Sunday, November 20, 2011

Part of an Idea for a d20 Mechanic

I really, really like Greg Stafford's core mechanic for King Arthur Pendragon.  It is both a universal d20 mechanic long before such became desirable and also a brilliant simplification of his own percentile system from Runequest.  As logically intuitive as percentile is, it's also way too fiddly for me: I can't believe that I need a game system to differentitate between a 52% chance to do something and a 53% chance.  KAP's system breaks it all down into 5% which is just right for me.

There is only one drawback to going from d100 to d20, which are the neat tricks you can play with criticals and fumbles in a percentile system.  Again, the native BRP idea isn't for me; it's intensely displeasing to my aesthetic to have a mechanic where you want to roll either high or low.  But there are other, niftier methods; my favourite is Unknown Armies' one of calling doubles that roll under as crits and doubles that roll over as fumbles (I have blogged about this before).

Obviously, you can't do this with a d20.  I think KAP's native system is that you crit when you roll your score exactly (my games are all still packed away from the move, so I can't check).  That is nice in one sense, since it maintains the blackjack mechanic of wanting to roll high, but, as a flat 5% loses any distinction between someone who is barely adequate scoring a crit and one who is highly-skilled.  I was thinking of how to change the odds a bit and, to do so, found myself putting all skill ratings into one of five broad bands:

  • Poor = 1-5
  • Not Good = 6-10
  • Good = 11-15
  • Great = 16-20
  • Super-human = 21+

The ways to get a critical success would be:

  • Poor = no crits (you stink; just try not to die).
  • Not Good = crit when you roll your rank.
  • Good = same
  • Great = crit when you roll your rank-1or your rank.
  • Super-human = same (but recall that you add the amount by which your rank exceeds 20 to all rolls).

Now the crit thing is nothing earth-shaking in of itself.  But coming up with the five types sort of seems clever to me as it addresses my number one problem with Runequest, KAP, and most other skill-based games: long stat blocks.  At the sacrifice of some detail, you could actually drop the base stats and just use the average for each type, plus the actual median, like so:

  • Poor = 3
  • Not Good = 8
  • Average = 10 (or 11; both are slightly wrong so just pick one)
  • Good = 13
  • Great = 18
  • Super-human = 23

Notice two things here. First, the averages map quite nicely onto the stats for BRP or, indeed, Ye Auld Game.  3 is piss-poor and 18 is the best you can hope for as a human.  Second, and less obviously, I think what I just did was translate the Moldvay/Cook unified modifier scheme from YAG to KAP.

When writing a stat-block, you could therefore assume that any ability is at Average and only mention the others.  Getting you something like this:
Ogre (STR 18, CON 13, INT 3, Claws 13)
That's still more involved that YAG (who's simple stat-blocks remain an almost Platonic ideal), but it's a helluva lot better than you usually see.


  1. I like this. I'm a fan of only things that matter being enumerated, like with your slimmed stat block.

  2. One of the advantages of a d% system is that you can use the same roll for multiple pieces of data. For example, I use a d% in combat, then break the 1s out as a damage roll and 10s as hit location. It's possible to also use the individual dice as randomizers for expanded description of scenes - if the last digit is even, give clue A, else B. Then there's the variable likelihood of critting (the UA thing)...

    I like what you're describing here. It's pretty light, though I could see it easily being even more granular. Set a standard TN (10 for a d20) and calculate a bonus/penalty for a stat (I'm thinking something like the 0e +3/-3 range). Your Ogre would become something like: Ogre: STR +3, INT -3 (animal-like), Attack +1 (claws), Defense +1 (bulky). If I wanted to add any further details about the this character, I could easily do that on the fly.

    Simple is good...

  3. You play a lot of ways without the more I think. Another option, for the FUDGE-inclined, is to just go the names:

    Ogre (Great STR, Good CON, Poor INT)