D&D 5E The Philosophy Behind Randomized and Standardized Ability Scores

[Yardiff]

So I thought I would try this and see what I got:
4 tries until first 2x 15 and
11,17,15,13,12,12
3 tries:
10,15,15,9,16,9
2 tries:
17,15,13,11,13,14
3 tries:
16,11,14,9,14,18

Interesting, Did Colville say, place in order or rearrange to suit?

Arrange as wanted.

 

[Yardiff]

Nice try, but math don’t cheat, if you don’t put a minimum you will eventually roll the two 15+ and a bunch of 3. It may not happen at your table but a world wide used method will get it.

Unclear what this reply is about. I was just letting tetrasodium know that in the video he was talking about from Matt Colville, Matt used 4d6 drop lowest not the 3d6.

 

[Krachek]

Unclear what this reply is about. I was just letting tetrasodium know that in the video he was talking about from Matt Colville, Matt used 4d6 drop lowest not the 3d6.

Even if you roll 4d6 drop lowest you can still achieve a 3. Rarer but still possible.
So unless you put an explicit rule for lowest score, you can still roll 15,15,3,3,3,3!
And should also buy a lottery ticket!

 

[Jack Daniel]

3d6 in order is where it's at for me, but (1) I prefer to use a variation on the Swords & Wizardry White Box table, where modifiers hardly ever exceed ±1, so the difference between (say) a Strength 3 warrior and a Strength 18 warrior who are fighting shoulder-to-shoulder is two points of hit and damage at most; and (2) I run old-school, which presupposes that as the campaign goes on, each player will eventually have a stable of differently-leveled characters to choose from, suitable for tackling a wide variety of differently-leveled adventures.

The idea that if you roll poorly, you get "stuck" with your poor stats for a whole, entire campaign is a concept thoroughly wedded to traditional, single-character, plotline-driven, and OC-centric modern play-styles.

 

[Yardiff]

Even if you roll 4d6 drop lowest you can still achieve a 3. Rarer but still possible.
So unless you put an explicit rule for lowest score, you can still roll 15,15,3,3,3,3!
And should also buy a lottery ticket!

I believe Colville had a minimum of 8. Not sure if you reroll lower then an 8 or made it an eight. For myself if someone rolled 2 or more 3's I'd have them start over.

 

[Maxperson]

I mention this because I think that this is often given short shrift in discussions of ability score generation- not the effect on a singular PC, but on the effect on the table. If someone gets really really low scores, and someone gets really really high scores, what will happen? Is this going to be seen as good fun- an opportunity for memorable characters? Or is this going to be viewed by the players as unfun and unfair? The type of players and the reaction you get will most likely determine the best methods of character generation.

I would go so far as to say that this sense of fairness is what has likely driven the adoption of standardized ability scores over time (point buy, array). Even some randomized methods allow for this sense of fairness (such as allowing players to bid on scores at an auction, or to chose among a number of random scores).

Anyway, I wanted to put this out for general discussion- the social and table aspect of ability score generation. What do you think? Do you consider this, and if so (or if not), why?

I think it comes down to table personalities. My group doesn't really care if they roll higher or lower than others at the table, so it's really not a concern of ours. However, I know a few players and I've seen many here who do view the kind of equity arrays give as very important.

The method of stat generation that I use is assign two stats to roll 5d6-2L, two stats to roll 4d6-L, and two stats to roll 3d6. They kinda sorta get to place stats for high intelligence as a wizard for example, by assigning a 5d6-2L to it, but as you note above with the old UA method, you can still roll poorly that way or highly on a 3d6. I do want them to be able to play the class they want, though, so I allow one swap of a pair of numbers. I also have a floor of 72 points, so if the total stats don't add up to 72, they roll a d6 and assign a +1 to that stat(usual stat order) and continue to do so until they are up to 72. Max of 18 of course.

 

[Maxperson]

I believe Colville had a minimum of 8. Not sure if you reroll lower then an 8 or made it an eight. For myself if someone rolled 2 or more 3's I'd have them start over.

He has them re-roll if lower than 8. I watched part of the first episode of his game and saw them make characters.

 

[Lanefan]

I like how The Black Hack does it: you roll 1d20 and if the result is ABOVE your current stat, it goes up by 1 point. That way it gets harder the higher your stat already is.

Interesting idea. How often is such a roll given? (not familiar with The Black Hack at all)

 

[Krachek]

Score improvement for those who like to roll but don’t like risk.

Use standard array or use point buy.

Then choose one option:

If you have a 15, you can roll six times 4d6 drop lowest and try to improve the 15. If you fail you keep the 15.

you can pick a racial feat for free.

You improve your two lowest score by 2.

You can turn a 13 into a 14

Using the standard array you can finish with

15, 14, 13, 12, 10, 8 and a racial feat
18, 14, 13, 12, 10, 8, high roll
17, 14, 13, 12, 10, 8
16, 14, 13, 12, 10, 8
15, 14, 13, 12, 10, 8 no luck
15, 14, 13, 12, 12, 10, safe play
15, 14, 14, 12, 10, 8, safe play

 

[Oofta]

Like others, the issue I have with a lot of options presented is that they're almost all "safe". You can't really end up with a "bad" PC. You can end up with something close to what point buy gets you or a PC that has sky-high scores. If sky-high scores are what you're after why not just increase the point buy budget and allow purchasing ability scores up to 18? Use D&D 3.5's point buy method with 32 points (high powered) will give everyone an exceptional PC. Random? Use the list I had back on the first page.

For me, I like not starting out with super high scores. It give me a sense of growth and accomplishment, a feeling that PC is improving in multiple ways. We did try higher scores for our first 5E campaign that my wife ran we used the aforementioned "high powered" option but ended up kind of regretting it. I like playing off the fact that my paladin while strong and has a charisma a bit above average but kind of a klutz.

Of course YMMV and all that.

 

[Helpful NPC Thom]

3d6 down the line, no rerolls, human fighter only, final destination.

 

[reelo]

Interesting idea. How often is such a roll given? (not familiar with The Black Hack at all)

You get ONE try for each stat at each levelup (which I personally find excessive) If it's your class' main stat, you can try twice.

 

[Chaosmancer]

I like how The Black Hack does it: you roll 1d20 and if the result is ABOVE your current stat, it goes up by 1 point. That way it gets harder the higher your stat already is.

I'd like to hear more about this, same as Lanefan

Edit: Guess I should have refreshed the page first.

But, actually, that makes sense to me considering the risk reward and that you only get to increase them by 1 pt

 

[Lanefan]

You get ONE try for each stat at each levelup (which I personally find excessive) If it's your class' main stat, you can try twice.

Which means someone starting with a standard-ish run of stats - say, 15-14-13-12-10-8 - has a pretty good chance of seeing two of those go up at level 2 and (maybe not the same) two go up at level 3; after which the odds very much favour one stat going up each level for quite a while.

I agree, that's excessive - unless you're playing a level-capped variant similar to the E6 version of 3e, in which case I could live with it.

 

[tetrasodium]

You get ONE try for each stat at each levelup (which I personally find excessive) If it's your class' main stat, you can try twice.

What do they begin at after character creation/before the first roll? Do they roll at level 1 or not till 2?

 

[Chaosmancer]

Which means someone starting with a standard-ish run of stats - say, 15-14-13-12-10-8 - has a pretty good chance of seeing two of those go up at level 2 and (maybe not the same) two go up at level 3; after which the odds very much favour one stat going up each level for quite a while.

I agree, that's excessive - unless you're playing a level-capped variant similar to the E6 version of 3e, in which case I could live with it.

True, but if you could only choose one per level, you'd go with the fifteen. And then you'd have a 75% chance of nothing happening, meaning it could be multiple levels until you saw any growth. So, they have to have it be for every stat just on that, and if you assume an average game of ten levels, you probably need every level to get that fifteen to go up two or three times.

 

[reelo]

What do they begin at after character creation/before the first roll? Do they roll at level 1 or not till 2?

In "TBH" at character creation, you roll 3d6 in order. If you roll 14+ for one stat, the next stat will automatically be a 7. Once you have 6 stats, you can swap 2 stats to your liking.
First time you roll for ASI is when you reach lvl 2.

 

[Lanefan]

True, but if you could only choose one per level, you'd go with the fifteen. And then you'd have a 75% chance of nothing happening, meaning it could be multiple levels until you saw any growth. So, they have to have it be for every stat just on that, and if you assume an average game of ten levels, you probably need every level to get that fifteen to go up two or three times.

