Worlds of Design: Rolls vs. Points in Character Building

Let’s talk about methods of generating RPG characters, both stochastic and deterministic.

Let’s talk about methods of generating RPG characters, both stochastic and deterministic.


Picture courtesy of Pixabay.
"Life is like a game of cards. The hand you are dealt is determinism; the way you play it is free will." Jawaharlal Nehru

When creating character attributes, there are two broad approaches to generating them: stochastic and deterministic. The stochastic method involves chance, while the deterministic method does not. Most any other method is going to be one of the other, whatever the details. The pros of one method tend to be the cons of the other.

The classic method is rolling dice, usually D6, sometimes an alternative like percentage dice. There are various ways do this. For example, some of the old methods were to sum the roll of 3d6 six times in a specific order of six character abilities. A variation was 3d6 and change the order as desired, another was roll 4d6, don’t count the lowest die, and then you might be able to change order or not; and so forth.

What are the pros of rolling the dice? First of all and primarily, variety (barring cheating). You get a big range of dice rolls. Dice rolling promotes realism, you get a big variation in numbers so you get some 3s, in fact you get as many 3s as 18s, and with some methods you have the opportunity to play characters with “cripplingly bad" ability numbers. Further, it's always exciting to roll dice, whether you like it or not. (Keep in mind, when I first saw D&D I said “I hate dice games.”)

One of the cons of rolling dice is that it's unfair in the long run, a player can get big advantages lasting for years of real-time throughout the campaign just by getting lucky in the first dice rolls. This can be frustrating to those who didn't get lucky. Perhaps even more, rolling dice encourages cheating. I've seen people roll one character after another until they get one they like - meaning lots of high numbers - and then they take that to a game to use. That’s not possible with point buy. Another con is that you may want to play a particular character class yet the dice just won’t cooperate (when you’re rolling in specific order).

The other method which I believe has been devised independently by several people including myself (I had an article for my system published a long time ago) is the one used in fifth edition D&D. A player is given a number of generic points to buy ability numbers. The lowest numbers can be very cheap, for example, if you are using a 3 to 18 scale, when you buy a 3 it may cost you one point, while an 18 may cost 20-some points. You decide what you want, for which ability, and allocate until you run out of points.

Point buy is very fair (FRP is a game, for some people). No one need be envious of someone who either 1) rolled high or 2) rolled many characters and picked the best one. It prevents the typical new character with sky-high abilities, it prevents cheating, so the player has to supply the skill, not rely on bonuses from big ability numbers. Of course, the GM can choose the number of points available to the players so he/she can give generally higher or lower numbers on average as they choose.

But point buy lacks variety for a particular class. The numbers tend to be the same. It's not exciting, it’s cerebral, and as such it takes a little longer than rolling dice. That's all the cons I can think of. Keep in mind I'm biased in favor of point buy. It's clean, fair and simple.

I haven’t spent much time trying to figure out yet another method of generating a character. The only other method I can think of that isn’t one or the other is to have some kind of skilled contest determine the numbers, such as pitching pennies or bowling. Then the question becomes why use one kind of skill over another?

Do you favor one method over the other? And has anyone devised a method that is not stochastic or deterministic?

log in or register to remove this ad

Lewis Pulsipher

Lewis Pulsipher

Dragon, White Dwarf, Fiend Folio

Jacob Lewis

Ye Olde GM
Rolling for attributes is exciting for the five minutes you spend hoping to beat the odds. After that, you'll spend the rest of the campaign with the outcome, good or bad, and eyeing everyone else's sheet to compare their results. Nobody ever really goes into this method hoping for less than or average numbers, and many will come up with numerous ways to circumvent the odds to ensure better results anyway, like more dice, discard low numbers, arrange the order, etc.

Point buy makes more sense to me. Everyone is given the same pool of resources and freedom to make decisions for themselves. If players want to play with low scores, it's determined by their choices, not chances.

As an active D&D 5e DM, I don't use dice rolling for ability scores in my game. Firstly, on average, what does it really get you? If you use the 5e standard array of [15,14,13,12,10,8] as representative of a typical point-buy character (remember, the standard array is built using the 27-point point buy method), that gives you 72 total ability score points. If you roll 4d6 (drop the lowest) six times, which I suspect is the most common method of rolling ability scores in D&D, you end up with an average of approximately 73.5 total ability score points. Frankly, there isn't much difference there.

