Bront
The man with the probe
The roll and add mods is nice, simple, and generaly fairly predictable without being too much so. Could be adjusted by using more than 1 die, or using precentile for more flexable modifiers, depending on the feal (Rolemaster, D20)
Roll all keep some can be ok, but the inherant flaw is that keeping is so much better than rolling an extra unkept die, it needs to be balanced accordingly (5 keep 2 is not as good as 4 keep 3 in many situations). It's not bad, but has some odd problems, mostly stemming from how the roll and keep is calculated. However, because you usually keep more than you roll, the probability curve is fairly good. (7th Seas, L5R)
I've seen a modified version of this system where you earn different dice and keep the best (Best die is a d12), so you could roll a d6 + 2d8 + d12, and you simply take the highest roll of all of the individual dice (Ironclaw/Jadeclaw)
Roll all keep all is better and worse. The level of diferentiation in skill/ability levels takes a sharp turn up as you increase dice here, and the level of randomness is irritating, though the curve leans towards the average. But the fact that on 5d6 you can range from a 5 to a 30 can have a huge effect on the game, and adding an extra d6 doesn't help as much as it could in smoothing this out. (Star Wars D6)
Roll under target method is ok, but the roll and add works a bit better, because you don't have the floor of a minimumn roll, and you can always raise your modifier. (Warhammer Fantasy, 2nd Ed Skills)
Roll many each one against a number isn't too bad, but can bog down if counting a success, and makes it hard to find the target if you have a floating target roll, or adjusting to successes on a flat target could be better as well. (Shadowrun, Trinity)
Diceless... well... let's not go there (Amber).
Roll all keep some can be ok, but the inherant flaw is that keeping is so much better than rolling an extra unkept die, it needs to be balanced accordingly (5 keep 2 is not as good as 4 keep 3 in many situations). It's not bad, but has some odd problems, mostly stemming from how the roll and keep is calculated. However, because you usually keep more than you roll, the probability curve is fairly good. (7th Seas, L5R)
I've seen a modified version of this system where you earn different dice and keep the best (Best die is a d12), so you could roll a d6 + 2d8 + d12, and you simply take the highest roll of all of the individual dice (Ironclaw/Jadeclaw)
Roll all keep all is better and worse. The level of diferentiation in skill/ability levels takes a sharp turn up as you increase dice here, and the level of randomness is irritating, though the curve leans towards the average. But the fact that on 5d6 you can range from a 5 to a 30 can have a huge effect on the game, and adding an extra d6 doesn't help as much as it could in smoothing this out. (Star Wars D6)
Roll under target method is ok, but the roll and add works a bit better, because you don't have the floor of a minimumn roll, and you can always raise your modifier. (Warhammer Fantasy, 2nd Ed Skills)
Roll many each one against a number isn't too bad, but can bog down if counting a success, and makes it hard to find the target if you have a floating target roll, or adjusting to successes on a flat target could be better as well. (Shadowrun, Trinity)
Diceless... well... let's not go there (Amber).
