OSR: Converting D&D to Utilize d6s Instead of that Blasted d20

I have recently been on an OSR D&D kick recently, and it has come to my attention that the system is basically a d6 engine. Consistent through several elder editions:

• Ability scores are generated via rolling 3d6.

• Starting gold is calculated via rolling 3d6.

• Reaction rolls are determined via rolling 2d6.

• Morale rolls are determined via rolling 2d6.

• Surprise is determined via rolling 1d6.

• Initiative is determined via rolling 1d6.

• Forcing open and listening at doors are determined via rolling 1d6.

• Searching for traps is determined via rolling 1d6.

• Accidentally setting of a trap is determined via rolling 1d6.

• Wandering monsters are encountered via rolling a d6.

• Undead are turned or destroyed via rolling 2d6.

• Hear noise (thieves) is tested on a d6.

• Dwarves detect sloping passages, traps, and new construction with a roll of 1d6.

• Adventurers may detect secret doors via a roll 1d6.

• Resting restores 1d3 Hit Points (rolled on a d6).

• Starting encounter distance is determined via rolling 2d6.

There are two areas in which d6s are not utilized:

• Attack rolls and saving throws, both of which use a d20.

• Damage and Hit Dice/Point rolls, both of which utilize other polyhedrons.

Presumably, these utilize non-d6s for one reason or another (I suspect both were chosen as an attempt to offer more granularity within the system), but I am nonetheless pondering: why bother? The d6 affords several benefits over the d20. It is ubiquitous, it resolves more swiftly than a d20, it is used throughout the rest of the system, and its small range of numbers allow for modifiers to have a greater impact upon its outcomes (e.g., bless, instead of being a paltry +5% chance of success, becomes a +17% chance of success).

All of these are objective improvements, save for the last one (which is a subjective improvement, one that I subjectively perceive as nearly objective).

Armor Class Essentials

Translating the existing system of attack vs. Armor Class is a bit tricky. Not impossible, but tricky. Each +1 on a d6 translates to about a +3 bonus on a d20, which means rounding is in order. (As an aside: if Armor Class is better the lower it is, characters should simply roll 1d20 under the enemy's Armor Class rather than consulting a matrix.)

Before we go any further, let's deal with the existing ranges of Armor Class in B/X D&D (what I view as a refinement of OD&D, despite it missing some valuable bits and bobs.)

The worst and best possible mundane (non-magical) Armor Classes, are delineated below, with brackets converting them to ascending Armor Class:

• Worst: 12 [7] AC.

  • Unarmored: 9 [10] base.

  • Dexterity adjustment: +3 [-3].

    
    

• Best: -1 [20] AC.

  • Plate Mail + Shield: 2 [17].

  • Dexterity adjustment: -3 [+3]

• Unarmored: 9 [10] AC.

  • No Dexterity adjustment--the average 9-12 Dexterity is assumed.

This likewise provides us a sensible range based upon d6s.

Armor Class Conversion I: Divide by 3

With each 3 points of Armor Class equating to 1 point on a d6, we can simply divide these by 3. If you prefer doing things the long way, calculate one's existing Armor Class and then divided it by three, rounding upward or downward. For instance, a character wearing Leather Armor with a Dexterity score of 13 (1-point of AC improvement) would have an AC of 6, translating it to an AC of 2.

Using the limits as noted above:

• Worst AC: 12 [7] becomes 4 [3].

• Best AC: -1 [20] becomes a 0 [6].

• Unarmored AC: 9 [10] becomes 3 [4].

The attack matrix in B/X includes Armor Classes of up to ("down to") -3 [23], which is trivial: -1 [7]. You'll note that the numerics here for ascending Armor Class are less clean because they start at 10 rather than 9.

This then becomes a matter of patience; the process of converting an ability score into a modifier, then calculating one Armor Class, then converting that value into another value is tedious.

There is a further thorniness raised in that existing Armor Classes are incremented by 10% for each "class" of armor (shields being 5%) whereas the smallest increment of a d6s is 16.67%, meaning that the benefits of armor will necessarily be diminished or enhanced, depending on the rounding choices one makes.

I lean toward rounding to worsen AC until certain "breakpoints" are reached; that is AC 9 [10] converts to AC 3 [4], and so to do AC 8 [11] and AC 7 [12]. It is only at the threshold of AC 6 [13] that the conversion improves to AC 2 [5].

Armor Class Conversion II: Simple d6

Now, all that being said, let us consider an alternative (and simpler) method wherein we set "unarmored" to a default value then improve it by increments of 1.

• Unarmored: 3 [4]

• Leather: 2 [5]

• Chain: 1 [6]

• Plate: 0 [7]

• Shield: -1 [+1]

That, of course, is the simplest and fastest solution to this problem, but it does run into an issue whereby the listed Armor Classes run into impossible values: namely, rolling a 0 or 7 on a d6. This issue is rectified by simply saying that a 6 automatically hits any Armor Class (much as an unmodified 20 on the d20 does), but it leads to an awkward situation where there is no purpose to wielding a shield when wearing chain or plate (at least not until later levels).