Most importantly, though, it puts too much weight on the DM. If a player rolls substandard scores, the weight falls on the DM to decide that the character is unplayably weak. If a player were to roll all 3s, that's a no-brainer. Reroll. If a player were to roll all 5s or all 7s, also a no-brainer. Reroll. What if the player rolled all 10s, though? Not an exciting D&D character, but is it unplayable? Surely, the player wouldn't want to play that character and would be looking to his or her DM to nix it and allow a reroll.

What about on the other end? What if the player rolled all 18s? Is it okay for the DM to say that the character is too powerful and would unbalance the campaign? Imagine if you accomplished the miraculous, staggeringly improbable act of rolling all 18s for a character and the DM told you that you had to reroll.

So if we're asking DMs to adjudicate a rolled character as being either too weak or too powerful, what we're really saying is that we need a mechanism for determining the acceptable upper and lower bounds of character ability scores. Well, we already have that. It's call a point buy method.


Well, that was fun
Staff member
Point buy is certainly fairer. However, you get a very limited number of character builds -- they all have a maximum primary attribute, and they all tend to look the same.

I do agree, though, that that similarity is outweighed by the unfairness of a single unlucky roll at the start of a campaign affecting you for the entire campaign.

Rolling ability scores is very exciting, and full of possibility and wonder. The tension is high. It's like being a kid, when the world is full of infinite variety, the opportunities endless and the future unknowable. It's like opening a present, when you don't know what's inside. But it only lasts a few minutes and then you have those scores for months.


"Diegetics", by L. Ron Gygax
There are definitely methods that split the difference between the two, a kind of stochastic with boundaries.

Off the top of my head, Beyond the Wall's playbooks generate stats that add up to the same values, but are varied by what you roll during the playbook. Like the first table of character creation gives you a total of +4 to your stats, but the actual stats that increase depend on the entry of the table.

One method I've used before is roll with boundaries. You roll 2d6+5 five times, and the 6th stat is 75 minus the sum of the other 5 stats. (You do some adjustments is the last stat is out of bounds for a viable stat.)

Another method I've tried recently (suggested by @FaerieGodfather) is point-buy combined with rolling. You do standard point-buy first for your character. Then you roll 4d6kh3 6 times in order, and you keep whichever value is higher for each stat. Fairly high powered, but everyone has at minimum a viable character and sometimes you get some lucky rolls that make you re-evaluate your concept.


I have to admit, I'm far more in the point buy camp now. Die rolled characters just don't really do it for me anymore.

Now, in 5e, since ability scores don't actually make a whole lot of difference, it really doesn't matter that much. Whether you've got a +8 to attack or a +10 isn't going to matter often enough that it is going to bother me.


Small God of the Dozens
I have enough experience playing characters with truly awful stats that I have no great desire to back. Not for a long-term campaign anyway. For shorter games or one-offs its a little different. I also have less issues with rolling in OSR and similar games where that's a key part of the experience that you've bought into before you start.

Rolling is fun though, there's no doubt about that. Hmm, now that I'm thinking about it, rolling would be cool if for every X number under a set total stat number you got something else, like a feat. SO for every three or four points you are under the total you get a feat or something from a list, and totals under X (something really bad) get binned. That could be fun too.


I prefer rolling dice and using the results to create the concept and background for the PC.

In my most recent campaign as a player in DCC I rolled a 5 for my Agility and a 5 for Personality.

Ian, my handsome halfling with a wooden peg leg and absolutely no filter to his blunt opinions is among the most fun PCs I’ve ever played. :)


I have the BEST way, it involves some randomness but statistically binds the stats together, so in the end the power curve looks a lot like point buy. It is literally the best of both worlds, all my groups use it now.

The cards:
Take Ace-6 cards from a normal deck.
Deal into 6 piles
Drop lowest from each pile (ace is a 1)
You are done.

The result is that you can have an 18 or a 3, but you only have 4 6's so you will never have two 18s or two 3s. One bad stat makes the others more likely to be high. One great stat makes the others more likely to be low.
All the gambling fun of random, but with a much better statistical distribution so all players are on an even playing field.

Voidrunner's Codex

Related Articles

Remove ads

Voidrunner's Codex

Remove ads