I must create a system, or be enslaved by another man's. I will not reason and compare: my business is to create.

- William Blake

Thursday, April 11, 2024

Where AD&D is better than 3d6 in order

Being a fan of basic D&D, I always thought that rolling 3d6 in order was the cleanest, fastest way to generate PCs. Also the most fair, so PCs start, on average, as a normal person.

In addition, the -3/+3 modifier spread is beautiful and intuitive, while in AD&D you need to consult lots of tables to fill your character sheet.

In comparison, the AD&D methods were not only more complex - involving more dice rolling, sometimes to ridiculous lengths - but also made starting PCs stronger than average humans (but they could still be knocked out and maybe dying from a 10-foot fall).

In addition, this added complexity, redundantly, often got you to the same modifiers you'd get in B/X. 

For example, in B/X you get abilities of 10.5 on average, while you needed 13 Strength to get a +1 bonus to damage. 

In AD&D the average is 12.24... But you still get +0 to damage even with 13! 

So what is the point?

Well, the point is that despite these things, AD&D got a few things better than B/X here.

- The default modifier is +0, like in B/X, despite higher abilities.
- However, the bell curve in 4d6 is "higher" - averages (12, 13) are more likely and negative results are a lot rarer, which makes the game a bit faster since subtraction is uncommon (addition is quicker).
- Maybe PCs should be a bit stronger than the average human? And, specially, avoid PCs that are extremely weak in any area (e.g., PCs that can barely speak).

In short, this 4d6 methods was adopted in subsequent editions for good reasons - and there are plenty of B/X players using it too.

However, in B/X this makes PCs a bit too strong for my taste.

There are several methods to combine the advantages of both systems. In B/X, these are some of my favorites:

A) Roll 2d6+4 to each ability score. This avoids extreme results and gives you an average of 11.
B) 3d6 in order, but replace one result of your choosing by 15.
C) Some standard array to make things even faster (for 3d6, it could be 14, 12, 11, 10, 9, 7 or 13, 12, 11, 10, 9, 8; add 1 point to each to get an average of 11.5). 

I'd let the players swap two attributes to play the PCs they want (or assign to taste in "C").

[One thing to note is that I use feats and ability scores improvement, so that eventually you can raise your ability scores. If I didn't, I'd probably use "B" - leaving the possibility of 3s and 18s and also allowing you to play basically any class you want].

Notice that this is quite close to the averages in the D&D B-series pregens - who have average ability scores around 11-12 or a bit higher, as explained here.

This range looks satisfying to me.

Of course, it is ultimately a matter of taste - do you want PCs to be ordinary, a bit better, or heroic form the start? 

But for my B/X(ish) games, this average of 11.5 - just a bit above a normal human - is what I like.


  1. I agree that preferred method is absolutely a matter of personal taste, though system details influence this. If you only give modifiers for stats of say >15 (+1) or <5 (-1) and don't do many roll under ability score checks, the game will be quite forgiving of any method including 3d6 down the line. Conversely if you give up to 18 (+4) like D&D 5E and/or do many roll under ability checks (particularly 3d6 rather than d20 roll under) then you have much lower tolerance for variance I think, and either weighting scores higher (like 4d6 drop lowest) or otherwise ensuring more balance seems best.

    I happened to recently post about another method of ability score generation that might be of interesting, particularly if you like roll under ability checks: https://illusorysensorium.com/12d6-dtl/

    1. Yes, this makes sense.

      Read your post, this is a cool way of generating abilities I hadn't read before!