4d6 drop the lowest is evil [Archive] - Wizards Community

4d6 drop the lowest is evil [Archive] - Wizards Community

Post/Author/DateTime Post
Sphyre
11-03-04, 06:06 PM So, I'm joining this online game, right? DM says allocation method is 4d6 drop the lowest, so we join an AIM chat to roll stats...

*OnlineHost* (4:44:39 PM): Sphyre San rolled 4 6-sided dice: 2 3 3 4
*OnlineHost* (4:44:40 PM): Sphyre San rolled 4 6-sided dice: 1 6 3 1
*OnlineHost* (4:44:41 PM): Sphyre San rolled 4 6-sided dice: 1 1 4 6
*OnlineHost* (4:44:42 PM): Sphyre San rolled 4 6-sided dice: 1 2 5 4
*OnlineHost* (4:44:43 PM): Sphyre San rolled 4 6-sided dice: 5 2 2 3
*OnlineHost* (4:44:45 PM): Sphyre San rolled 4 6-sided dice: 4 4 2 5
Sphyre San (4:44:54 PM): wow, i just have no luck with that method, huh?
Sphyre San (4:45:28 PM): 10 10 11 11 10 13!!!
Sphyre San (4:45:29 PM): woot!
Sphyre San (4:46:01 PM): "congratulations, you're joe six pack!"
Sucineri
11-03-04, 06:14 PM It's not so much evil as it is random. I personally like going with 4d6 drop lowest and reroll. Keep 3 highest.
Nom
11-03-04, 06:20 PM I don't see a score above 13. Invoke the rerolling rule, PH pg 8.
Sphyre
11-03-04, 06:23 PM I don't see a score above 13. Invoke the rerolling rule, PH pg 8.

I did reroll. Actually, the DM rolled for me... got me the stats: 10, 16, 11, 13, 9, 17
Ongorth
11-03-04, 06:27 PM I'm not seeing the problem here, could you explain it to me better?

I've got a level 13 character who started with 9,9,10,11,11,13, so I'm kinda lost. Are those stats too low, because I assume your not complaining that they are too high.
Lanky Ogre Baggy Pants
11-03-04, 06:50 PM according to page 8 in the PHB a character with their highest stat as a 13 or a total modifier of +0 is considered to have too low of stats and may reroll.
StormKnight
11-03-04, 06:52 PM Last time I tried pure random character gen, we were rolling 5d6, drop two lowest (we wanted a powerful character game). My high score was a 12 and I had one 4.
StevenO
11-03-04, 08:51 PM You have another example of why point buy stats can be good. This is a good illustration of how using four average d6s still produce barely above average scores. You may strike gold rolling stats or you can also just get dirt but with point buy you always know what you get.
Morwen
11-03-04, 08:59 PM It would also be possible to use some form of rolling dice, but add to the rules of when rerolling stats would be applicable, beyond the "highest is 13 and/or total ability mods 0 or less" rule. For instance, you could have a rule saying you could reroll if you don't have at least two scores of 14 or higher, or you could have a rule stating you can reroll if you don't have total ability score modifiers of +2 or better, or anything of your own choosing.

It is possible to have the same general powerlevel as pointbuy without sacrificing randomness (and everybody having 16-14-14-12-10-10 does get tedious after a while :)).
twoDtwenty
11-03-04, 09:29 PM Just be glad he didn't decide to use 3d6 place the in the order they fall, my personal favorite method from the days of 2nd ed.
kliate
11-03-04, 10:20 PM I've been nearly so cruel...

I set up a game as 3d6 in order, option to swap one pair. I reserve the right to determine arbitrarily if the stats qualify for re-roll...only one serious complaint thus far (naturally from the "...But I didn't get 3 18's" player)
Sphyre
11-03-04, 11:18 PM You have another example of why point buy stats can be good. This is a good illustration of how using four average d6s still produce barely above average scores. You may strike gold rolling stats or you can also just get dirt but with point buy you always know what you get.

actually, i think both methods have their good points and thier lesser points. It's not so much that being joe six pack is a bad thing... it's just when the player to player difference is so large that you're always taking a backseat when your stats come into play (obviously, when you're role playing it doesn't matter what stats you have, in fact I rolled a pretty low for another game and i'm having a blast role playing that character)

I much prefer blending the pointbuy and rolling methods.
Ricky
11-04-04, 10:25 AM For those who are interested, I wrote a short computer program to determine average results when rolling up a character with 4d6 drop lowest. According to this program, the sum of all your ability modifiers is +5.2 on average. This works with the "elite" array:

15: +2
14: +2
13: +1
12: +1
10: +0
8: -1
-------
+5

So if you have a character who total ability modifiers is +5, your character is average according to the 4d6 drop lowest method.