True.

I suppose the bigger question* is, on average how many total stat points, if any, should a character expect to gain** over 5 levels, or 10, or 20? Then once that one's answered the next is should those gains be at predictable points (e.g. every x levels), or random, or a combination of both? After that we can ask should the gains be in stats of the player's choice, or forced by class (i.e. gains must be in your prime stat), or randomized, or some combination?

Personally, I'm not sold on the need for much stat gain at all in 5e, largely because as far as I can tell there's no means of permanently losing stat points in this edition thus any cancel-out-the-losses rationale evaporates. I could see over 20 levels a gain of maybe 2-5 points in your prime stat plus the same range again in total across randomized other stats, with the gains in all cases coming unpredictably. One way of doing this might be to take that roll-over-the-stat method propsed upthread and modify it such that if you miss at a level your roll at next level is at +5, cumulative each level until you hit; but once you hit you don't get to roll for that stat on the level following such that a stat cannot advance on two level-ups in a row. (so if the stat is 15 and you roll 13 (miss) then next level you roll 6 [+5] (miss) then next level you roll 12 [+10] (hit, stat advances) then next level there's no roll, after which you start the process again by rolling straight-up against 16)

* - and one can argue this question can be applied in some form to any edition, not just 5e; and the answer would probably be different for each.
** - via pure game mechanics, ignoring wish effects or other in-game means of permanently boosting a stat.

 

[Chaosmancer]

True.

I suppose the bigger question* is, on average how many total stat points, if any, should a character expect to gain** over 5 levels, or 10, or 20? Then once that one's answered the next is should those gains be at predictable points (e.g. every x levels), or random, or a combination of both? After that we can ask should the gains be in stats of the player's choice, or forced by class (i.e. gains must be in your prime stat), or randomized, or some combination?

Personally, I'm not sold on the need for much stat gain at all in 5e, largely because as far as I can tell there's no means of permanently losing stat points in this edition thus any cancel-out-the-losses rationale evaporates. I could see over 20 levels a gain of maybe 2-5 points in your prime stat plus the same range again in total across randomized other stats, with the gains in all cases coming unpredictably. One way of doing this might be to take that roll-over-the-stat method propsed upthread and modify it such that if you miss at a level your roll at next level is at +5, cumulative each level until you hit; but once you hit you don't get to roll for that stat on the level following such that a stat cannot advance on two level-ups in a row. (so if the stat is 15 and you roll 13 (miss) then next level you roll 6 [+5] (miss) then next level you roll 12 [+10] (hit, stat advances) then next level there's no roll, after which you start the process again by rolling straight-up against 16)

* - and one can argue this question can be applied in some form to any edition, not just 5e; and the answer would probably be different for each.
** - via pure game mechanics, ignoring wish effects or other in-game means of permanently boosting a stat.

One aspect I'm thinking about is linear growth and training. With the way that the system is designed, with 5e, you end up in a quite odd situation.

Let's say that you ended up with the Standard array and applied racial bonuses. So 16, 16, 13, 12, 10, 8. Now, at level two you roll 6d20 and what are the likely results? It is highly likely that the two 16's remain the same, if the 12, 10 or 8 get a +1 it doesn't change anything mechanically, and so the only potential increase is the 13 to a 14.

Now, next level you might see a bit of a cascade, you might see those bottom three numbers go up all at once. But the core of your character's power is still those two 16's, and they are still unlikely to have changed. It may take you until level 5 or 6 to get those to raise up to a 17. Only hitting an 18 around level 10.

So, what seems likely is that your lower stats, which usually are ignored in DnD, start mellowing out into that +1, +2 territory as you start reaching the end of the typical campaign, while you high stats remain largely unchanged (barring luck). This gives a narrative of training to overcome your weaknesses, and combatting the issues with MAD classes who generally struggle with needing a lot of stats. It would also actually slow down the changes to the ceiling of power, because getting to an 18 wouldn't be nearly guaranteed by level 4. Meanwhile, I don't think it would feel... bad, to not be reaching those heights, because your breadth is increasing.


The issue I can see is whether or not you are okay with the lower stats getting up to the +2 range pretty regularly in every campaign. But for me, that seems perfectly reasonable.