I posted some months ago in First Thoughts about a lightweight system I was building, called OneShot.  I received no replies, likely because I presented the system as a great wall of text with little analysis.  I've since playtested variants of the system in multiple campaign settings, and purified a few of the concepts; rearranged the stats, and a couple of other things.  For now, though, I'd like to share a bit of discussion my girlfriend and I had about the ramifications of the dice rolling system I incorporated, and why I consider it "non-negotiable" (at least short of introducing house rules).

I'll mention the relevant points of the system here.  Stats (Strength, Speed, Perception, Aptitude) are ranked from 1 to 8, where 4 is human average.  Player characters get 20 points to spend, so they'll have an average of 5.  Specializations are the fortes of a character -- not the exclusive limit of what they're capable of, but what they're especially good at.  Player characters get four of these.

The main design goal is similarly aligned to systems like Microlite; that is, distill standard task-resolution gaming to its core, strip out nonessentials, and keep the focus on the roleplaying if possible.  Ideally, GMs need to prepare the framework of a story, but the specific stats of NPCs and enemy characters can be generated on the fly.  "Hmm, I need a tough bruiser guy with a bat.  Okay, Strength 6, Speed 4, Perception 5, Aptitude 3, a baseball bat (power Strength +0), and specializations in heavy melee combat and toughness."

Dice are rolled as follows; the total set of numbers and modifiers are quite simple:

• Roll 1d10.  If it's a 10, stop; the action succeeds.  If it's a 1, stop; the action fails.
• Otherwise, add the most relevant of the four stats.
• If the character has a specialization in the field, add 1 if it's an attack or defense roll, and 2 otherwise.  Specializations don't stack.
• If the character has an advantage, add 2.  If the advantage is severe, add 4.
• If the character has a disadvantage, subtract 2.  If the advantage is severe, subtract 4.
• Advantages and disadvantages do stack.
• If the total meets or exceeds either the designated target number if a task in the environment, or the relevant stat roll of the target if it's an action taken directly against an opponent, the action succeeds.  Otherwise, it fails.

There's the health system, weapons and their damage, difficulty charts, blah blah blah.  That's material for another thread.  The main focus I want to attack here is:  why I chose +1/+2 for specializations, +2/+4 for advantages, and why I don't think any other numbers are necessary.

Specializations can be anything narrowly defined; that includes heavy melee attacks, pistol shooting, cover, cooking, short con, seduction, intuition, and so forth.  Originally, specializations provided advantage (+2) if you had a specialization that was relevant, with no other explanation. 

There were two problems with that:  first, specializations could be stacked.  If someone had a specialization in motorcycles, and full-auto shooting, should I give them +4 to attack rolls with a machine pistol while on a motorcycle?  I don't think so.  Second, it was likely that most players would take at least one specialization in both an attack and defense form, mainly because you'd be left behind if you didn't.  You were sitting on a huge difference (possibly as much as 15%) to your success rate on attacks or defenses, and of course, many more task rolls are made in combat than out of combat!  Because I wanted to make sure the system didn't devolve into nothing more than a combat simulator, I decided to disincentivize combat specializations by reducing their bonus to +1; still worthwhile, since combat rolls occur more frequently, but not a game-breaker.  I'll get back to this decision later.

Advantages and disadvantages are options built into the game for two main purposes:  first, to give the GM (whether that's a single person, or the collective decisionmaking process of the group) the ability to consider external circumstances, like attacking from high ground, or sneaking in a thunderstorm (thus muffling footstep noise), or similar things; second, to prevent the need for an exhaustive list of gear and explicit bonuses for each one.  For example, a scope attaches to a rifle, and does what a scope does.  I'd apply advantage (+2) to attack rolls at long range, where it can be brought to bear, and disadvantage (-2) to attack rolls at close range since the targets enter and leave the scope more quickly, and because instinctive, quick-kill shooting is much more difficult.  Nowhere is this, or even the concept of a telescopic scope, explicitly stated in the rules.

The idea is that all the GM has to do, for any circumstance like that in the game, is to determine is whether there exists a situation that is advantageous or disadvantageous, and if it's a serious benefit or complication.  Spend two seconds, plunk down a modifier, and make the roll.  I didn't leave room for any more granularity (+1 or +3) because I honestly think it doesn't make enough difference to matter; laser and reflex sights should ostensibly provide a +1 to firearm attacks at close range, but truthfully, not too many people use them (outside of the military, I suppose).  The benefit doesn't outweigh the costs in most cases.  However, reflex sights would essentially be "easy money" for most characters; they'd take them, almost without exception, because there's no downside; a flat +1 to all your close-range gunshots, essentially for free, is too good to pass up.