My house rule is that if you roll and your average modifiers total to less than +5, you can reroll all 6 scores. But that's just me. :)
digoraccoon
11-04-04, 11:57 AM Computer dice rollers aren't quite as random as real dice. I had that issue as well a few nights ago. nothing over a 12 and it liked rolling 7 a lot.

So the DM gave me some actual dice and we watched me roll nothing under a 15. Um... wow. :D

So now all the players are using real dice.
Stuntman
11-04-04, 12:55 PM Dice rolling is not evil. It is random.
WihHizwind
11-04-04, 03:07 PM In the game I'm currently in, were' going 4d6, drop lowest. You can do that twice, if you don't like either set, you can do a 24 buy.
Of course, when everyone rolls 15 as the average stat, no-one goes for the point-buy.

I love my dice =)
StormKnight
11-04-04, 06:16 PM Our current system (I'd credit this but I don't remember who posted it originally) is as follows:
1) Each stat begins at 6.
2) Roll one die for each stat and add the roll to the stat. Leave the dice after rolling them!
3) Now, flip each die directly over (so it shows 7 - original roll)
4) You can apply these new die rolls to any stats you wish. You can add any number of dice to one stat, as long as the result does not exceed 18.

It gives great results. I've found it usually lets you make approximately the character you want, but almost never lets you be exactly optimized or predictable the way a pure point system does. Best of both worlds.
Nom
11-04-04, 07:07 PM For those who are interested, I wrote a short computer program to determine average results when rolling up a character with 4d6 drop lowest. According to this program, the sum of all your ability modifiers is +5.2 on average. This works with the "elite" array:

15: +2
14: +2
13: +1
12: +1
10: +0
8: -1
-------
+5

So if you have a character who total ability modifiers is +5, your character is average according to the 4d6 drop lowest method.Ah, the classic munging of statistical terminology. It's the same problem of saying that the average household has 1.3 children. In fact, there's no such thing as a household with 1.3 children; rather the average number of children across all households is 1.3.

Without including rerolls (which affects about 1 in 4 characters rolled), the expected ability modifier is approximately 0.873, which when multiplied by 6 is approximately 5.2, which matches your result.

The problem is a huge variance (standard deviation of bonus is about 1.45) with a positive skew. If you roll 100 ability scores, you will get a nice spread. If you roll only 6, they can be all over the place.
My house rule is that if you roll and your average modifiers total to less than +5, you can reroll all 6 scores.So you expect to reroll just under half of all characters? This introduces massive positive skew, though it does reduce the variance by upping the lower bound without moving the upper bound.
One Wolf
11-04-04, 08:56 PM Ah, the classic munging of statistical terminology. It's the same problem of saying that the average household has 1.3 children. In fact, there's no such thing as a household with 1.3 children; rather the average number of children across all households is 1.3.

Without including rerolls (which affects about 1 in 4 characters rolled), the expected ability modifier is approximately 0.873, which when multiplied by 6 is approximately 5.2, which matches your result.

The problem is a huge variance (standard deviation of bonus is about 1.45) with a positive skew. If you roll 100 ability scores, you will get a nice spread. If you roll only 6, they can be all over the place.
So you expect to reroll just under half of all characters? This introduces massive positive skew, though it does reduce the variance by upping the lower bound without moving the upper bound.

Statisticians! You gotta love 'em!!!

-One Wolf
darczero
11-04-04, 11:29 PM My house rule is that if you roll and your average modifiers total to less than +5, you can reroll all 6 scores. But that's just me. :)

So that means that I get to reroll any character with Ability Modifiers across all stats totaling less than 30. Nice. Is that before Racial/Template adjustments?

IMX, with that 4d6 Drop Lowest method, computer dice-rolling programs are designed to give statistically average rolls (every possible result), usually over an obscene number (like 10,000+) of rolls; this does, if used for long enough, give "Average" results, but if used for short-term, those programs often give wildly unfavorable results (one player may get 18's across the board, while another's Total Ability Modifiers for all six stats may be -8). Statistically fair, but in practice, very bad idea. Even if you tell it to Reroll under the prescribed circumstances.

IMC, we don't use Point Buy, mostly because my players like the small RP challenge that low rolled numbers can provide, and because I (as a DM) prefer something a little bit more powerful than a 40 point buy would offer, and really can't settle on a decent amount (I've personally rolled NPC's/characters that had about 71 "Points" in their Ability Scores). We use 4d6, reroll all 1's, drop the lowest. It increases the probability of a better character (bumps the average up to about 13 instead of ~11).

And as for the "Elite" array presented in the DMG, that's all well and good for average PC-classed NPC's (NPC Fighter, NPC Rogue, NPC Cleric, etc.), but I don't use it for specific NPC's (like any BBEG). With them, I still roll; occasionally, just for variety, I'll apply the "Elite Array" to a Monster, just to change things (all Monsters have either 10 or 11 before Racial Adjustments, and while this does give a theoretical advantage to PC's, it's lame).
fractal
11-05-04, 01:23 AM My group has had the most fun with stat draft, where each player rolls a str, dex, con, etc in order, and puts them into a communal pool of str's, dex's, etc. Then players take turn drafting a stat for their character, until each player has one of each stat.

Stat draft is fun, because it keeps the randomness of rolling, together with the fairness of everyone having comparable stats.
Mr. Wilson
11-05-04, 08:09 AM I like that idea. I may have to think about using that.

Last time I used 4d6 drop the lowest I managed to hit the bottom of the barrel, 4, 5, 10, 11, 13, 14. My DM (to his credit) took one look at that an called for an immediate reroll. This time, it was much better 7, 9, 14, 16, 17, 17. I almost felt bad using the second set.
Ricky
11-05-04, 10:15 AM The problem is a huge variance (standard deviation of bonus is about 1.45) with a positive skew. If you roll 100 ability scores, you will get a nice spread. If you roll only 6, they can be all over the place.

I thought of that. The computer program rolled 6 stats for a character and summed the ability modifiers. It then repeated this process for one million characters and calculated the average of the sums. I figured one million samples would eliminate any transient variations. I can show you the code if you're interested.

So you expect to reroll just under half of all characters?

Yup! It works, I tell ya. :)

So that means that I get to reroll any character with Ability Modifiers across all stats totaling less than 30.

I think you misunderstood. It's not +5 for each ability score. It's a total of +5 for all your 6 ability scores.
Sphyre
11-05-04, 07:23 PM Dice rolling is not evil. It is random.

lol

well you know... people throw those terms around so easily... i mean D&D isn't evil... it's just random ;)
Palin the Wizard
11-05-04, 09:52 PM Our preferred method of rolling is 4d6 drop lowest, reroll 1s. It gives a minumum of 6, rather than 3.

Of course, I generally prefer a simple point-by-point system (not the point-buy system!), with 84-86 points total. D&D characters are meant to be heroes.

I've also seen other systems implemented with great success. 1 18, 1 15, 50 points to distribute on each of the other stats (and the 18 and the 15 may not be increased or decreased) is one I like a lot.
Morwen
11-05-04, 10:01 PM Of course, I generally prefer a simple point-by-point system (not the point-buy system!), with 84-86 points total. D&D characters are meant to be heroes.

That produces an average ability score of 14 across the board.. Seems a bit powered :)
fractal
11-06-04, 12:04 AM <84-86 Total stats> That produces an average ability score of 14 across the board.. Seems a bit powered :)
It's worse than that, really, because there's no incentive for players not to generate wildly skewed stats. 18, 18, 18, 18, 8, 6 can happen easily with this system. I strongly recommend Point Buy over a 1 for 1 stat selection like this.
Ras Kass
11-06-04, 01:03 AM Another option to use is 4d6, taking the highest 2 and add 6 because the lowest score you can have is 8 and you get a free 6. Plus, the average is around 13 to 14, which is pretty good.
fractal
11-06-04, 01:29 AM Another option to use is 4d6, taking the highest 2 and add 6 because the lowest score you can have is 8 and you get a free 6. Plus, the average is around 13 to 14, which is pretty good.
Yeah, that will lead to very high powered characters, though. The average will be a lot higher than 13-14, the average would be 13 if you rolled 2d6+6. It also has the same problem as standard rolling - some people will roll a lot better than others.
Ras Kass
11-06-04, 01:50 AM Yeah, that will lead to very high powered characters, though. The average will be a lot higher than 13-14, the average would be 13 if you rolled 2d6+6. It also has the same problem as standard rolling - some people will roll a lot better than others.

