people, better on average than 215 comrades in one attribute. That's not such a stretch, it it? I mean, if you're reading this forum, and especially if you play D&D, you're already in an elite subset of people, right?

Criticisim welcome.
Click to expand...

I do something similar... If you want that 18 stat you can have it. 4 of your other scores is rolled with 3d6, and the last is rolled using 3d4, meaning the character is *very* likely to have a subpar stat to go with that 18.

I also let players use the 4d6 method. Most players opt out for a 4d6 approach.
 

I allow the players to choose between the following methods:

Two stats at 14, four stats at 13 (before racial modifiers) - simplified point buy for newbies.

32 point buy system as laid out in the DMG.

Roll 4d6 7 times. Drop the lowest roll each time, drop the lowest total, no reroll of ones. Assign to abilities as you like. May reroll if the total is below 80 (that's what the simplified point buy yields).

GM generated stats, for complete newbies. I generate the stats race and class and hand it to the PC.

Again the PCs get to choose the method used. I've had PCs use all four offered and I've had PCs be happy and upset with all the methods offered.
 

IIRC, the last edition of D&D that explicitly said population stats were distributed according to the 3d6 curve was 1E.
 

hong said:
IIRC, the last edition of D&D that explicitly said population stats were distributed according to the 3d6 curve was 1E.
Click to expand...

Well, I think it's a very very safe assumption that NPC attributes in the D&D world are Normally distributed about the population mean. I know it's fantasy, but it strikes me as absurd and cartoonish if this wasn't the case. Does this mean 3d6 is the right distribution? No--maybe the variance is a lot tighter. Or the mean should be 8-9, not 10-11. But given that 10s and 11s have no associated penalty or bonus, it seems logical to put the mean there. Then among the distributions available to us: 3d6, 4d6, 5d6, point buy (a very funny distribution), only 3d6 is Normal (symmetric) about its mean, which happens to be 10.5. So while it doesn't explicitly state that the 3d6 distribution is so, I'd argue that implicitly, it is. Make it so in your world, if you want.
 
Last edited:

frisbeet said:
Well, I think it's a very very safe assumption that NPC attributes in the D&D world are Normally distributed about the population mean. I know it's fantasy, but it strikes me as absurd and cartoonish if this wasn't the case. Does this mean 3d6 is the right distribution? No--maybe the variance is a lot tighter. Or the mean should be 8-9, not 10-11. But given that 10s and 11s have no associated penalty or bonus, it seems logical to put the mean there. Then among the distributions available to us: 3d6, 4d6, 5d6, point buy (a very funny distribution), only 3d6 is Normal (symmetric) about its mean, which happens to be 10.5. So while it doesn't explicitly state that the 3d6 distribution is so, I'd argue that implicitly, it is. Make it so in your world, if you want.
Click to expand...

Chapter 1 of the PH states that the typical NPC scores are 10-11 (at least twice). I'm not sure if the 3d6 rule is explicit anywhere, though.
 

Remove ads

Top