The 5% Principle

If we look at the 1E AD&D Dungeon Master's Guide combat tables, we see a table for Fighters (and rangers, paladins, and bards) which improves by 2 points every 2 levels. Immediately below it, there is this note:

Special Note Regarding Fighters' Progression: This table is designed to allow fighters to advance by 5% per level of experience attained, rather than 10% every 2 levels, if you believe that such will be helpful in your particular campaign. If you opt for a per level advancement in combat ability, simply use the table but give a +1 "to hit" bonus to fighters who attain the second level of experience shown in each group of 2 levels, i.e. 1-2, 3-4, etc. You may, of course, elect not to allow per level combat advancement. [DMG p. 74]

I bring this up partly because some people overlook it, and partly because it provides credence for Gygax being amenable towards combat charts with a "smooth" progression in them. In fact, we might hypothesize that the small, "jumpy" charts we see in the classic game are merely a concession to the limited page space available in OD&D and AD&D. A mechanic where we check d20 + level + AC ≥ 21 would perform the task identically, without any need for table lookups (i.e., one pip difference from what I call "Target 20"). This idea is further expanded upon in a Dragon Magazine article by Len Lakofka titled "New charts, using the '5% principle'", which begins thus:

© 1983 E. Gary Gygax & Lenard Lakofka
The following material is not official, but is provided for your study and comment. Gary Gygax has said that an expanded combat results table is certainly desirable, so perhaps that part of the following information will eventually be made part of the official rules. However, the suggestions on how to change the experience-point chart are entirely of my own devising. [Dragon #80, December 1983, p. 48]

Thereafter, consolidated charts are presented for attacks and saving throws which increment by exactly 5% (1 in 20) any time an improvement is made -- interpolating the existing tables in the AD&D DMG. This gives at least the impression of Gygax approval, due to (1) the conspicuous copyright notice, (2) the Gygax attribution, and (3) the distinction from the experience tables at the end of the article. I present the former tables below.

Would I want to use these tables? Actually, no -- I think they've become too complicated visually. Once you have more than 7 categories per axis, a person's speed at cross-referencing is sure to fall off. The attacks table here is exacerbated by 6 different lines of numbers across the x-axis (the classes) that you need to distinguish between on a case-by-case basis.

What I really want to argue is that the fundamental intent behind these tables would, in fact, be better served by a simple, much shorter, numerical core rule, but for some reason the Gygax/Lakofka team lacked the mechanical creativity to invest the game with such. While it's been argued that tables can present complicated mechanics in an easy-to-reference fashion, the fact is, there's nothing complicated here for the tables to simplify -- the presented progressions are all inherently linear, and tablature may in fact be about the clunkiest way to express the desired, smooth system seeking a "5% principle".


  1. Total agreement + fascinating discovery of something that had slipped past my radar.

  2. Well, our DM used that table. He even provided each player with his own photocopy for inclusion in their PHBs.

  3. We did use the 5% rule as presented in the DMG, though IIRC we often used it for fighters only and not Rangers or Paladins.

  4. While it's been argued that tables can present complicated mechanics in an easy-to-reference fashion, the fact is, there's nothing complicated here for the tables to simplify

    However, a tabular presentation can make patterns more readily apparent and allows for at a glance comparisons... especially the old chunked tables in the DMG.

    Once you drop the table and have the single number, you go on the add attribute bonuses and then magical weapon modifiers and suddenly you have to deconstruct the final number and reconstruct a table in order to retrieve certain kinds of information.

    That is to say, tables aren't necessarily there to simplify. They can be there to reveal.

  5. Vincent said: "However, a tabular presentation can make patterns more readily apparent and allows for at a glance comparisons... especially the old chunked tables in the DMG."

    Definitely have to disagree with that. "d20+L" is far more communicative than a 17x20 matrix of numbers. Confirmation of that relationship can't really be done "at a glance".

  6. FWIW, the number 7+/-2 (i.e., 5-9) isn't really applicable in visual page/web site design any longer (or more likely, it was misapplied in the first place), so it probably doesn't really apply to the table columns/structures of the DMG, either. See Edward Tufte and Miller himself @ http://www.edwardtufte.com/bboard/q-and-a-fetch-msg?msg_id=0000U6 if you're curious.


  7. grodog, I'm aware of the critique and in this case it doesn't get much traction from me. A table like this deals with discrete data, whereas Tufte tends to be dealing with continuous data (and therefore has more smooth options available).

    Admittedly I'm only using this as a very rough guideline; nonetheless, more categories are harder to deal with, and my opinion is that the Dragon #80 tables are too much. Especially so when the whole thing could be replaced by "d20+L" or the like.

  8. Yup. This is one of those places where upon examination replacing the tables with a formula or BAB system really isn't much of a change, other than smoothing out basically the same progression.

  9. I can't help but wonder if there wasn't also a belief that a table as a visualization, had more impact than a formula, and therefore more useful. Table commands attention, where a formula often gets disregarded. In that that, I think the early reliance on tables was critical for making RPGs engaging. I vaguely recall games that included formulas for various operations instead of tables, but have no recollection of every using that information.

    Also, considering the standard at the time, in academia and war-gaming, was to consult tables. Even the slide-rule, the defacto calculating machine of the time, was nothing more than a table disguised as an elaborate variable (sliding) mechanism, so relying on tables instead of easily overlooked formula probably made perfect sense, through the lens of the time.

    1. Yeah I agree with most of that -- certainly the tradition in wargaming was to use tables, which makes the most sense when you started with actually military-science research on hit and penetration rates, and the functions are inherently nonlinear.

      But I do think for the sake of the linear D&D mechanics it was time to make the switch and not need those tables, which in practice slows everything down compared to the alternative. They were already fiddling with text formula and THACO in the 1E DMG but couldn't quite find the right, simple mechanic.