For my part, I consider the benefit to be so small as to be inconsequential.  That considered, there's probably a laundry list of situations that could, ostensibly, provide a +1 bonus or -1 penalty to a roll.  Sun in your eyes?  -1 to your Speed roll for driving that car.  Had a beer?  -1 to your Perception rolls to notice things, +1 to your Aptitude rolls to maintain a cool head.  Once you start considering a couple of these things, you kind of have to consider any situation that provides a small edge like this.  A point I firmly believe in with respect to RPG design and GM technique is consistency, consistency, consistency.  If a small edge gives you +1, you better get a +1 from all your small edges.  It's too much work, in my opinion, to have to improvise that.  So I ignore it.

When I bring these two points together, I arrive at a small amount of incongruity:  why allow a +1 bonus for combat specializations, but nothing else?  Do my thought processes and analyses hold water, or should everything be +/- 2, 4, or nothing?  Am I making mountains out of molehills?  I think my choices are sound, and failing any convincing evidence to the contrary, they're what'll go into the finished product, such as it is.  (It's not something I consider developed enough to publish for purchase, but I might slap together a rulebook as a free download and see what happens.)  That said, I'd love to hear what you all think on the decisions I made.

My question would be how much granularity do you think works best for the game. I imagine your GF disagreed with your static modifiers, and that you're now trying to defend them. You don't have to, the basic principal at work here seems to be abstraction v/complexity, and thats a choice thats split first by the goals of the design and second by personal taste. Some people won't like it, they can either modify the game or play a different game. If it works for you, in the type of game you want to play, I say go for it, and if you do lose anything in the way of realism or whatever, you gain a lot in flexibility. As for specifics, its not really an issue for you. Since you're not going to nail down what gets how much of an advantage (at least thats the impression I got from your post), these things are going to automatically tailor themselves to the opinions of the group, and while I'm sure many modern experts on things like laser scopes can give you statistics and theories of their effectiveness and applicability, I'm also sure there's room for debate, and with many RPG settings and players, genre conventions are going to outweigh such points (for example, I don't really care about how effective laser scopes are in or outside of the context of the game and I imagine as a GM I would just give everyone a plus 2 for any kind of scope, medieval weaponry on the other hand... I would be a lot more nitpicky, and thats just a question of how much I know, as a person, about each and how much I enjoy discussing/exploring their particulars). Go were the fun is for you, try to help others find it and as long as actual play does what "its suppose to" stick with what works.

Sounds logical enough. There are of course infinite ways to do this, each of them having their own merits and flaws. "GM chooses a difficulty number" has a few practical difficulties in play, since this decision can often feel arbitrary or difficult, but this problem can be solved with guidelines and examples.

I have a question about a more fundamental aspect of your resolution system.

Say may character wishes to climb into a window two stories up a brick building. He's trying to find a way into the building that bypasses the locked security door at the front. The GM chooses a difficulty number, I roll my dice, and succeed. Has my character made it in the window? What if the window is locked? What if it's nailed shut? What if there's a guard standing right on the inside of it? My character wants to get inside the building. It seems like there are a lot of deciding factors in that conflict that rely on the GM's decision. It seems like the ultimate arbiter of success is not the dice, but the GM.

Is that an accurate description?

Here's another example:

My guy wants to kill a dude. He's got his baseball bat out, and he waits behind a corner. When the dude walks past, he leaps out, bats the dude in the head, and runs away. The GM chooses a difficulty number, I roll my dice, and fail. What happens? Has my guy already run away, but his victim is still alive? Does his victim avoid the blow entirely? Does he catch my guy? Did he spot who did it? All this seems to be up to the GM.

What I'm saying is that systems that resolve tasks only leave a lot of decisions in the hands of the GM. They give the illusion of control to players, but actual responsibility for determining the outcome of events lies with the GM. You may be comfortable with this, but I suggest to you that there are simpler ways of achieving the same effect. I wrote a kind of parody of this style of game on my blog, here: http://simoncarryer.blogspot.com/2010/06/master-of-illusion.html

Nolan:  That's what I expected might happen.  My GF and I actually had a pretty healthy debate on this one, and she was coming from the standpoint of, "sure, that's how you and I would do it, but what about GMs who really want to provide all those bonuses?"  That granularity isn't something I consider a big enough deal to care about.  If someone else thinks it's relevant, house rule.  Seems reasonable enough.

