Monitors: you win, Dan, you win

So after not inconsiderable thought (as John Major might have said, if he designed RPGs) I have decided to move Monitors over to a dicepool system.

Initial thoughts for the system are as follows:

• D6-based.
• Each point of Attribute lets you roll one die.
• Traits may grant additional dice.
• Lingering effects may grant or remove dice.
• Penalty dice remain in use. They are rolled at the beginning of the turn as previously. If the die hits or exceeds the target number, it negates one success on all relevant tasks this turn; on a roll of 1 the effect dissipates and the penalty is removed. Die size varies, so a 1d4 penalty is mild, but a 1d12 penalty is likely to hit hard and often.
• If you have an appropriate Skill, you gain one automatic success.
• More difficult tasks require more successes.

Because Skill grants an automatic success, a skilled character will automatically succeed at simple tasks. They will also be much more likely to succeed at more challenging tasks, since they need to roll fewer successes. This system seems preferable to allowing rerolls or changing target numbers, either of which would increase the chance of success but still allows a chance of abject failure - with this system there's only so badly an expert can do. It's also simple, which is a big factor.

Opposed rolls work in the same way - roll Attribute plus relevant traits, highest score wins. On a tie, a skilled character wins. If there's no skill discrepancy, break the tie however you like unless a tie seems appropriate.

The question is, what target number do we use, and how big an Attribute pool? I did some calculations.

I get the feeling that the sweet spot for this system might be target 3+, 3 dice and 2 successes as the default "moderate skill, average difficulty" challenge. This would give a 75% chance of success - not what I'd pick for a crunchy game, but it seems appropriate in a cheery adventure game of highly competent professionals. A better attribute or having training offers a very high chance of overall success, while even someone with a low attribute has a 45% chance of success on normal challenges. Meanwhile, simple challenges are only an issue for the least competent (67% with a single Attribute and no training). The low target number means difficult scales fairly sharply (which multiple successes always does) but maintains at least a bit of a curve; compared to even the 4+ probabilities there's far less of a gulf.

In this case, we could reasonably assume that Attributes go from 1-5, but Monitors (due to exacting recruitment rules) can't have any lower than 2. This immediately offers the possibility of NPCs just being bad at things and getting shown up. There might be the odd Monitor PC with an Attribute of 2, but in most cases they'll be at least a little higher. On the other hand, training continues to be relevant whatever your Attributes are; on an average task an Attribute 5 PC will fail one time in twenty, while a trained Attribute 5 PC will fail less than one time in two hundred. I would expect dicepools to generally cap out at 8 (Attribute 5 + 2 traits + assistance).

Difficulties will go up several steps but down only one. Most die rolls should require 2 successes or more, but the Easy 1 success roll is there for when failure might be interesting and/or diepools are very small. Anything particularly easy just works. Harder tasks are much more interesting, because they let you highlight the competence and heroism of your arcane cyber-lizard by doing things that are really freakin' hard.