Celebrim
Legend
Back in the day I used 4D6 take the best three in order as my means of producing ability scores. This was never wholly satisfying. By 1990 I had moved to 4d6 take the best three in order, but you could choose 5d6 take the best three for one ability score of your choosing at any point when rolling. This worked a bit better. Then I started playing heavily with a different group that used method V from the Unearthed Arcana. I disliked Method V because it just rewarded you for picking the classes with the highest required ability scores, resulted in pretty cookie cutter characters. It did allow you to play what you wanted rather than what the dice gave you, but the main thing I liked about random ability score generation was that it forced you to build around the array and play something you might not have thought about playing. But then again, I rarely got an opportunity (before joining this new group) to be a player for more than two or three sessions, so it wasn't like I was getting tired of being forced into something I didn't want to play.
Even with the Method V group, there was a considerable amount of cheating. Some players were pretty obviously showing up with characters that hadn't really rolled. Other players - like myself - were making 15 characters for fun (it legitimately was) under whatever system, and then picking the one that they wanted to play the most. If this happened to correspond to the character with the best stats, that was just a coincidence right? Interesting, the night I had joined the group that I found out had used method V, I was using method I and had made like 15 characters, ending up with an elven thief/M-U (which I'd never thought to play before). I had thought my ability scores, which included a 17 and a 16 would seem suspiciously good and was a little worried, but they complained that they were too low. So I rolled up Method V in front of them and ended up with a lucky 18 and a 17 - but all suspiciously still the lowest stats in the group (I think I was the only one who had an 11 in any ability score).
Over the next few years I grew increasingly frustrated (as a DM, not necessarily as a player) with 1e AD&D because of its limitations. By the mid-90s when I finally abandoned the game, I had decided that if I did run the game again I would use Method III for ability score generation based of some experiments with each of the four DMG methods. I liked that it very rarely generated an unplayable character and that it had a low standard deviation, with not too many 18's but almost no low scores. You'd generally have the two 16's that I considered necessary to build a viable long term character that was more than just the thespian skill you brought to the role.
When I went back to D&D in 3e I used a point buy variant I had picked up somewhere that I was really happy with. My evolving homebrew system had no dump stats, and just taking a lot of 14's was very viable, while you could buy up an 18 to shine in one area, you'd suffer in others. It made interesting characters and ensured even spotlight and that everyone got someone that they wanted to play. Early on I had allowed my old 1e random generation method, but I dropped it within a couple of years because 90% of the time if they rolled in front of me they ended up disappointed. Hoping for more points than what you were allocated was usually a losing proposition. It could happen, but it was like 1 in 6 times and even when it did happen, it really meant everyone who hadn't gambled was losing. Random generation was over for me.
But 1e is a different proposition, and giving up random generation would be giving up on an essential feel. So, here are the methods I would endorse if I was playing again.
a) Method III: 3d6 six times, take the best one, done six times in order
b) 5d6 take the best 3, six times, arrange to taste.
c) 6d6 take the best 3, six times, in order
d) Take the array 18, 17, 14, 13, 12, 10; arrange to taste.
Players could choose the method they prefer, but they have to roll in front of the group and play what they get. If you seem to be trying to evade the stipulation of playing what you get (by for example committing suicide), you are forced for future characters to take Method IX, which makes a perfectly good fighter or M-U; nothing fancy, but a solid character. Overall ability averages are quite high, and while I concede that each method has its drawbacks, they are all generous enough that we shouldn't have too much whining or cheating.
Which method would you prefer?
And, how to you handle ability score generation in 1e/2e AD&D?
Even with the Method V group, there was a considerable amount of cheating. Some players were pretty obviously showing up with characters that hadn't really rolled. Other players - like myself - were making 15 characters for fun (it legitimately was) under whatever system, and then picking the one that they wanted to play the most. If this happened to correspond to the character with the best stats, that was just a coincidence right? Interesting, the night I had joined the group that I found out had used method V, I was using method I and had made like 15 characters, ending up with an elven thief/M-U (which I'd never thought to play before). I had thought my ability scores, which included a 17 and a 16 would seem suspiciously good and was a little worried, but they complained that they were too low. So I rolled up Method V in front of them and ended up with a lucky 18 and a 17 - but all suspiciously still the lowest stats in the group (I think I was the only one who had an 11 in any ability score).
Over the next few years I grew increasingly frustrated (as a DM, not necessarily as a player) with 1e AD&D because of its limitations. By the mid-90s when I finally abandoned the game, I had decided that if I did run the game again I would use Method III for ability score generation based of some experiments with each of the four DMG methods. I liked that it very rarely generated an unplayable character and that it had a low standard deviation, with not too many 18's but almost no low scores. You'd generally have the two 16's that I considered necessary to build a viable long term character that was more than just the thespian skill you brought to the role.
When I went back to D&D in 3e I used a point buy variant I had picked up somewhere that I was really happy with. My evolving homebrew system had no dump stats, and just taking a lot of 14's was very viable, while you could buy up an 18 to shine in one area, you'd suffer in others. It made interesting characters and ensured even spotlight and that everyone got someone that they wanted to play. Early on I had allowed my old 1e random generation method, but I dropped it within a couple of years because 90% of the time if they rolled in front of me they ended up disappointed. Hoping for more points than what you were allocated was usually a losing proposition. It could happen, but it was like 1 in 6 times and even when it did happen, it really meant everyone who hadn't gambled was losing. Random generation was over for me.
But 1e is a different proposition, and giving up random generation would be giving up on an essential feel. So, here are the methods I would endorse if I was playing again.
a) Method III: 3d6 six times, take the best one, done six times in order
b) 5d6 take the best 3, six times, arrange to taste.
c) 6d6 take the best 3, six times, in order
d) Take the array 18, 17, 14, 13, 12, 10; arrange to taste.
Players could choose the method they prefer, but they have to roll in front of the group and play what they get. If you seem to be trying to evade the stipulation of playing what you get (by for example committing suicide), you are forced for future characters to take Method IX, which makes a perfectly good fighter or M-U; nothing fancy, but a solid character. Overall ability averages are quite high, and while I concede that each method has its drawbacks, they are all generous enough that we shouldn't have too much whining or cheating.
Which method would you prefer?
And, how to you handle ability score generation in 1e/2e AD&D?
Last edited:
