I have a question for all the other DMs/GMs/World Creators out there on DDB, regarding methods for determining ability scores.  First let me say that this is a variation of a system used by Bill Allen. It is not one that I have developed on my own, rather it is a system that I asked if I could use and then modified. So, without further adieu, here it is:

First, you roll 4d6, if you get any 1s - you reroll those. You keep rerolling 1s until you get 4 numbers between 2 and 6.

Next, you take the three highest numbers and add them together so you get a single number that is the combination of the three highest numbers you rolled from 4d6.

Repeat the steps above until you have a set of 6 numbers.

Now you repeat the whole process two more times, so you have three sets of 6 numbers.

Add together each of the three sets of six numbers so you get a total for each set and the set with the highest total is one you use for your ability scores. Here's an example:

Now you would add the totals up like this:

So in this case, it would be Set 3 with a total of 81 that you would use as your ability scores. In Set 3, we can see that we have two really high numbers - two 18s but the other numbers are all relatively low - 10, 12 and 13 but when we add in our racial modifiers and well as any increases we get from our class(s) from Ability Score Increases (if we start at higher levels), this rounds out some of the lower numbers.

Using this system, we can create truly heroic characters that start out with stats determined largely by chance, while avoiding the standard array and point buy systems - which I personally hate.   I am looking forward to all you comments below and please do feel free to answer the Quick Poll up top too

So what do you think of this Heroic System for determining ability scores?

Seems like a lot of work just to ensure you have high numbers. Why not just start with a bunch of points to dole out for scores if you want heroic-level characters?

EDIT: I'm not negging your idea and I should consider that some people really enjoy the dice-rolling process for character creation.

I really hate the point buy system because it is not really down to chance.  You get these points and can put them wherever you like, there's not really any chance involved.  The same with the standard array, although that gets around the problem of players deciding they are going to have 20 in this stat and this stat and 8 in everything else (simplified from what I have seen players doing with point buy)

The Heroic System, increases the chance that you will get higher number to put in your stats but it doesn't mean you will.  There is an element of chance involved.  In the dice I rolled for my example, Set 3 had two really high numbers but the rest were pretty average.  If I rolled that again, I might not get 18s but I might get 14s, 15s and 16s (which would be a more rounded character), so it is giving ability stats an element of chance, while increasing the chance of getting a higher number.

This system also encourages more roleplay as you have to roleplay more to make up for any low numbers you get.  Or if you rolled your dice and got all 10s, 12s and 13s - well that's a totally different character mechanically and is going to involve a lot of roleplay to pull it off.

In short, the Heroic System encourages players to delve deeper into the game and actually play their characters, instead of just being concerned with stats.

Or that is my intention anyway.

I think the point trying to be made is that it is convoluted when you can get similar results by using something like 3d6 drop the lowest, add 6. You will still have an average roll of 14.46 versus 13.43 using the method you described, and the lowest roll possible is an 8 versus 6 using what you described. It retains all of its chance aspects while not requiring you to roll a d6 72 times just to build a character. The one interesting thing that what you described does do well though is that it forces you to take the array with the larger total score, which could have one or two horrendous scores and some really epic ones...but if you were to crunch the math I would be willing to bet that you'd end up taking well rounded characters more often than characters with large distribution gaps between stats, implying that the system only promises to encourage roleplay in theory, while in reality the players often end up with balanced stats.

An example: the average total score for the system you described is 80.58, while the median is 72, meaning that a point distribution of 18, 18, 18, 6, 6, 6, which matches your median score, would be less likely to appear. This goes to show that scores at or below the median are much less likely to occur while something like 18, 18, 18, 9, 9, 9  (81 points) is going to be something much more likely to occur. So basically what you end up with is a system that edges on having characters with extreme distributions that have average or barely below average stats, while having godly stats in others. More or less taking the highest score set is forcing you to take something approaching a more homogenized distribution when it sits close to or above the average hence the illusion of the system encouraging roleplaying. Essentially I would say that the further that the median diverges from the average, the more homogenized the characters become, meaning that epic systems are bound to fail at reliably generating characters with the kinds of extreme divergences that really do influence roleplay.

I've been using roll 4d6 take the highest 3 and put them where you want since the 70s.

Everybody has their own system. Use the one you like best =)

Quote from Wysperra >>

I've been using roll 4d6 take the highest 3 and put them where you want since the 70s.

Everybody has their own system. Use the one you like best =)

 
Yeah, the difference is that they are doing 4d6, drop the lowest, reroll ones AND doing it for each stat 3 times.. That is a lot of rolling to generate a character...hell it is basically just rolling 3 characters using the method you described and taking the one with the highest score total. Like it can make for some strong characters which is the intention, but based on other things the OP said I think that it fails since the stat distribution chances more towards having evenly distributed stats which defeats the purpose of allowing randomness to create interesting roleplay characters.

