I assess a task on three axes.
First, how important is a success for the gameplay:
- if it has to happen, the roll is usualy 3d6 + stuff vs stat, and simply gives degree of success.
- If it will ruin the campaign, don't allow it in the first place.
- If it is a grey area, in that it could happen, but won't kill the game if it fails, read on.
Second, consider if it is it likely to be relevant to the character's class or, if they've taken the time, their personal .
- If class appropriate, add their level to the stat.
- If just a backstory skill, add half.
Be generous -for example, bargaining for a good horse is probably a fighter related skill, although reading may not be. etc etc.
Third, do I have any idea how likely success is.
For the record, despote my deep and obsessive love of skill based games (Traveller Mon Amor) I think they can be a pain in class based games. Plus, I hate (hate, hate) having elaborate difficulty lookup charts -(besides wondering if "amazing" is really better or worse than "incredible" ).
In general, i use the well worn and reliable, roll under stat + stuff to succeed.
If I can take a guess at difficulty, I use an xd6 roll -
- For easy, I just let it happen.
- If it isn't much of a challenge and if I wouldn't mind it happening , 3d6 ;
- if it is hard, but okay 4d6.
- if it seems really hard, or I don't want it to happen, but can't justify it as a game killer, 5d6
If the player is coming up with an overly complicated plan, or missing the easy or obvious way, I use 4 d6.
If he has outsmarted me (bastard ! ), and/or will really complicate my life (like, solving the mystery in game one before the pizza arrives) , or is being astonishingly stupid, stubborn, obtuse or all three at once on purpose (or at least it seems to be) I punitively go with 5d6.*
I use xd6 becuase I like the bell distribution -not only is this a phoilosophical thing for me, it is a more reliable distribution - if you have the target a point or two above the average, you have a really good chance of success -in other words, the chance of failure diminishes faster once you get good at somthing.
If I don't have the slightest idea of weather its likely or not, use a D20, and I have autosuccess and auto failure. If I don't know (or care) about difficulty, I like a linear roll; since I don't know the liklihood of failure, it may as well be a constant factor.
If I'm remembering, I call a bell roll (xd6) a standard roll, and a d20 a wild (or whimsey) roll.
There we go. What else ? I generally assume that tools either allow it in the first place (lack of which makes it impossible), or make the task a gimmie. Lockpicks would be an example of the first, climbing gear, the second. I try not to add too much situational stuff -I assume that most of that is noise, and cancels out. Finally, I always try to reduce things to a single roll; permutations are a stone bitch, and eventually get you a player who invents a maxim gun in the stone age, or a master ranger who cuts his foot off setting a snare. And, while fun........well. Y'know. Table flips and all.
*Remember, I said this is how I work, and I'm a fossilized self aggrandized old grognard, so no irate responses about how my needs deprotagonize the colocated storytelling player's meta-experience, okay? I have the heavy lifting in the campaign, so I get a few selfish perks. Go empower the reprotagonized bloody albatross as much as your want on your own dime......
PPS. My job is keeping me very busy and interested. So, thus the lack of posts; fact is, I ain't cutting what little time I have for gaming so that I can write about gaming that I'm not doing; that way lies madness, and arguing on Traveller and D&D boards...;) .