Balancing Rolled Attributes [Archive] - Wizards Community

Balancing Rolled Attributes [Archive] - Wizards Community

Post/Author/DateTime Post
OmiKapSi
02-22-06, 11:22 PM In my endless quest to create a system for a balanced but random attribute generation system, I've come up with the following iteration, which I decided to post here for feedback.
Warning, basic math ahead. :)

This is based around using the '4d6 drop the lowest' option for stat rolling, but in theory can be used with any system that generates random numbers between 3 and 18.
The target is characters with a range of attributes who's total modifier is +4 to +6. (If you add up the modifiers of each characters attributes, your result will be +4, +5, or +6.)
With this in mind, I decided that if all the characters attributes added up to 73, they were gauranteed to fall in the target range, depending on how many attributes were odd.

So, the system. A player generates 5 attribute values using whatever system is preferred and records them in the order that they were generated. Like I mentioned, 4d6 works best for this part, but it's quite variable.
Once the attributes have been generated, follow the steps. After each step is complete, use the modified attributes rather than the original rolls for following steps:
Step A; Add the first and second attributes, if the total is less than 9, increase the second attribute by the difference between the total and 9 (making their new total 9).
Step B; Add the first three attributes. If their total is is less than 25, increase the third attribute by the difference between the total and 25 (like in step A).
Step C; Add the first four attributes. If their total is less than 41, increase the fourth attribute like in previous steps. If their total is greater than 63, then reduce the fourth attribute in the same manner.
Step D; Add the five attributes. Adjust the fifth attribute as previous steps to make sure that the total is 57-68.
Step E; Subtract the total after Step D from 73. This is the sixth attribute.

e.g. Barak is rolling up a character. His rolls for the first 5 stats are 3, 4, 17, 15, 13.
Step A: Bilbo's total for his first two attributes is only 7 (3+4). He increases his second roll, the 4, to a 6.
Step B: Barak's total is 26 (3+6+17), so no adjustments are made.
Step C: Barak's total is 41 (3+6+17+15). No changes.
Step D: Barak's total is 54 (3+6+17+15+13), which is 3 lower than 57. He adjusts the 13 to a 16.
Step E: Barak subtracts 57 from 73 to get 16 for his final attribute. His attributes are therefore: 3, 6, 17, 15, 16, 16.

Silk then rolls up a character. His rolls for the first 5 stats are 14, 13, 16, 16, 17.
Step A: No change.
Step B: No change.
Step C: No change.
Step D: The total of the five attributes is 76, which is too high, the 17 gets bumped down by 8, making it a 9.
Step E: 73-68=5. Silk's attributes are 14, 13, 16, 16, 9, 5.

I think that this system works best for when attributes are arranged as desired after generation, however it can work just fine with them not being rearranged.
So, what would be your response if your DM told you that the new campaign would be using this system for attribute generation?
Seeker95
02-23-06, 07:35 AM Just an observation: What you are doing is a dice-driven point-buy. Not that this is a bad thing, but I would prefer point buy if the risk and benefit of rolling was gone (the real randomness of it all).
Thelandrach
02-23-06, 10:37 AM Yikes almighty. Why not just use a 24 point buy? You can get 6 12's, which is +6, or two 14's, two 12's, and two 10's, which is also +6. Was it your intent to have the numbers evaluated in order, lowest to highest, so the lowest stats stay (some poor schmuck gets stuck with a 3 in your first example), and the highest ones get clobbered (if somebody legitimately rolled an 18, his chances of ending up with that stat are pretty much nil). If I introduced this to my players, they would (a) not understand it at all, (b) kill me, or (c) both.

also: You have a player named Barak???
Grand Theft Otto
02-23-06, 11:32 AM Let go of the security blanket. What you want is a balanced method for creating character stats. A covoluted random rolling isnt the answer.

If point buy isnt your thing, I can suggest another option. Pick a number. Everyone's stat mods must add up to this number. Up to 3 of your ability scores can be odd.

Some groups may desire to set low ranges for stats (ie, nothing below 6 or whatever).

We used +7, and its worked out fine.
OmiKapSi
02-23-06, 11:57 AM Seeker: So you'd rather take the risk of rolling a character who would be weaker than the group average for the potential benefit of rolling a character who could be stronger than the group average? I don't consider either result desirable.
If there was a great need to have a more variable result (characters ranging from +2 to +8, for example), it would be a very simple matter to use 66+2d6, for example, instead of the base of 73, and modify all other numbers in the system appropriately.

My reasons for not using point buy (or other choice based systems) are numerous. In this system, I'm not just seeking to have all attributes fall into a middle range. I actually kind of like having a few attributes at extremes, it makes for interesting characters, who have significant strengths and weaknesses. My goal is to have random attributes, but without the risk of having players dissatisfied with their rolls, or having players who's rolls are so high that they end up dominating the game.

And the numbers are not evaluated in order from lowest to highest, they are evaluated in the order that they are rolled. If someone rolls an 18 in their first two rolls, the system guarantees that it won't be touched. It's simply a coincidence that the numbers in the second example are 'in order', the numbers in the first example are not.
sherretz
02-23-06, 02:48 PM Your system has a few flaws also and I think the dice and/or point buy systems are better.

First, if the player doesn't roll an 18 in the first two stats, they cannot save their 18 roll. It will take a while for the players to be wise on the system before they realize they can rearrange their rolls so that their higher rolls are preserved. Also, with a bit of work, an intelligent player can use the system to min/max his character. Since your Step C has a total range, they can max it to 63 points total and get higher stats through.

Plus, above about 5th level, feats and skill points take over for ability scores with regards to the aptitude of the characters. So really, you're only guaranteeing a "balanced" character for each player until level 5 to 7. Thus in the end your system doesn't really make much of an impact on twinked characters.

Really though, you should just talk to the players about min/maxing and twinking. If you understand why they are apt to do it, then you can give them reasons why it's not that great. Just because you like characters with single-digit ability scores, doesn't mean every player does. This appears to players like you are imposing your will on their PCs. Also keep in mind that this system will discourage players from playing characters that require a large number of average to above-average stats, such as Paladins and Monks.
Arrowhen
02-23-06, 03:22 PM One thing I've been thinking about lately is rolling ability scores and then balancing them with action points: the lower your stats are, the more action points you get. This way, you'd still get the wider variety of ability scores afforded by rolling, but players who rolled poorly would have a larger pool of luck to fall back upon when it really mattered.

I haven't played around with the numbers yet, because we've got a couple of months yet to go in my current campaign and the next on I run probably won't be D&D, but it's something I've been kicking around in the back of my mind.

Any of you math-types want to take a stab at figuring out how much ability score points and action points are worth in relation to one another?
OmiKapSi
02-23-06, 03:58 PM Since people have twice missed the fact that attributes are evaluated in the order that they are rolled, I've bolded it, hopefully this will prevent further confusion.
Seeker95
02-23-06, 10:18 PM Seeker: So you'd rather take the risk of rolling a character who would be weaker than the group average for the potential benefit of rolling a character who could be stronger than the group average? I don't consider either result desirable.I prefer point buy.

But *if* we're going to roll, then yes... I'd rather take the risk of a poor / low stat character to gain the possibility of a high stat character.

Your method just makes the outcome similar to point buy, but complicated.
If you want something other than point buy without the randomness of rolling, just hand the players 72 points (or whatever amount) and tell 'em to distribute as they wish.
sherretz
02-24-06, 09:01 AM Since people have twice missed the fact that attributes are evaluated in the order that they are rolled, I've bolded it, hopefully this will prevent further confusion.

Ok, I misread your post above mine. In that case, it makes the system even worse because you are penalizing players who roll an 18 after their first two rolls. I seriously think you're handicapping the players and PCs by doing this.
danzig138
02-24-06, 04:22 PM I prefer point buy.

But *if* we're going to roll, then yes... I'd rather take the risk of a poor / low stat character to gain the possibility of a high stat character.Personally, I like random generation. The risk of a low-stat character is half the fun of the random method. But then, I am the type of person who likes to go at the game with a blank space, and fill in the character based on what I roll. It's very rare for me to start with a single concept that I really want. So random = good for me. I can't get down with a complex random method though. Or one that attempts to balance randomly generated stats. If I want balanced stats, I'll lobby for PB.