Well, if it is important that everybody have relatively the same power in attributes you could always go all 18's :D . This way no one can complain that you are detracting from optimum starting attributes and everybody is the same. hehe
Ras Kass
11-06-04, 02:11 AM One more thing, about the averages of 4d6, highest 2 plus 6.

When you said the average was more than 13-14, I decided to look into it. I did 100 rolls, 10 sets of 10 to be exact. They were as follows:

16.3 14.9 16.1 15.3 15.8 14.9 15.1 14.7 14.4 15.3

Average = 15.28, or a 15.3

Things of note:
- No roll was ever below 10, though this is statistically possible, ranging from 8 to 18 in possibilities.
- No average set was below 14
- This data is all based on 100 rolls, which isn't very specific but still good to use.
- The rolling was used by computer, site being Invisible Castle (http://www.invisiblecastle.com/)

Well, uh, that's all for now.

- One love
Sphyre
11-06-04, 06:00 AM Well, if it is important that everybody have relatively the same power in attributes you could always go all 18's :D . This way no one can complain that you are detracting from optimum starting attributes and everybody is the same. hehe

Actually... i'd rather be the joe six pack i rolled up than everyone being 18s. Nothing cheapens the game for me like overpowered characters. I always think up really powerful combos... and never use them because once I use them... the DM is entitled to challenge us, and if he were to ever pull something that was immune to the one trick pony... I have nothing left in my chararcter's disposal to deal with it.
Palin the Wizard
11-06-04, 01:43 PM It's worse than that, really, because there's no incentive for players not to generate wildly skewed stats. 18, 18, 18, 18, 8, 6 can happen easily with this system. I strongly recommend Point Buy over a 1 for 1 stat selection like this.
That's funny, because I've played in over a dozen of games with this exact method of stats, and I've never seen what you're proposing.

I've seen 10, 18, 12, 18, 10, 18, but that's the worst case so far.
86 points on a point-by-point basis ranges from 40 to 70 point-buy, unless I miss my guess.

Point-buy is not for me, nor for the people I usually play with. We prefer this method, and we're doing just fine, thankyouverymuch.
BenjCano
11-06-04, 02:20 PM My DM uses the "elite plus five" method. We start with the elite array, and then we have five points by which we can increase any stat on a 1-to-1 basis, subject to normal racial modifiers.
fractal
11-07-04, 05:37 PM I've seen 10, 18, 12, 18, 10, 18, but that's the worst case so far.
86 points on a point-by-point basis ranges from 40 to 70 point-buy, unless I miss my guess.

Point-buy is not for me, nor for the people I usually play with. We prefer this method, and we're doing just fine, thankyouverymuch.
Out of curiousity, how come you prefer this system (1:1 stat purchases) over a comparable point buy (50-60 points)?

Not that I think Point Buy is perfect, as it certainly isn't (it would be much better if it started one point lower, on 7, an odd stat), but it does encourage a more normal distribution of scores than a 1:1 purchase. Given that your objection to it is obviously not the usual (that you want to roll dice to get your stats), I'm interested to hear your thoughts.
Veton
11-07-04, 09:40 PM You have another example of why point buy stats can be good. This is a good illustration of how using four average d6s still produce barely above average scores. You may strike gold rolling stats or you can also just get dirt but with point buy you always know what you get.

Yup. I actually have have one character (who I'll be playing soon) with really great stats I rolled with 4d6 drop lowest, but I also had another character that I rolled rather poorly on in a different campaign with the same method.

So now I opt for point buy, except that the DM I play with seems to dislike it, and only allows me to use it if I have at least one score be a 8 or 9. I still like the stats for my new character better then the previous one though.
Nom
11-08-04, 12:12 AM I thought of that. The computer program rolled 6 stats for a character and summed the ability modifiers. It then repeated this process for one million characters and calculated the average of the sums. I figured one million samples would eliminate any transient variations. I can show you the code if you're interested.Considering that we know the exact generation method (best 3 of 4 d6) and that enumerating all possible samples requires only 6^4 = 1296 samples, why did you run a million sample monte-carlo simulation? :confused:
Ricky
11-08-04, 11:15 AM Considering that we know the exact generation method (best 3 of 4 d6) and that enumerating all possible samples requires only 6^4 = 1296 samples, why did you run a million sample monte-carlo simulation? :confused:

Because I'm not a statistician and so I didn't think of that, I guess. :)
It only took a few minutes for my computer to run the simulation, so I guess it's not a big deal.