I would rectify this in part by worsening unarmored AC by 1 and adjusting thusly:

• Unarmored: 4 [3]

• Leather: 3 [4]

• Chain: 2 [5]

• Plate: 1 [6]

• Shield: -1 [+1]

This doesn't fully resolve the difficulties, but I think it won't matter overmuch in actual gameplay. Reducing AC across the board will make wizards and thieves a bit squishier, however.

Armor Class Conversion III: Attack Matrices

The earliest editions of D&D used attack matrices rather than formulas: THAC0 was an invention for 2e, and 3e converted this to Base Attack Bonus, which 4e standardized into a half-level bonus + weapon proficiency, which 5e has now simplified into proficiency bonus.

There is a formula here; 1st-level characters must roll 19 - the defender's Armor Class or higher on a d20 to hit. Given that the modern player is mystified and enraged at the suggestion of THAC0 and struggles to add an ability score modifier + proficiency bonus to a d20 in 5e, I think concealing the formula and just showing a matrix was the smart thing to do. This way, players don't need to read the rules (because they seem allergic to doing so).

To convert this to a d6 system, we can start with the easy stuff:

• 1s always miss (the equivalent of an unmodified 1 on the d20).

• 6s always hit (the equivalent of an unmodified 20 on the d20).

Yes, this alters the probabilities significantly, tripling the rates of an automatic hit or miss. I do not view this as a negative--even if a character automatically hits 1-in-6 times, that's still a very low probability. Doubt me? Wager me $100 and roll a d6. On a 6, you win $200. On a 1-5, I pocket the cash.

Now that we've settled that matter, let's move onto the rest.

• AC 9 [10] has a 55% hit chance (10+ on the d20).

• AC 5 [14] has a 35% hit chance (14+ on the d20).

• AC 1 [18] has a 15% hit chance (18+ on the d20).

These values are important because they roughly correlate to certain digits on the d6:

• AC 9 is 4+ on a d6.

• AC 5 is 5+ on a d6.

• AC 1 is 6+ on a d6.

Now we need to fill in the blanks.

• AC 11 [8] has a 65% hit chance (8+ on the d20, or 3+ on the d6).

• AC 14 [5] has an 80% hit chance (5+ on the d20, or 2+ on the d6).

We have the baseline for our attack matrix. For our everyman, it looks like this:

d6 Roll and Lowest [Highest] Armor Class Hit
1	2	3	4	5	6
miss	14 [5]	11 [8]	9 [10]	5 [14]	hit

What to do with Strength and Dexterity adjustments for melee and ranged weapons, respectively? Subtract [add] the adjustment from the Armor Class hit.

For instance, an everyman with a 16 Strength (a 2-point improvement to-hit) would look like this:

d6 Roll and Lowest [Highest] Armor Class Hit
1	2	3	4	5	6
miss	12 [7]	9 [10]	7 [12]	3 [16]	hit

Magical enhancements to-hit function similarly.

While the simplicity of the AC Conversion II appeals to me, AC Conversion III is growing on me.

Saving Throw Conversion

Saving throws are preposterously easy to convert to d6 rolls.

In this chart, one can see the expecting saving throw values for 1st-level characters. To convert:

• A 16-17 saves on a 6+.

• A 13-15 saves on a 5+.

• A 10-12 saves on a 4+.

• A 7-9 saves on a 3+.

• A 1-6 saves on a 2+.

One can adjust these numbers downward by 1 increment (e.g., a 16-17 saves on a 5+) to increase the heroism of the characters involved (I would probably do so, but I am a soft touch). As noted before, a 6 automatically succeeds on a save and a 1 automatically fails.

Hit Dice, Weapon Damage, and Other Considerations

These are likewise trivial to convert, though the variance of such will be diminished.

• 1d4 converts to 1d6-1 (min. 1).

• 1d8 converts to 1d6+1.

• 1d10 converts to 1d6+2.

• 1d12 converts to 1d6+3.

• 2d6 remains as-is.

• 2d8 converts to 2d6+2.

And so forth.

Ability Score Adjustments to Other Areas

Several methods occur to me.

• Ability scores of 3 and 18 confer a -1 and +1 adjustment, respectively. This leads to maximally viable PCs, though there will be little differentiation between them in certain areas.

  
  

• Ability score of 6-8 and 13-15 convert a -1 and +1 adjustment, respectively, and 3/18 modify this to -2/+2. This leads to more variance in PCs, but they will be significantly stronger (and weaker!) than their standard counterparts.

  
  

• Ability scores of 3-5 or and 16-18 confer a +1 bonus. This strikes the middle ground between the two.

Now, I believe fully that the second method produces more interesting play, and it is quite possible that the more variable power level would be attenuated by the aforementioned 1/6 fail/success rule.