So, I've got a game here. It's an Apocalypse World hack for just two players. It's an adventure game about children on the run, in a quest that's beyond them. You play a goblin child, maybe 6 years old, cast out from their tribe. You discover that the Demon King, the ruler of all goblins, has kidnapped a princess from another dimension and intends to sacrifice her, which will allow him to open the gates of hell onto the mortal world. You save her, and you're on the run, travelling the world over to fight mythical beasts that the Demon King hid his soul in, so that you can confront him directly, then open the doorway so the princess can return home.

It's about children being thrown out into a world that's much too big for them, and about growing up way too fast.

Here's the playtest file:

https://docs.google.com/open?id=0BwQFfBABRn2BeFJVZUtJbmhSc0NWY0RrbVVnQnhaQQ

And a sheet for it, since one isn't in the document yet:

https://docs.google.com/open?id=0BwQFfBABRn2BejJBdFpUR0dSN2FuRDJmenJ2a2JLdw

In playing it this past weekend, the player felt that the way the dice were assigned raised some issues.

Basically, you get a d4 for a skill you're bad at, a d6 for one you're okay in, and a d8 for one you're great in. You roll off against the GM's d6. You're looking to get higher than the GM. If you don't, the GM gets to make a move against you (for those of you not versed in Apocalypse World, that means the GM gets to make the circumstances in the fiction sticky, difficult, maybe dangerous, maybe really bad, depending). If you get the higher roll, it depends on what you've rolled. A 1-3 is an okay result, a 4-5 is a pretty good thing, and a 6+ is the best.

You'll always have two stats at a d6, one at a d4, one at a d8. Currency in the game allows you to temporarily boost certain stats, and when you Advance (roughly meaning, you level up), you get to change stats around, raising one as long as you lower another.

The player, for one, hated that last bit. Because it's not really advancement, right? You don't really improve, you're just shifting around where you're bad at. My most recent idea was to change the opening stat array to a d4, and 3 d6s, and whenever you Advance, you raise one of them. You Advance 12 times in the game before it's over, so that means at five Advancements, you're as high as you can go. I don't know, I don't feel like that's a valid avenue either. I'm really at a loss here.

The other thing that came up was the player felt like her d4 trait was ultimately useless. It couldn't get the best result, it was literally impossible to do so. It was kind of disheartening that way.

Here's a list of the dice rollery, for reference:

https://docs.google.com/open?id=0BwQFfBABRn2BRm9DamM3SmZUWENISnBuQ0I2dEZuUQ

I've been spinning my brain trying to figure this out! I really like the simplicity of just changing die sizes, it's really important to me. But I can't figure out how to even out some of lumps.

Maybe turn the dice upside down so 1 is the best result and larger numbers are worse. 

1-2 : great result
2-4 : result with complication
5+  : poor result

I really do like that. I think it solves one problem, but raises another - the d4 goes from kind of weak to super strong. Maybe some tinkering with the numbers. But I dig it!

Whats' wrong with having a weakness?

I'm not sure what the skills cover, but the d4 seems perfectly adequate for something you're not good at.
How are"untrained" skills handled in the game? Are the 4 skills the players receive broad enough to cover most interactions?

Moreover, as a game with 2 player characters, I think it's important for the characters to be able to cover for each other's weaknesses.

I know if I were going to break into a government weapons cache, and the guards could only be defeated with Calculus... I'd bring a friend along.

Oh no, it's not for two player characters - it's for two players total. One GM, one player.

There's four stats which are plugged into a variety of moves. "When you want to __________, then roll ___________", for example. The moves are very broad:

- Going into danger
- Trade a few blows in combat
- Learn something about a person, place, or thing
- Cast a spell
- Persuade someone of something
- Sneak around or go unnoticed.

Whenever the player wants to do one of those things, they find the stat listed for that move (for example, the move for going into danger is called "Face certain doom," and when you do that, you roll Courage).

So when the player says they want to do something, they say something specific: "I want to slip around the corner without the guard noticing," and the GM then interprets that broadly, as it applies to the moves above. Go unnoticed, obviously, but it could also be going into danger, or even a combination of the two. It the player does something that falls under the purview of one of the moves, they roll for it. If they do something not related to any of the moves, then they don't roll for it. Usually it just happens.

"I move the boulder" - Okay, great, it's moved. But compare that to...

"The giant scorpion is bearing down on me? I try to move the boulder to get out of the way" - This is going into danger, because it's dangerous to try and spend time moving this boulder with a big scorpion about to getcha.

The moves are meant to broadly cover the important and dramatic things the player wants to do, with wiggle room to adapt. If the GM decides it's necessary, she can make a custom move, which is a move invented on the spot just for a specific circumstance.

The problem that came along with the d4 as a weakness was when you take, let's say, Courage as your d4 stat, you're making every move that involves Courage suddenly become a problem. It's not that there's a weakness - it's how that weakness influences several different aspects of the character that was the problem.

Here's the going errata:

Instead of white 1-3, each move works on a white 1-2.
Instead of a white 4-5, each move works on a white 3-4.
Instead of a white 6+, each move works on a white 5+.

This lets the d4 remain weak, but still allow you to make powerful hits. Someone over the AW forums mentioned a statistical issue with the "lower die gets a +1" problem, but I like what it does here. So, for now, I'm still working with it.

Could you consider increasing the size of all the dice to d6/d8/d10 vs d8 or d8/d10/d12 vs d10?

This would achieve two things:

1) The result range of the dice overlap more, so there might be some room for giving the lowest die access to the best result.

2) The "bad" die has a greater chance of success.

To expand on the last point, if the player has to roll higher than the GM, a d4 vs d6 will only succeed 6 times out of 24, or 25% of the time, while a d6 vs d8 will succeed 15 out of 48 (31%) and a d8 vs d10 28 out of 80 (35%). 1 out of 3 is still risky, but a lot better than 1 out of 4.

Sure, I don't see why not.

Really all I want to do is play with the dynamic of "player rolls a white die vs. a black die, highest one wins, and if the player wins, what they rolled means something". I don't mind it being risky, but I want people to be comfortable with it. Odds-wise, I want it to be challenging but not bleak.

If instead of varying die sizes, I used multiple dice, would that alleviate the problem? Maybe like, you get two dice to roll versus the bad die on an average roll, but only one die to roll versus the bad die on a weak roll? The 6+ result in the moves, in play, has shown up largely unnecessary, so I wouldn't mind doctoring it.

for advancement simply have the ability to add 1 to that attribute, with it costing more advancement each level.
for instance,

d4
d6
d6
d8

on level up you spend 1 'rank' to gain

d4
d6
d6+1
d8

then if you wanted to raise it to d6+2, you spend two 'ranks', meaning you have to save up. You wouldn't rearrange your skills but could improve them slowly, balancing as you saw fit. do you raise the d4 to offset penalty or improve the d8 for great ability?

simple but it should work ok for what you are talking about. I like the premise well enough, very Dark Crystal.