Post/Author/DateTime	Post
Sphyre 11-03-04, 06:06 PM	So, I'm joining this online game, right? DM says allocation method is 4d6 drop the lowest, so we join an AIM chat to roll stats... OnlineHost (4:44:39 PM): Sphyre San rolled 4 6-sided dice: 2 3 3 4 OnlineHost (4:44:40 PM): Sphyre San rolled 4 6-sided dice: 1 6 3 1 OnlineHost (4:44:41 PM): Sphyre San rolled 4 6-sided dice: 1 1 4 6 OnlineHost (4:44:42 PM): Sphyre San rolled 4 6-sided dice: 1 2 5 4 OnlineHost (4:44:43 PM): Sphyre San rolled 4 6-sided dice: 5 2 2 3 OnlineHost (4:44:45 PM): Sphyre San rolled 4 6-sided dice: 4 4 2 5 Sphyre San (4:44:54 PM): wow, i just have no luck with that method, huh? Sphyre San (4:45:28 PM): 10 10 11 11 10 13!!! Sphyre San (4:45:29 PM): woot! Sphyre San (4:46:01 PM): "congratulations, you're joe six pack!"
Sucineri 11-03-04, 06:14 PM	It's not so much evil as it is random. I personally like going with 4d6 drop lowest and reroll. Keep 3 highest.
Nom 11-03-04, 06:20 PM	I don't see a score above 13. Invoke the rerolling rule, PH pg 8.
Sphyre 11-03-04, 06:23 PM	I don't see a score above 13. Invoke the rerolling rule, PH pg 8. I did reroll. Actually, the DM rolled for me... got me the stats: 10, 16, 11, 13, 9, 17
Ongorth 11-03-04, 06:27 PM	I'm not seeing the problem here, could you explain it to me better? I've got a level 13 character who started with 9,9,10,11,11,13, so I'm kinda lost. Are those stats too low, because I assume your not complaining that they are too high.
Lanky Ogre Baggy Pants 11-03-04, 06:50 PM	according to page 8 in the PHB a character with their highest stat as a 13 or a total modifier of +0 is considered to have too low of stats and may reroll.
StormKnight 11-03-04, 06:52 PM	Last time I tried pure random character gen, we were rolling 5d6, drop two lowest (we wanted a powerful character game). My high score was a 12 and I had one 4.
StevenO 11-03-04, 08:51 PM	You have another example of why point buy stats can be good. This is a good illustration of how using four average d6s still produce barely above average scores. You may strike gold rolling stats or you can also just get dirt but with point buy you always know what you get.
Morwen 11-03-04, 08:59 PM	It would also be possible to use some form of rolling dice, but add to the rules of when rerolling stats would be applicable, beyond the "highest is 13 and/or total ability mods 0 or less" rule. For instance, you could have a rule saying you could reroll if you don't have at least two scores of 14 or higher, or you could have a rule stating you can reroll if you don't have total ability score modifiers of +2 or better, or anything of your own choosing. It is possible to have the same general powerlevel as pointbuy without sacrificing randomness (and everybody having 16-14-14-12-10-10 does get tedious after a while :)).
twoDtwenty 11-03-04, 09:29 PM	Just be glad he didn't decide to use 3d6 place the in the order they fall, my personal favorite method from the days of 2nd ed.
kliate 11-03-04, 10:20 PM	I've been nearly so cruel... I set up a game as 3d6 in order, option to swap one pair. I reserve the right to determine arbitrarily if the stats qualify for re-roll...only one serious complaint thus far (naturally from the "...But I didn't get 3 18's" player)
Sphyre 11-03-04, 11:18 PM	You have another example of why point buy stats can be good. This is a good illustration of how using four average d6s still produce barely above average scores. You may strike gold rolling stats or you can also just get dirt but with point buy you always know what you get. actually, i think both methods have their good points and thier lesser points. It's not so much that being joe six pack is a bad thing... it's just when the player to player difference is so large that you're always taking a backseat when your stats come into play (obviously, when you're role playing it doesn't matter what stats you have, in fact I rolled a pretty low for another game and i'm having a blast role playing that character) I much prefer blending the pointbuy and rolling methods.
Ricky 11-04-04, 10:25 AM	For those who are interested, I wrote a short computer program to determine average results when rolling up a character with 4d6 drop lowest. According to this program, the sum of all your ability modifiers is +5.2 on average. This works with the "elite" array: 15: +2 14: +2 13: +1 12: +1 10: +0 8: -1 ------- +5 So if you have a character who total ability modifiers is +5, your character is average according to the 4d6 drop lowest method. My house rule is that if you roll and your average modifiers total to less than +5, you can reroll all 6 scores. But that's just me. :)
