Assumptions:
A skilled craftsman has a skill of 6. A greenhorn has a skill of 1 (I assume Dex+3 for both). This means that the chance for the greenhorn to succeed at a task of difficulty 12 is about 20%, and 70% for a professional.
Frankly, luck seems to play too important a role in this.
If we took d6+2 instead of d10, we'd have about the same average rolls. Of course this opens up a whole new filed of dealing with botches and exceptionally good rolls.
Alternative:
Import the "take 10" and "take 20" rules from D&D3.5. That game is played with a d20, so it would be "take 6" or "take 10".
I have been toying with this for the past week or so (since it appeared in the forum as an idea). I would make the roll for botch a 1 followed by another 1, and use added sixes for success (6+6+6+...) instead of explosive dice. It reduces the number of botch rolls (1 in 36 times you check for a botch, instead of 1 in 10 times) and decrease the explosive value of the dice. Makes skill level much more important.
Why stop at just d06.
Ability Score 01 uses a d10 , A.S. 02 to 05: d08 , A.S. 06 to 08: d06 , A.S. 09 & 10: d04.
Add a fixed numeric value to the base die as determined by xp spent on the Ability.
Using d4, d8 and d10 sort of scares away the newbies.
d6 is for everyone. the Volks-dice.
d10 is easy for doing statistics. The dumb-dice.
The may be other ways of dealing with this problem, but in a near perfect game, it is a problem.
From D&D (modified for d10): Take a number instead of rolling one.
"Take 5": Player assumes he rolls a five and doesn't roll at all (non-stressful situations only). Less roll-playing. Less chance.
"Take 10": Player assumes he tries until he does his best. Consumes 10 times the normal time. Typical application: lock-picking.
Using a d10 works well for Arts, though. It is just with Abilities that it introduces too much variability.
Maybe a simple fix would be to divide the die roll by 2 (rounded up) when applied to an ability? Keep the same rules for botches (0) and extraordinary success (1), simply divide the result of the die roll.
If you want to reduce variability but dont want to radically change the skill numbers, try using 2D6-1 instead.
Double 1>botch...
D10 average, 5.5, 2D6-1 average 6 but much more average or close to average rolls.
This doesnt really reduce the chance to get results in the outer zones of variability much. You still have equal chance to get each result even if you move it to 3-8.
Unless doing something special, i tend not to bother with a dieroll at all for things like crafts...
And it would be "take 5" or "take 10" not 6 unless you wanted to change its level.
Anyway, the "take x" alternative is actually quite good and very useful. I think it would be a good addition in AM.
Yeah that probably works somewhat as well, even though it also reduces the max normal result that can be achieved which means difficulty may need to be changed for rolls, which is a bit too big of a change for me at least.
It's a little clumsy but it does express the fact that target numbers for (Art+Art+Attribute) effects are on a different scale from (Ability+Attribute) effects. Going below d10 on a casting total makes the roll almost pointless, while on most skill rolls the die imposes far too much randomness for my taste. I think you're going in the right direction.
Instead of halving the die, you could just double the Ability and Characteristic. Juggling with the relative multipliers of Abilities and Characteristics (e.g. multiply only the Ability and not the Characteristic in situations where knowledge and training should matter more than raw talent) was, incidentally, suggested in the Corebook of Ars Magica 3rd edition.
A similar possibility, is to just roll 2d6, but, if you roll 12, to add 1d6 with exploding dices mechanics. Slightly easier to calculate, slightly higher average results, less frequent and extreme high results.
besides, using a single plain old D6 is easier if you want to reduce variablility.
