Here is one way to use it: roll 2d6 and add attribute modifiers, etc. Check the result:
|3-5||Failure. Something bad happens.|
|6-8||Nothing happens, partial success, as expected, uncertain results, success with a cost, etc.|
Let us see how a "modern" RPG such as Apocalypse World handles conflict resolution (the quotation marks is both because AW uses the classic 2d6 method of resolution but also because there is some old school philosophy in the game, IMO; but this is the matter for a different post... or posts) .
Here is how it goes: you roll 2d6. Add modifiers (that are just slightly higher that you would expect on Moldvay). An then:
"On a 10+, the best happens. On a 7-9, it's good but complicated. On a miss, it's never nothing, it's always something worse."
The idea of "good but complicated", or "success with a cost" is quite popular in modern games; many old games used a pass/fail system instead.
I never played AW, but I like Dungeon World, which uses the same base.
DW has no "critical failures" rule that I can remember right now (although there are soft moves and hard moves that can happen on a failure), but some classes have "critical success" on a 12+. The resulting matrix would look like this:
|2-6||Failure. Something bad happens.|
|7-9||Good but complicated.|
|10+||Success! / "the best"|
|(12+)||Critical success (sometimes)|
Anyway, the thing I want to to discuss is this idea that when you "fail" with the dice, there must be consequences. "If there is no consequence, why are you rolling?", or so the zeitgeist says.
The thing is, coming up with three possible outcomes whenever you roll the dice, with "nothing happens" out of the table, takes some effort, when compared to "if you succeed you get what you want, if you fail you don't".
Now, I have to say that, in my opinion, "critical success" and "critical failure" are less complicated than, well "good but complicated" (if you pardon the pun). They add less complexity than some kind of "complicated" middle ground, because the extremes are just "more success" and "more failure", most of the times.
What I am trying to say is that "nothing happens" is often a good solution, at least in two situations.
When time management is important, like it happens to be in OD&D, "nothing happens" really means you wasted time, which is enough consequence by itself. The same idea applies when you're wasting other resources.
The other situation is combat. Coming up with clever complications and consequences for every to-hit roll seems like a waste of time when you're actually inside a dangerous situations where not getting out of it is often bad enough.
It is worth mentioning that Dungeon World has you covered, most of the time; it already says what to expect from a 7-9 roll, for example. You still have to come up with good ideas for failures, but they are somewhat implied by 7-9 results or the situation at hand.
But these emphasis on success at a cost cause some odd situations in combat. Very often, you find yourself choosing, after you roll, if you want to put yourself in danger to hit the enemy or run out of ammo, etc. This not only destroys my immersion in the game (yes, I like avoiding making decisions my character couldn't make - your taste might be different), but also makes combat a bit too cumbersome for my tastes - one must choose what to do, then roll the dice, and then choose what to do with the result of the dice.
"Say what you want to do, roll the dice to see you succeed" is often good enough for me.
What's the point?
Yeah, I'm rambling again.
Anyway, I like old school games, and I like modern games, but some ideas are just not to my tastes. One single resolution method for skills and combat are a bad idea, in my opinion: a d20 makes for insanely uneven skill rolls, while things like a bell curve and "success with a cost" seem to slow down combat. I played a lot of GURPS back in the day, and 3d6 made skills reliable, but combat got too predictable after a while.
All in all, I think old school and new school methods are tools, and nothing else; there is nothing inherently better about "unified mechanics" or "success at a cost". Just use what you want, when you want, and see what suits you better.