Friday, July 16, 2021

Tweaking the ability score modifiers for importance and impact

As of last Wednesday, my group and I are eleven sessions into our current (red box D&D) mini-campaign. And in this time, as I always do when I run D&D, I've tried to pay attention to how the rules are working. I look for areas where the experience could be improved. I can't help doing that; I like to tinker.

And as this campaign has gone on, I've noticed something yet again that has bothered me in the past: I don't like the Basic/Expert ability modifier table that ranges from −3 to +3. I really don't like it. In fact, I kind of hate it—because it makes ability scores too important, high scores too desirable, and low scores too punitive.

Now, since this is just a mini-campaign that's already in full swing, I'm not going to implement any additional house rules mid-stride. This bit of theorizing is strictly to inform the next occasion that I run some form of OD&D. But it solidly reaffirms that I have to do something about that modifier table. ±3 is just too much variation, despite the expected rarity of extremely high or low ability scores. Such scores do occur, via both the dice and the acquisition of magical items (gantlets of ogre power, gloves of dexterity, amulets of health, periapts of wisdom, headbands of intellect, and cloaks of charisma), and they have a psychological effect on the players. They make the ability scores a focus of play in a way that seriously rubs me the wrong way.

In fact, it reminds me of what I hate about 5th edition D&D, where scores are central to play, high ability scores are downright essential, and even a mediocre score in a key area can cripple a character. It's a paradigm that happens to be totally at odds with the ability-scores-as-prime-requisites model,  generating scores on 3d6 in order, letting them inform (but almost never dictate) your choice of character class, and play decisions being the primary determinant of success. It's the first step on that slippery slope to bloated stats and slanted methods of character generation that eventually become character creation or the despised build.

But I digress. To briefly summarize my current thoughts on the matter of ability score modifier tables:

• The d20 System table, where modifiers for ability scores in the 3–18 range run from −4 to +4, is basically the worst of the worst. The ability scores have far too much mechanical impact on gameplay.

• The classic D&D table, which runs from −3 to +3, looks beautiful with its "exaggerate the already-present bell curve" distribution of modifiers: −3 at 3, −2 at 4–5, −1 at 6–8, no modifier at 9–12, +1 at 13–15, +2 at 16–17, and +3 at 18. But I now firmly believe that this is one of those all-too-common examples of an RPG rule that looks pleasing to the eye and pretty on paper, while also being actively detrimental to actual gameplay.

• The table I use in Engines & Empires (and which I'm also given to understand is used by Kevin Crawford in Stars/Worlds Without Number) is essentially the d20 System table halved and dropping fractions, so that modifiers run from −2 to +2. Specifically, they are: −2 at 3, −1 at 4–7, no modifier at 8–13, +1 at 14–17, and +2 at 18. This table has a lot to recommend it. I like that the no modifier range of 8–13 means that fully two-thirds of all naturally rolled scores will fall into this average band, with only one-sixth of scores having a bonus and one-sixth having a penalty, nearly always just ±1 (but scores of 3 and 18 are still as impactful as their rarity merits). I think this table is superior to the ±1 spread used by Swords & Wizardry. But, while it works well in Engines & Empires (with its four scores and very limited list of things that the modifiers apply to), I'm not certain it's the best fit for classic D&D.

• The Swords & Wizardry table is the simplest of all: −1 for scores of 3–8 (which is about 25% of rolls on 3d6), no modifier for 9–12 (about 50% of rolls), and +1 for scores of 13–18 (about 25% again). It's simple, clean, and doesn't burden the game with excessive stat modifiers. It's just… difficult to wrap one's head around, the notion that a Strength of 13 has the same mechanical impact as a Strength of 18, for anyone used to the very simulationistic idea that the 3–18 score range is supposed to represent the range of human variability. But this idea could very well be wrong, and maybe Swords & Wizardry has the right of it: maybe Strength 3 doesn't represent sickly invalids, and maybe Strength 18 doesn't represent Olympic-class weight-lifters. And there certainly doesn't need to be Super-Strength of 18/100% in the game to represent the likes of Samson and Hercules. Maybe scores of 3–18 are just there to represent some broad cross-section of human variability, and it's enough to sort everyone's abilities into below average, average, and above average, and call that good?