Quote from AtlaStar >>

Quote from Wysperra >>

I've been using roll 4d6 take the highest 3 and put them where you want since the 70s.

Everybody has their own system. Use the one you like best =)

 
Yeah, the difference is that they are doing 4d6, drop the lowest, reroll ones AND doing it for each stat 3 times.. That is a lot of rolling to generate a character...hell it is basically just rolling 3 characters using the method you described and taking the one with the highest score total. Like it can make for some strong characters which is the intention, but based on other things the OP said I think that it fails since the stat distribution chances more towards having evenly distributed stats which defeats the purpose of allowing randomness to create interesting roleplay characters.

I have actually further refined this system to get a very similar result with less dice rolls and less math.  Upon further considering why I was doing what I was doing - I decided to change the system entirely.

The system that put in its place now seems more polished and doesn't seem as though I am gaming the dice as much.

What I did was say, everyone gets commoners stats as standard - 10 in everything automatically.  Then you take 1d4 and roll it (you don't reroll 1s this time) and add the number to your 10.  So the maximum you can get before any racial/class modifiers are applied is 14 and the lowest is 11.

So if you were playing a standard Human your stat range after racial modifiers, could be between 11 and 15.  With an average of 13.

There is also a lot more to the system, regarding ASIs and feats.  And along with the system for generating stats, I also overhauled the magic scaling system, for a lower magic setting.

I will get all my notes into logical order, put them into digital form and then post the overhauled stat generation/feat/magic scaling systems for a low magic (heroic) world setting.

Quote from SocialFoxes >>

I really hate the point buy system because it is not really down to chance.  You get these points and can put them wherever you like, there's not really any chance involved.  The same with the standard array, although that gets around the problem of players deciding they are going to have 20 in this stat and this stat and 8 in everything else (simplified from what I have seen players doing with point buy)

The Heroic System, increases the chance that you will get higher number to put in your stats but it doesn't mean you will.  There is an element of chance involved.  In the dice I rolled for my example, Set 3 had two really high numbers but the rest were pretty average.  If I rolled that again, I might not get 18s but I might get 14s, 15s and 16s (which would be a more rounded character), so it is giving ability stats an element of chance, while increasing the chance of getting a higher number.

This system also encourages more roleplay as you have to roleplay more to make up for any low numbers you get.  Or if you rolled your dice and got all 10s, 12s and 13s - well that's a totally different character mechanically and is going to involve a lot of roleplay to pull it off.

In short, the Heroic System encourages players to delve deeper into the game and actually play their characters, instead of just being concerned with stats.

Or that is my intention anyway.

I don't find that rolling for stats increases group fun, because the ability scores do matter in play, quite a lot. Someone who happens to be the only one at the table who doesn't get any 18s while everyone else gets two is probably not going to feel like "ooh, awesome -- the dice gave me a chance to role-play!".

This can be fun if you tend to do many short sessions where characters don't last long — either everyone rerolls frequently and you take turns shining or sucking, or else you've got a highly-lethal game where the low-ability characters tend to die off anyway and there's a literal survival of the fittest until everyone has high stats they're happy with. But for a single event — or for a long campaign, might as well just replace some people's d20s with d12s.

I like the point buy system because, as others have said, random roles can make some in the group feel cheated.  (I do standard progression hit points too)

I know there's a "sameness" to the point buy system, so I'm one of the folks that gives everyone a feat at 1st level.  It feels like that's a good way for characters to distinguish themselves.

Quote from Generic_Poster >>

I like the point buy system because, as others have said, random roles can make some in the group feel cheated.

I know there's a "sameness" to the point buy system, so I'm one of the folks that gives everyone a feat at 1st level.  It feels like that's a good way for characters to distinguish themselves.

Me too, sometimes. I think this is one of the best ways to add a heroic feel to low-level characters.

SocialFoxes: The only thing I would be concerned about is that while a straight 1d4 roll has equal distribution on a per stat basis, the total score will follow a bell curve that is really heavily weighted towards the median due to the narrow range of a d4 having similar issues of homogenization I mentioned before. The larger the die used, the wider and flatter the overall curve becomes. So something like setting the base scores at 8 and adding a d10 is going to result in a distribution curve for your total score that is literally the same curve as rolling 3d6 for all stats, but phase shifted so that the median score is 81 rather than 63.

Basically I made an anydice link that shows the probability distributions for all the systems I just mentioned; your standard 3d6 for all 6 stats, 1d4 + 10, and 1d10 + 8, so that you can visualize what I am talking about. https://anydice.com/program/115b4

If you put that into graph view instead of table view, you will see how heavily 1d4 + 10 (actually 6d4+60 for the total score) gets weighted, leading to a very heavy likelihood that all players scores after rolling are pretty equal. Now if that is the intent, then your system works wonderfully for normalizing the power between players. If not though, then I'd suggest playing around with things further cause the idea you presented initially wasn't bad, just different, but had some hard to spot pitfalls imo.