Post/Author/DateTime	Post
OmiKapSi 02-22-06, 11:22 PM	In my endless quest to create a system for a balanced but random attribute generation system, I've come up with the following iteration, which I decided to post here for feedback. Warning, basic math ahead. :) This is based around using the '4d6 drop the lowest' option for stat rolling, but in theory can be used with any system that generates random numbers between 3 and 18. The target is characters with a range of attributes who's total modifier is +4 to +6. (If you add up the modifiers of each characters attributes, your result will be +4, +5, or +6.) With this in mind, I decided that if all the characters attributes added up to 73, they were gauranteed to fall in the target range, depending on how many attributes were odd. So, the system. A player generates 5 attribute values using whatever system is preferred and records them in the order that they were generated. Like I mentioned, 4d6 works best for this part, but it's quite variable. Once the attributes have been generated, follow the steps. After each step is complete, use the modified attributes rather than the original rolls for following steps: Step A; Add the first and second attributes, if the total is less than 9, increase the second attribute by the difference between the total and 9 (making their new total 9). Step B; Add the first three attributes. If their total is is less than 25, increase the third attribute by the difference between the total and 25 (like in step A). Step C; Add the first four attributes. If their total is less than 41, increase the fourth attribute like in previous steps. If their total is greater than 63, then reduce the fourth attribute in the same manner. Step D; Add the five attributes. Adjust the fifth attribute as previous steps to make sure that the total is 57-68. Step E; Subtract the total after Step D from 73. This is the sixth attribute. e.g. Barak is rolling up a character. His rolls for the first 5 stats are 3, 4, 17, 15, 13. Step A: Bilbo's total for his first two attributes is only 7 (3+4). He increases his second roll, the 4, to a 6. Step B: Barak's total is 26 (3+6+17), so no adjustments are made. Step C: Barak's total is 41 (3+6+17+15). No changes. Step D: Barak's total is 54 (3+6+17+15+13), which is 3 lower than 57. He adjusts the 13 to a 16. Step E: Barak subtracts 57 from 73 to get 16 for his final attribute. His attributes are therefore: 3, 6, 17, 15, 16, 16. Silk then rolls up a character. His rolls for the first 5 stats are 14, 13, 16, 16, 17. Step A: No change. Step B: No change. Step C: No change. Step D: The total of the five attributes is 76, which is too high, the 17 gets bumped down by 8, making it a 9. Step E: 73-68=5. Silk's attributes are 14, 13, 16, 16, 9, 5. I think that this system works best for when attributes are arranged as desired after generation, however it can work just fine with them not being rearranged. So, what would be your response if your DM told you that the new campaign would be using this system for attribute generation?
Seeker95 02-23-06, 07:35 AM	Just an observation: What you are doing is a dice-driven point-buy. Not that this is a bad thing, but I would prefer point buy if the risk and benefit of rolling was gone (the real randomness of it all).
Thelandrach 02-23-06, 10:37 AM	Yikes almighty. Why not just use a 24 point buy? You can get 6 12's, which is +6, or two 14's, two 12's, and two 10's, which is also +6. Was it your intent to have the numbers evaluated in order, lowest to highest, so the lowest stats stay (some poor schmuck gets stuck with a 3 in your first example), and the highest ones get clobbered (if somebody legitimately rolled an 18, his chances of ending up with that stat are pretty much nil). If I introduced this to my players, they would (a) not understand it at all, (b) kill me, or (c) both. also: You have a player named Barak???
Grand Theft Otto 02-23-06, 11:32 AM	Let go of the security blanket. What you want is a balanced method for creating character stats. A covoluted random rolling isnt the answer. If point buy isnt your thing, I can suggest another option. Pick a number. Everyone's stat mods must add up to this number. Up to 3 of your ability scores can be odd. Some groups may desire to set low ranges for stats (ie, nothing below 6 or whatever). We used +7, and its worked out fine.