• Finally, of course, there's the mish-mash of tables we see in the LBBs and Greyhawk and Holmes (and retro-clones of the same). No uniform modifier table. Just bespoke modifiers that fit the particular sub-systems they're modifying, without letting things get as overly-detailed and unwieldy as AD&D. There's something that I instinctively like about this approach, but I don't think that I can embrace it fully yet. Instead, I'm more inclined to come up with a compromise between this method and the Swords & Wizardry method. The bespoke, organic, "perfect fit" of Holmes meshed with the clean simplicity and low impact of S&W seems, to my mind, ideal.

To that end:

Here's how I think I'll revise the six ability scores the next time I run D&D.

Ability Score





−1 to melee hit & damage

−1 to Armor Class

−10% hp


−1 to open doors

−1 to hit with missiles

−5% hp






+1 to open doors

+1 to hit with missiles

+10% hp


+1 to melee hit & damage

+1 to Armor Class

+20% hp

Ability Score






−1 to find secret doors

−1 to reaction rolls


Literate (Common)

−1 to saving throws

−1 to followers & morale


Literate (Common and Alignment Tongue)

(base 2-in-6 to find secret doors)

(4 followers, ML 7)


+1 Bonus Language

+1 to saving throws

+1 to followers & morale


+2 Bonus Languages

+1 to find secret doors

+1 to reaction rolls

Ability Score

Prime Requisite (one per class)

Secondary Requisites (two per class)


−20% to earned XP


−10% to earned XP




+5% to earned XP

Virtual +1 to prime requisite for XP only


+10% to earned XP

Virtual +2 to prime requisite for XP only


+10% to earned XP

Virtual +3 to prime requisite for XP only

Taken altogether, these tables represent something of a culmination of the house rules I've discussed on this blog in recent months. In particular, I've retained the house rule that I cooked up earlier for leaving prime requisite ability scores unmodified, but "virtually adjusted" by a character class's "secondary" requisites (which is, in the end, an instance of that +3 modifier surviving after all, albeit in a very low-impact form). Also, for the vast majority of classes, the secondary requisites are Int and Wis, and it makes sense to have these stats—particularly Int—impact the rate at which experience points are gained!

Just sitting back and considering these tables, I think they're almost as clean as Swords & Wizardry, almost as flavorful as Holmes Basic or Greyhawk, and decidedly functional. They do to D&D characters more or less what I want them to do. There's a good range of variation, but nothing excessively overpowering or punishing. 

Guess I'm going to have to another page to my D&D rules document after all… ∎

• • •

FOOTNOTE: It occurred to me a bit later that a house-rule like this would still necessitate a slight tweak to stat-enhancing magical items, to make them suitably desirable and effective. Items like gauntlets of ogre power and such ought, under a system like this, to both set the relevant attribute at 18 and magically double any bonuses granted by such a high score. So, for example, a character wearing gauntlets of ogre power would enjoy +2 to open doors rolls, melee to hit rolls, and melee damage. An amulet of health would boost a character's hit point total by +40%. A headband of intellect would impart to a character a total of four bonus languages. And so forth.


  1. I like it. I always felt the +3/-3 modifiers are too big of a spectrum and tend to direct game into flat and shallow ability dependency which is 5e's biggest problem.

    1. 5e uses the same ability score table as 3e and 4e, but stat dependency is markedly worse in 5e because of bounded accuracy. In 3e and 4e, a character's experience level still does a lot of heavy lifting, but in 5e, it was deliberately decided that the ability scores would do most of the mechanical heavy-lifting. And the effect is a focus on the scores that really detracts from player choice and self-determination.

      (Old-school D&D can be just as bad, when a DM relies on roll-under d20 ability checks for task resolution. I prefer to avoid that mechanic like the plague that it is.)