Card Based Stat Gen: A more equitable random stat gen method

Recently, the topic of random vs. point buy ability score generation has come up again. A large segment of the D&D is really enamored with it, but I find the sorts of over-optimized and/or repetitive character builds point buy often engenders to be antithetical to my playstyle.

That said, I do agree that it really sucks to be the guy who rolled crappy in a party of players who rolled well.

Though there are a couple of methods I came up with to deal with this, here's one I came up with that I have stuck with for a while because it's great for Play-by-Posts as well as for making pregens for gamedays/conventions.

This was originally posted here, but there are a few updates that introduce some tweaks I have been using lately.

Note that this was intended for use in D&D 3.5, D20 Modern, and Spycraft 2.0. I have no idea how suited this is for D&D 4e; use at your own risk.

Psion's D20 System Game Card Generation Method

1) Get a deck of standard playing cards. Take the 4-9 cards of two suits out of the deck. This should give you 12 cards, 2 of each numbered 4-9.

2) Shuffle the 12 cards. Deal them out in pairs. Flip them over and total the pairs. This gives you 6 numbers that can range from 8 to 18.

(If you are happy with this, just allocate these scores to the 6 stats at the player's discretion. I, however, was shooting for a slightly higher average to match the assumptions of Spycraft 2.0 and D20 Mars, and I wanted to avoid 8's, so I added this step.)

3) Sort the pairs from highest to lowest. Add 1 to the 2nd, 4th, and 6th highest numbers. Allocate statistics to ability scores to taste.

Note:
In Spycraft, to compensate for the fact that some players have more odd scores than others, I allow players to subtract 1 from two odd scores in exchange for 1 bonus feat, similar to Spycraft's Modular campaign quality.

In D20 Mars, you get more frequent attribute boosts than in standard D20 games, so odd scores are less of a "sunk cost."

Example:

The cards are shuffled and dealt, resulting in these pairs...

6, 5 ( = 11)
9, 4 ( = 13)
7, 7 ( = 14)
5, 9 ( = 14)
6, 4 ( = 10)
8, 8 ( = 16)

The scores are ordered, and 1 is added to the 2nd, 4th, and 6th highest
16
14 + 1 = 15
14
13 + 1 = 14
11
10 + 1 = 11

Resulting in
16
15
14
14
11
11

"But I want to be able to pick my stats!"

Okay, these sorts of players are still out there. So here's an option I provided: instead of shuffling to get pairs of cards, allocate them to taste. However, you have these additional restrictions (to make things more fair to the randomizers):
1) You may not deal yourself an 18 or 2 17's.
2) You don't get the 3 "kicker points" from step 3 above.

Comments on method

Having been using this for a bit, I have observed a few properties that differ here from random dice rolling.

Obviously, by intent, the spread between PC power is much less than random, but unlike point-buy, you don't see pattern builds and a flurry of ugly fighters and you don't see the big point cost hit for higher stats.

While dealing the cards to make scores, it's psychologically a different experience than rolling. If you roll a crappy score using a dice method, there's the definite feeling that you have been set back, because you can't expect the rest of your scores to make up for it. When using the card method, you deal out some low cards, you know a good score is coming, because those cards are out of the deck.

The main randomization between characters is that some characters will have higher peak stats than others. But that's okay by me. I find forcing a spate of hyper-optimized characters is detrimental to character variety. By giving the player a higher score, you force them to consider credible sidelines. This sort of adds some variety to character design.

UPDATE: Excel Tool

Irda Ranger has put up a quick excel tool to generate characters using this method. Check it out in this post in the original:
http://www.enworld.org/forum/d-d-3rd...ml#post4483802