OmiKapSi 02-23-06, 11:57 AM	Seeker: So you'd rather take the risk of rolling a character who would be weaker than the group average for the potential benefit of rolling a character who could be stronger than the group average? I don't consider either result desirable. If there was a great need to have a more variable result (characters ranging from +2 to +8, for example), it would be a very simple matter to use 66+2d6, for example, instead of the base of 73, and modify all other numbers in the system appropriately. My reasons for not using point buy (or other choice based systems) are numerous. In this system, I'm not just seeking to have all attributes fall into a middle range. I actually kind of like having a few attributes at extremes, it makes for interesting characters, who have significant strengths and weaknesses. My goal is to have random attributes, but without the risk of having players dissatisfied with their rolls, or having players who's rolls are so high that they end up dominating the game. And the numbers are not evaluated in order from lowest to highest, they are evaluated in the order that they are rolled. If someone rolls an 18 in their first two rolls, the system guarantees that it won't be touched. It's simply a coincidence that the numbers in the second example are 'in order', the numbers in the first example are not.
sherretz 02-23-06, 02:48 PM	Your system has a few flaws also and I think the dice and/or point buy systems are better. First, if the player doesn't roll an 18 in the first two stats, they cannot save their 18 roll. It will take a while for the players to be wise on the system before they realize they can rearrange their rolls so that their higher rolls are preserved. Also, with a bit of work, an intelligent player can use the system to min/max his character. Since your Step C has a total range, they can max it to 63 points total and get higher stats through. Plus, above about 5th level, feats and skill points take over for ability scores with regards to the aptitude of the characters. So really, you're only guaranteeing a "balanced" character for each player until level 5 to 7. Thus in the end your system doesn't really make much of an impact on twinked characters. Really though, you should just talk to the players about min/maxing and twinking. If you understand why they are apt to do it, then you can give them reasons why it's not that great. Just because you like characters with single-digit ability scores, doesn't mean every player does. This appears to players like you are imposing your will on their PCs. Also keep in mind that this system will discourage players from playing characters that require a large number of average to above-average stats, such as Paladins and Monks.
Arrowhen 02-23-06, 03:22 PM	One thing I've been thinking about lately is rolling ability scores and then balancing them with action points: the lower your stats are, the more action points you get. This way, you'd still get the wider variety of ability scores afforded by rolling, but players who rolled poorly would have a larger pool of luck to fall back upon when it really mattered. I haven't played around with the numbers yet, because we've got a couple of months yet to go in my current campaign and the next on I run probably won't be D&D, but it's something I've been kicking around in the back of my mind. Any of you math-types want to take a stab at figuring out how much ability score points and action points are worth in relation to one another?
OmiKapSi 02-23-06, 03:58 PM	Since people have twice missed the fact that attributes are evaluated in the order that they are rolled, I've bolded it, hopefully this will prevent further confusion.
Seeker95 02-23-06, 10:18 PM	Seeker: So you'd rather take the risk of rolling a character who would be weaker than the group average for the potential benefit of rolling a character who could be stronger than the group average? I don't consider either result desirable.I prefer point buy. But if we're going to roll, then yes... I'd rather take the risk of a poor / low stat character to gain the possibility of a high stat character. Your method just makes the outcome similar to point buy, but complicated. If you want something other than point buy without the randomness of rolling, just hand the players 72 points (or whatever amount) and tell 'em to distribute as they wish.
sherretz 02-24-06, 09:01 AM	Since people have twice missed the fact that attributes are evaluated in the order that they are rolled, I've bolded it, hopefully this will prevent further confusion. Ok, I misread your post above mine. In that case, it makes the system even worse because you are penalizing players who roll an 18 after their first two rolls. I seriously think you're handicapping the players and PCs by doing this.
danzig138 02-24-06, 04:22 PM	I prefer point buy. But if we're going to roll, then yes... I'd rather take the risk of a poor / low stat character to gain the possibility of a high stat character.Personally, I like random generation. The risk of a low-stat character is half the fun of the random method. But then, I am the type of person who likes to go at the game with a blank space, and fill in the character based on what I roll. It's very rare for me to start with a single concept that I really want. So random = good for me. I can't get down with a complex random method though. Or one that attempts to balance randomly generated stats. If I want balanced stats, I'll lobby for PB.