So, I've been toying with just using 1d6 and manipulating the TN. That is, if you're "trained," the odds of success are 4/6 (66%), if you're "untrained," the odds are 2/6 (33%). You can move in intervals of 16%, and can generally give up to +/- 1 or 2 in modifiers.
How do the odds work out if you just start with a fixed TN of 5 and then add dice to a pool, taking just the best result? A "success" requires at least one die showing a 5 or higher.
This does not let you modify the odds of success down much below 33% -- although you could make a rule that says that if you're down to 1 die in your pool, additional penalties jack the TN up to 6 and then its impossible.
The first +1 modifier (adding a die to the pool) gives a +22% boost. More than a +1 on the linear D6 roll (+16%), but not by much. The second +1 modifier (getting to the third die) adds a +15% boost. Again, close to our other odds. You then start to get diminishing returns. This is nice because it allows you to add many more positive modifiers. You can give more than +2 out as a bonus without having to worry about auto-success.
Another advantage is that you can use this for degree of success pretty easily. For example, a character using the Athletics skill can leap a number of squares equal to their dice pool roll. Or, you can count successes to see if they can run a race faster than someone else.
Dice Pool Odds stolen from here: http://www.darkshire.net/~jhkim/rpg/systemdesign/dice-silhouette.html
OD&D Experience Levels
8 hours ago