Quote:
Originally Posted by Ze'Manel Cunha
the statistical reality was often miles away from the perception the game writers had while writing the rules.
|
this.
d100 vs. TN may let OP avoid that, however. The worst statistical stupidities I've seen involve dice pools and sets of multiple dice. With a flat distribution, it may be easier to keep your probabilities feasible. Then again, d20 has a flat distribution and D&D 3rd edition breaks down whenever PCs are over or under 10th level.
This brings me to another important point to decide: what's the ratio of PC skill versus the die roll in your system, Jim? In d20, for example, the random element can be four times as significant as a talented beginner's skill. As the game progresses, the random element becomes superfluous (ie. you succeed as long as you don't roll a 1). In any case, you need to settle somewhere between these two extremes.
I have a homebrew system for a terrorist RPG I'm running where PC scores and dice roll results run from 1 to 10, rolled to hit a target number with an average difficulty of 10 (so average people can expect to succeed roughly half the time). Instead of a roll of plain 1d10, I use a median of 3d10 (roll three dice, discard the highest and lowest result) and a median of 5d10 (discard two highest and two lowest results). This keeps the spread similar (1 to 10) but lets me adjust the significance of the random element. It's simple and quick in practice because there's less addition and subtraction involved.