digoraccoon 11-04-04, 11:57 AM	Computer dice rollers aren't quite as random as real dice. I had that issue as well a few nights ago. nothing over a 12 and it liked rolling 7 a lot. So the DM gave me some actual dice and we watched me roll nothing under a 15. Um... wow. :D So now all the players are using real dice.
Stuntman 11-04-04, 12:55 PM	Dice rolling is not evil. It is random.
WihHizwind 11-04-04, 03:07 PM	In the game I'm currently in, were' going 4d6, drop lowest. You can do that twice, if you don't like either set, you can do a 24 buy. Of course, when everyone rolls 15 as the average stat, no-one goes for the point-buy. I love my dice =)
StormKnight 11-04-04, 06:16 PM	Our current system (I'd credit this but I don't remember who posted it originally) is as follows: 1) Each stat begins at 6. 2) Roll one die for each stat and add the roll to the stat. Leave the dice after rolling them! 3) Now, flip each die directly over (so it shows 7 - original roll) 4) You can apply these new die rolls to any stats you wish. You can add any number of dice to one stat, as long as the result does not exceed 18. It gives great results. I've found it usually lets you make approximately the character you want, but almost never lets you be exactly optimized or predictable the way a pure point system does. Best of both worlds.
Nom 11-04-04, 07:07 PM	For those who are interested, I wrote a short computer program to determine average results when rolling up a character with 4d6 drop lowest. According to this program, the sum of all your ability modifiers is +5.2 on average. This works with the "elite" array: 15: +2 14: +2 13: +1 12: +1 10: +0 8: -1 ------- +5 So if you have a character who total ability modifiers is +5, your character is average according to the 4d6 drop lowest method.Ah, the classic munging of statistical terminology. It's the same problem of saying that the average household has 1.3 children. In fact, there's no such thing as a household with 1.3 children; rather the average number of children across all households is 1.3. Without including rerolls (which affects about 1 in 4 characters rolled), the expected ability modifier is approximately 0.873, which when multiplied by 6 is approximately 5.2, which matches your result. The problem is a huge variance (standard deviation of bonus is about 1.45) with a positive skew. If you roll 100 ability scores, you will get a nice spread. If you roll only 6, they can be all over the place. My house rule is that if you roll and your average modifiers total to less than +5, you can reroll all 6 scores.So you expect to reroll just under half of all characters? This introduces massive positive skew, though it does reduce the variance by upping the lower bound without moving the upper bound.
One Wolf 11-04-04, 08:56 PM	Ah, the classic munging of statistical terminology. It's the same problem of saying that the average household has 1.3 children. In fact, there's no such thing as a household with 1.3 children; rather the average number of children across all households is 1.3. Without including rerolls (which affects about 1 in 4 characters rolled), the expected ability modifier is approximately 0.873, which when multiplied by 6 is approximately 5.2, which matches your result. The problem is a huge variance (standard deviation of bonus is about 1.45) with a positive skew. If you roll 100 ability scores, you will get a nice spread. If you roll only 6, they can be all over the place. So you expect to reroll just under half of all characters? This introduces massive positive skew, though it does reduce the variance by upping the lower bound without moving the upper bound. Statisticians! You gotta love 'em!!! -One Wolf
darczero 11-04-04, 11:29 PM	My house rule is that if you roll and your average modifiers total to less than +5, you can reroll all 6 scores. But that's just me. :) So that means that I get to reroll any character with Ability Modifiers across all stats totaling less than 30. Nice. Is that before Racial/Template adjustments? IMX, with that 4d6 Drop Lowest method, computer dice-rolling programs are designed to give statistically average rolls (every possible result), usually over an obscene number (like 10,000+) of rolls; this does, if used for long enough, give "Average" results, but if used for short-term, those programs often give wildly unfavorable results (one player may get 18's across the board, while another's Total Ability Modifiers for all six stats may be -8). Statistically fair, but in practice, very bad idea. Even if you tell it to Reroll under the prescribed circumstances. IMC, we don't use Point Buy, mostly because my players like the small RP challenge that low rolled numbers can provide, and because I (as a DM) prefer something a little bit more powerful than a 40 point buy would offer, and really can't settle on a decent amount (I've personally rolled NPC's/characters that had about 71 "Points" in their Ability Scores). We use 4d6, reroll all 1's, drop the lowest. It increases the probability of a better character (bumps the average up to about 13 instead of ~11). And as for the "Elite" array presented in the DMG, that's all well and good for average PC-classed NPC's (NPC Fighter, NPC Rogue, NPC Cleric, etc.), but I don't use it for specific NPC's (like any BBEG). With them, I still roll; occasionally, just for variety, I'll apply the "Elite Array" to a Monster, just to change things (all Monsters have either 10 or 11 before Racial Adjustments, and while this does give a theoretical advantage to PC's, it's lame).
fractal 11-05-04, 01:23 AM	My group has had the most fun with stat draft, where each player rolls a str, dex, con, etc in order, and puts them into a communal pool of str's, dex's, etc. Then players take turn drafting a stat for their character, until each player has one of each stat. Stat draft is fun, because it keeps the randomness of rolling, together with the fairness of everyone having comparable stats.
Mr. Wilson 11-05-04, 08:09 AM	I like that idea. I may have to think about using that. Last time I used 4d6 drop the lowest I managed to hit the bottom of the barrel, 4, 5, 10, 11, 13, 14. My DM (to his credit) took one look at that an called for an immediate reroll. This time, it was much better 7, 9, 14, 16, 17, 17. I almost felt bad using the second set.
Ricky 11-05-04, 10:15 AM	The problem is a huge variance (standard deviation of bonus is about 1.45) with a positive skew. If you roll 100 ability scores, you will get a nice spread. If you roll only 6, they can be all over the place. I thought of that. The computer program rolled 6 stats for a character and summed the ability modifiers. It then repeated this process for one million characters and calculated the average of the sums. I figured one million samples would eliminate any transient variations. I can show you the code if you're interested. So you expect to reroll just under half of all characters? Yup! It works, I tell ya. :) So that means that I get to reroll any character with Ability Modifiers across all stats totaling less than 30. I think you misunderstood. It's not +5 for each ability score. It's a total of +5 for all your 6 ability scores.
Sphyre 11-05-04, 07:23 PM	Dice rolling is not evil. It is random. lol well you know... people throw those terms around so easily... i mean D&D isn't evil... it's just random ;)
Palin the Wizard 11-05-04, 09:52 PM	Our preferred method of rolling is 4d6 drop lowest, reroll 1s. It gives a minumum of 6, rather than 3. Of course, I generally prefer a simple point-by-point system (not the point-buy system!), with 84-86 points total. D&D characters are meant to be heroes. I've also seen other systems implemented with great success. 1 18, 1 15, 50 points to distribute on each of the other stats (and the 18 and the 15 may not be increased or decreased) is one I like a lot.
Morwen 11-05-04, 10:01 PM	Of course, I generally prefer a simple point-by-point system (not the point-buy system!), with 84-86 points total. D&D characters are meant to be heroes. That produces an average ability score of 14 across the board.. Seems a bit powered :)
fractal 11-06-04, 12:04 AM	<84-86 Total stats> That produces an average ability score of 14 across the board.. Seems a bit powered :) It's worse than that, really, because there's no incentive for players not to generate wildly skewed stats. 18, 18, 18, 18, 8, 6 can happen easily with this system. I strongly recommend Point Buy over a 1 for 1 stat selection like this.
Ras Kass 11-06-04, 01:03 AM	Another option to use is 4d6, taking the highest 2 and add 6 because the lowest score you can have is 8 and you get a free 6. Plus, the average is around 13 to 14, which is pretty good.
fractal 11-06-04, 01:29 AM	Another option to use is 4d6, taking the highest 2 and add 6 because the lowest score you can have is 8 and you get a free 6. Plus, the average is around 13 to 14, which is pretty good. Yeah, that will lead to very high powered characters, though. The average will be a lot higher than 13-14, the average would be 13 if you rolled 2d6+6. It also has the same problem as standard rolling - some people will roll a lot better than others.
Ras Kass 11-06-04, 01:50 AM	Yeah, that will lead to very high powered characters, though. The average will be a lot higher than 13-14, the average would be 13 if you rolled 2d6+6. It also has the same problem as standard rolling - some people will roll a lot better than others. Well, if it is important that everybody have relatively the same power in attributes you could always go all 18's :D . This way no one can complain that you are detracting from optimum starting attributes and everybody is the same. hehe
Ras Kass 11-06-04, 02:11 AM	One more thing, about the averages of 4d6, highest 2 plus 6. When you said the average was more than 13-14, I decided to look into it. I did 100 rolls, 10 sets of 10 to be exact. They were as follows: 16.3 14.9 16.1 15.3 15.8 14.9 15.1 14.7 14.4 15.3 Average = 15.28, or a 15.3 Things of note: - No roll was ever below 10, though this is statistically possible, ranging from 8 to 18 in possibilities. - No average set was below 14 - This data is all based on 100 rolls, which isn't very specific but still good to use. - The rolling was used by computer, site being Invisible Castle (http://www.invisiblecastle.com/) Well, uh, that's all for now. - One love
Sphyre 11-06-04, 06:00 AM	Well, if it is important that everybody have relatively the same power in attributes you could always go all 18's :D . This way no one can complain that you are detracting from optimum starting attributes and everybody is the same. hehe Actually... i'd rather be the joe six pack i rolled up than everyone being 18s. Nothing cheapens the game for me like overpowered characters. I always think up really powerful combos... and never use them because once I use them... the DM is entitled to challenge us, and if he were to ever pull something that was immune to the one trick pony... I have nothing left in my chararcter's disposal to deal with it.
Palin the Wizard 11-06-04, 01:43 PM	It's worse than that, really, because there's no incentive for players not to generate wildly skewed stats. 18, 18, 18, 18, 8, 6 can happen easily with this system. I strongly recommend Point Buy over a 1 for 1 stat selection like this. That's funny, because I've played in over a dozen of games with this exact method of stats, and I've never seen what you're proposing. I've seen 10, 18, 12, 18, 10, 18, but that's the worst case so far. 86 points on a point-by-point basis ranges from 40 to 70 point-buy, unless I miss my guess. Point-buy is not for me, nor for the people I usually play with. We prefer this method, and we're doing just fine, thankyouverymuch.
BenjCano 11-06-04, 02:20 PM	My DM uses the "elite plus five" method. We start with the elite array, and then we have five points by which we can increase any stat on a 1-to-1 basis, subject to normal racial modifiers.
fractal 11-07-04, 05:37 PM	I've seen 10, 18, 12, 18, 10, 18, but that's the worst case so far. 86 points on a point-by-point basis ranges from 40 to 70 point-buy, unless I miss my guess. Point-buy is not for me, nor for the people I usually play with. We prefer this method, and we're doing just fine, thankyouverymuch. Out of curiousity, how come you prefer this system (1:1 stat purchases) over a comparable point buy (50-60 points)? Not that I think Point Buy is perfect, as it certainly isn't (it would be much better if it started one point lower, on 7, an odd stat), but it does encourage a more normal distribution of scores than a 1:1 purchase. Given that your objection to it is obviously not the usual (that you want to roll dice to get your stats), I'm interested to hear your thoughts.
Veton 11-07-04, 09:40 PM	You have another example of why point buy stats can be good. This is a good illustration of how using four average d6s still produce barely above average scores. You may strike gold rolling stats or you can also just get dirt but with point buy you always know what you get. Yup. I actually have have one character (who I'll be playing soon) with really great stats I rolled with 4d6 drop lowest, but I also had another character that I rolled rather poorly on in a different campaign with the same method. So now I opt for point buy, except that the DM I play with seems to dislike it, and only allows me to use it if I have at least one score be a 8 or 9. I still like the stats for my new character better then the previous one though.
Nom 11-08-04, 12:12 AM	I thought of that. The computer program rolled 6 stats for a character and summed the ability modifiers. It then repeated this process for one million characters and calculated the average of the sums. I figured one million samples would eliminate any transient variations. I can show you the code if you're interested.Considering that we know the exact generation method (best 3 of 4 d6) and that enumerating all possible samples requires only 6^4 = 1296 samples, why did you run a million sample monte-carlo simulation? :confused:
Ricky 11-08-04, 11:15 AM	Considering that we know the exact generation method (best 3 of 4 d6) and that enumerating all possible samples requires only 6^4 = 1296 samples, why did you run a million sample monte-carlo simulation? :confused: Because I'm not a statistician and so I didn't think of that, I guess. :) It only took a few minutes for my computer to run the simulation, so I guess it's not a big deal.