The biggest issue in any of the random stat assignment systems isn't so much the effect on the individual player but the significant variance player to player. Even with a generous system like you describe here ... you could get one player with a total of 95 points on their best set and another with 70.

This is the main problem with the random systems ... one player can have an absolutely amazing character that is good at everything while another has a character who is much less effective. A character starting with an 18 in their primary stat compared to a 14 is at least 10% more effective in terms of hitting, they do more damage ... but perhaps most imbalancing, they only require 1 ASI to max the stat compared with 3 ASI on the other character. The character with the 18 will usually end up much more powerful.

This is why the standard array and point buy exist ... so that different players can start out on the same footing but I will agree that it can feel harder to generate unique characters this way.

I can honestly say that I would have serious anxiety generating a character in the original format - pretty much as soon as I saw it, I threw it out. I don't mind having a crap stat but having a very real chance of getting more than one is BLERRHHHGG! 

In our group we are very generous but we have found fondness for the method of having 18 and a 8 and rolling the other 4 scores in normal manner while taking lowest rolled score and rerolling it choosing higher of the two rolls. It ensures your main stat is good but you are poor at something and the random of the other 4 scores gives you variety. We do offer rerolls of whole lot if the random four scores are all very bad.

As a DM I am not overly fussed when it comes to stats. If they just choose the scores they want and have a spread between high and low scores, I'm fine with it. To the DM player stats are not that important, we can easily adjust what they face to accommodate the stats based on how difficult we want something to be. They could have 20 in every stat and I can still make difficult things for them to face and they could have 8 in every stat and I can make things easy for them to face. The group are good at not taking spotlight from others and ensuring everyone gets a chance to shine no matter the stats.

In other games where I'm player the DM has let me use a rolled set as a sort of personal array: I can use it for any future characters I make. It has two 18s and lowest score is 11. The reason is partly because stats are not important to him for the same reasons given above and also because I have shitty luck normally and spend most of my games as a player missing and failing (there are multiple sessions where I never achieved a check or attack, even with a 20 in the stat). Nothing humbles you more when fate gives a **** you to your high stats by giving you three Nat-1s in a row. >.>

One way I found that is solid... and fair...

The entire group rolls 4d6b3 7 times.. keep the best 6.... 

Then the group looks at all the numbers and picks one set and the entire group uses that set....that way they all have the same stat pool numbers... just in different spots...

you could generate scores for players in your game and make a game it part of the campaign to see which of the scores they get depending on what they choose for example

if you lead with a attack then the player would have to decide to counter it with a defence or they would get the second strat. Or if you lead with a defend they would have to use a attack or they would automaticly get the first strat. or if you lead with a magic they would have to use either their magic because its for saving throws or maybe a attack or defence depending on what the dm decides is needed for the  situation or they would get the third strat and maybe they wouldn't like it.

what I think is most important is to give some clues about each situation so they can find out what kind of option to use if there good at interpreting so I would use a skill style to give some clue about something in dnd like

use adventurer at field research to stop the goblins

so the player could figure out the clue is that a wizard hopes to accomplish stopping the goblins and his motivation is to use field research so the adventurer must be a wizard if you look in the dnd 4.0 book for background of field research theyd know and learn that the must use magic in this situation to stop the goblins.

I've been planning to start a campaign where everyone picks a race but no class, and starts with 10 in every ability score plus the racial modifiers. Then everyone gets a set of little markers corresponding to the standard array. Whenever you use an ability score for a check, save, or attack roll, you may decide to assign one of the numbers permanently to that ability. (When you assign the last of 15, 14, 13, 12, you also have to decide which of the remaining ones is 10 and which is 8.)

There'd little cards with various backgrounds at the table — at any time, you could say that one of them fits you, and grab it and get the associated skill proficiency and minor feature.

And there would also be cards listing the first-level features of various classes (For example: "Intelligence-based spellcasting from a spellbook", "Battle Rage and unarmored defense", "Charisma-based spellcasting of spells you know, and the ability to inspire your comrades"). You could grab one of those, and become that class — but only if you're in a situation where you're actually going to legitimately use that ability. On the back would be the associated list of proficiencies and the hit die size (and corresponding hd increase from the 4hp everyone would start with). For skills, you'd mark the ones which are options for the class and only make them actually proficient on first use.

Someday! :)

Quote from mattdm >>

I've been planning to start a campaign where everyone picks a race but no class, and starts with 10 in every ability score plus the racial modifiers. Then everyone gets a set of little markers corresponding to the standard array....

Someday! :)

 Sounds like what you're really looking for is HERO system =)