Simon:  I have to admit, you raise a very good point.  While I'm not going to go and do anything crazy like switch to a conflict resolution system (honestly, that'd turn it into a completely different game!), it does suggest that I investigate things like that more closely.  Let me go over your examples and arbitrate each one, and maybe I can use that as a thought exercise.

Say may character wishes to climb into a window two stories up a brick building. He's trying to find a way into the building that bypasses the locked security door at the front. The GM chooses a difficulty number, I roll my dice, and succeed. Has my character made it in the window? What if the window is locked? What if it's nailed shut? What if there's a guard standing right on the inside of it? My character wants to get inside the building.

Let's start with this:  the chart for target numbers relies on the difficulty of the task at hand.

Trivial: 6
Easy: 8
Average: 10 (With an average stat value of 4, a character would succeed 50% of the time with no special training)
Challenging:  12
Very hard:  14
Impractical:  16

When I envision successive tasks like this, I view it to be several different steps in the same whole.  Resolve each individual concern one by one.  So:

Climbing the brick building:  tricky if there are no available handholds.  Let's say for the sake of sanity you've got a couple of railings to use.  I'd call that Strength vs. 13.  That gets you to the window.

Window's locked?  Well, if the lock is accessible on your side, that's Speed vs. say 10 (it's a pretty crappy lock).  Nailed shut?  Well, the boards are probably on the inside of the building, and you'd need a pretty significant blow to knock them in.  Strength vs. 14, and you're at a disadvantage (-2) due to precarious balance.

Guard there?  Well, he should be plainly visible from a few feet down; you'll know whether he's facing you or not.  (The GM could call for Perception vs. 6, but he's free to allow automatic successes on tasks that simple.)  The character is permitted to wait until the guard passes, or maybe open the window and pull him out, even!  (If the guard's facing away, opening the window initiates a combat round where the character could attempt a Strength-based maneuver to grapple and pull.  If he's facing toward the window, reaching to open it probably initiates combat right there.)

Thinking about this, I've come to an interesting realization about how I perceive the task resolution in this game:  anytime a character is faced with a new task or complication, a new die roll is needed.  I think it'd be a jerk move, as a GM, to say the character is automatically seen because he got to the top of the window, and "Oh!  There was a guard there!  Roll Speed for initiative!"

Jim, sure. Who decides if the window is locked? What's the basis for that decision? How do they decide if it's a simple lock, or some ultra-secure megalock?

Your comment about Conflict Resolution is revealing. Your game already has an implicit conflict resolution system. The system is "The GM decides". That's fine, but you may find it improves the experience of running and playing the game to provide guidelines to the GM for making those decisions.

Designers often tighten up their systems to make sure that increases or decreases in resources make differences in the way players interact with the system.

Example A: Robin Laws, Heroquest 1.0, Heroquest 2.0
- Bonuses to a roll on a d20 in 1.0 used the full range of intergers
- In 2.0 any lingering bonuses or penalties that result from conflict resolution, or from bringing in skills to supplement a main skill, or the giving of help between characters ALL work by jumps of 3.
- Laws experience and that of playtesters was that the constant haggling for little bonuses of +1 and +2 slowed down resolution.
- My experience was that too many little bonuses cluttered the shared imagined space too.
- Now in HQ 2.0 you can chose to give some significant help or go for synergy between abilities or measure your diminishing or increasing chances for success in easy-to-perceive steps. 
- A jump from +2 to +4 makes no appreciable change in the chance or frequency of sucess in Heroquest rolls, but a jump from +0 to +3, or +0 to -3, or +5 to +8 does.
- Players and GM now make a few significant choices rather than trying to tot up all sorts of insignificant ones in the hopes of making a perceptible change in chance or frequency of success.

Example B: Sandy Peterson at a game design seminar during Toronto's Gloranthacon a few years back
- He said why give your players just a bunch of measly +1% chances for success -- give'em something that produces really significant results.
- He drew our attention to Doom and why is inclusion of a few really special kinds of weapon make that game a blast.
- He didn't get the name Sandy "Quad Damage" Peterson for nothing!