In Pathfinder point buy, ability scores of lower than 7 (before adjustments from race) may not be bought with point buy.

What would be a balanced price for such low ability scores?

I am looking for personal experience from people who have done such an extension, or for a reasonably accepted view in the character optimization community.

Motivation for the question: How likely is to roll elite array or better with 3d6 in order?

In particular, I am not worried about players creating horrible malformed monsters with only threes and eighteens as their ability scores.

The question Is it allowed to buy ability scores lower than 7? addresses whether buying low ability scores is allowed (no), which is not the question here. The answer also argues that allowing buying such scores is a bad idea. My use case is very different - I want to have a method of measuring how good or bad ability scores are, and allowing anyone to buy them is not an issue here.


2 Answers 2


The formula is that you have to pay for an ability score the new modifier. So going from 7 to 6 gives you a price of -2, 6 to 5 and 5 to 4 each give you -3, and so on. Dropping an ability score to 3 would give you 16 points.

Score | Mod | Cost  
   10 |   0 |    0
    9 |  -1 |    1
    8 |  -1 |    2
    7 |  -2 |    4
    6 |  -2 |    6
    5 |  -3 |    9
    4 |  -3 |   12
    3 |  -4 |   16

It's uncertain whether those values would be balanced ; while they follow the pattern of ability score costs, nothing says that this pattern won't fall apart outside of the 7-18 range. Furthermore, in any case, this should never be an option available to players as discussed in the answer to this question.

  • 1
    \$\begingroup\$ This is better than a pure guess, but I would be surprised such a symmetrical result were true. Do you have play experience or serious character optimization people taking a look at this? \$\endgroup\$
    – Tommi
    Apr 7, 2018 at 18:38
  • 1
    \$\begingroup\$ Negative. I'm highly certain that the formula is correct, but as mentioned, nothing says this formula will work outside of the 7-18 range. Especially considering that standard roll is 4L6 and not 3D6. Sorry if I made it sound more confident than I am. \$\endgroup\$
    – Bielna
    Apr 8, 2018 at 8:23

If I were to houserule it, I would say that going below 7 grants you one extra point per reduction(or possibly less!). The reasoning is the same as it (presumably) is for there being such a limit in the first place: If you have a 7(or a 5, since someone who reduces their ability score to such degree in point buy may well go all in and pick a race that further penalizes it), the score might as well be 3 for most purposes and you're just not going to use it for anything unless absolutely necessary: A wizard with 7 strength isn't going to be making melee attacks, and if someone grapples you, you're pretty much automatically going to fail any opposed checks. A barbarian with 7 charisma is not going to make social skill checks - he'll let the bard do the talking for him. And a fighter with 7 intelligence is going to take his one skill point per level and ignore anything that requires intelligence based skills(constitution, dexterity and wisdom people aren't going to reduce too much anyway, because every character needs them to a degree - unless someone is planning to become undead). There are some marginal downsides, of course(carrying capacity and ability damage come to mind), but except maybe carrying capacity, they're not something you're really going to worry about (unless you know your GM is going to make sure to engineer situations where your weaknesses come into play).

  • \$\begingroup\$ Thanks for the answer. Have you actually tried this in play or could you refer to some people who have, or is there otherwise a general consensus that this is the right answer? \$\endgroup\$
    – Tommi
    Feb 11, 2021 at 14:46
  • 1
    \$\begingroup\$ I have not tried this, nor do I think there exists a consensus because, as has been pointed out, extending point-buy range downwards is generally considered to be a bad idea. For your usecase, I would consider 1-to-1 ratio to be adequate approximation, however. \$\endgroup\$
    – Spodah
    Feb 12, 2021 at 8:09

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .