D&D 5E How do you handle randomly rolling for stats

[CapnZapp]

@ezo, two comments:

I would like more information than just the expected value to make an informed decision. What does the bell curve (with the expected score "on top") look like for each of the numbers? (How likely is it that, say, that second number is a 15 instead of your 14? How big of a risk it ends up a 13 instead?)

Do you weigh the numbers of each array at all? I would say (15, 13, 12, 10, 8, 6) is at least as attractive as (14, 13, 12, 11, 9, 7) - your score in your most used ability is much more important than your three lowest scores. Getting one point higher in your "prime stat" is arguably worth getting even at the cost of having all three of your lowest scores lowered one point each.

 

[CapnZapp]

That's your opinion.

It is an universal aspect of risk assessment and valuation.

 

[ezo]

@ezo, two comments:

I would like more information than just the expected value to make an informed decision. What does the bell curve (with the expected score "on top") look like for each of the numbers? (How likely is it that, say, that second number is a 15 instead of your 14? How big of a risk it ends up a 13 instead?)

Do you weigh the numbers of each array at all? I would say (15, 13, 12, 10, 8, 6) is at least as attractive as (14, 13, 12, 11, 9, 7) - your score in your most used ability is much more important than your three lowest scores. Getting one point higher in your "prime stat" is arguably worth getting even at the cost of having all three of your lowest scores lowered one point each.

Frankly, that gets into more infromation and discussion than I am really interested in getting into.

But, I'll give you the quick-notes version of what I did:

It isn't practical to look at all the possible variants of rolling for all six scores, so I went with 100000 samples of generated arrays.

For each array, I calculated the total number of the scores as well as the total modifiers, sorted from using scores primarily and secondary by modifiers.

For any array which generated totals below the standard array (and modifiers below the standard array), I replaced those 41000+ with the standard array and recaluclated the expected values.

Of course, if a particular array is "bad" is subjective. In comparing our two "inferior arrays", your 15 is higher than my 14, and I'll require the full ASI, as you wouldn't (could go with a half feat, etc.), but both your negatives require two increases to reduce the penalty, as where one ASI could be split to reduce both negatives for mine...

It really just depends on how you weigh the positives vs. the negatives. I don't put as much emphasis on the "prime stat", and there is little real benefit in having a +3 vs. a +2. Casters probably would value it more than martials.

Anyway, I really don't like a -2 ability modifier, so even if I went with my array, my first ASI would be to raise my 9 to 10 and 7 to 8. Having more odd scores allow for faster improvement in overall ability modifiers.

FInally, I just noticed while writing this post if I had my inferior array of 14, 13, 12, 11, 9, 7 and choose regular human, it would be 15, 14, 12, 12, 10, 8... or the normal 5E standard array! That just strikes me as funny. All the math I did, and I could have just said, "Meh, just lower the standard array by 1 for each score" and gotten the same result! Sigh...

 

[Maxperson]

I've never had players choose the array. If I am using the RAW methods, Usually it is they choose to roll, occasionally they choose point buy.

However the most common way I do it is by rolling using an adaptation of an old official method from late 1E.

In this method players pick what class they start with at first level then they take their 6 abilities and rank them from most important to least important - so a Barbarian might pick something like this: 1. Strength, 2. Constitution, 3. Dexterity, 4. Wisdom, 5. Intelligence, 6. Charisma.

After they rank their abilities they roll them in order.
Ability #1 (most important): Roll 9d6 and save 3 highest
Ability #2: Roll 8d6 and save 3 highest
Ability #3: Roll 7d6 and save 3 highest
Ability #4: Roll 6d6 and save 3 highest
Ability #5: Roll 5d6 and save 3 highest
Ability #6 (least important): Roll 4d6 and save 3 highest.

After this you choose race and apply bonuses (I allow the Tasha's floating bonuses).

This generates very high scores overall on average, but you can't move the scores. The average score is a 16 in the primary ability before racial bonuses and it is not uncommon to have a 13+ in every ability but you can't move them, so if your Barbarian has a higher intelligence than strength, then she has a higher intelligence than strength. Rolling 9 dice for the most important ability virtually guarantees at least a 14 after the +2 for race and a 19 or 20 is relatively common.

This adds a lot of flavor to characters because it eliminates "dump stats" and the luck can give very high scores in "unimportant" abilities, while virtually assuring you have good scores in the important abilities.

For example we have a V Human player who rolled a melee fighter and after racial mods had a 16 Strength, 14 Con, 18 Intelligence and 18 Wisdom at 1st level. The Bugbear Barbarian had a 19 Strength, 18 Dex, 16 Constitution, 17 Wisdom at 1st level.

This is by far my favorite method of generating stats when I am a DM. I always play in small groups (3-4 PCs).

Wow! Through all of 1e I could never find a DM who allowed that method, including me! It was too broken. Do your players have to pick class first, or can they just rank the stats and pick a class after rolling?

 

[Rabbitbait]

I've never had players choose the array. If I am using the RAW methods, Usually it is they choose to roll, occasionally they choose point buy.

However the most common way I do it is by rolling using an adaptation of an old official method from late 1E.

In this method players pick what class they start with at first level then they take their 6 abilities and rank them from most important to least important - so a Barbarian might pick something like this: 1. Strength, 2. Constitution, 3. Dexterity, 4. Wisdom, 5. Intelligence, 6. Charisma.

After they rank their abilities they roll them in order.
Ability #1 (most important): Roll 9d6 and save 3 highest
Ability #2: Roll 8d6 and save 3 highest
Ability #3: Roll 7d6 and save 3 highest
Ability #4: Roll 6d6 and save 3 highest
Ability #5: Roll 5d6 and save 3 highest
Ability #6 (least important): Roll 4d6 and save 3 highest.

After this you choose race and apply bonuses (I allow the Tasha's floating bonuses).

This generates very high scores overall on average, but you can't move the scores. The average score is a 16 in the primary ability before racial bonuses and it is not uncommon to have a 13+ in every ability but you can't move them, so if your Barbarian has a higher intelligence than strength, then she has a higher intelligence than strength. Rolling 9 dice for the most important ability virtually guarantees at least a 14 after the +2 for race and a 19 or 20 is relatively common.

This adds a lot of flavor to characters because it eliminates "dump stats" and the luck can give very high scores in "unimportant" abilities, while virtually assuring you have good scores in the important abilities.

For example we have a V Human player who rolled a melee fighter and after racial mods had a 16 Strength, 14 Con, 18 Intelligence and 18 Wisdom at 1st level. The Bugbear Barbarian had a 19 Strength, 18 Dex, 16 Constitution, 17 Wisdom at 1st level.

This is by far my favorite method of generating stats when I am a DM. I always play in small groups (3-4 PCs).

Wow. Not a method I ever would have allowed unless I was really wanting a super-powered campaign. I don't remember that at all from 1e. Was it in a later supplement?

 

[Maxperson]

Wow. Not a method I ever would have allowed unless I was really wanting a super-powered campaign. I don't remember that at all from 1e. Was it in a later supplement?

It was in the 1e Unearthed Arcana.

 

[Zardnaar]

Generally several ways. Players roll in front of DM.

Small the group more liberal with revolts. Generally each 0layer rolls once and players can pick each others stats do no one's left behind.

DM warning if rolls are high you will get a lot less magical loot or lower owered.

Eg one player rolls an 18 everyone picks it. If they follow it up with the usual rates, feats etc to hit a 20 don't expect much beyond +1 items and scrolls.

 

[Yaarel]

After exploring many mathematically precise arrays:

Standard Array.


But for a future campaign, I am considering: 13, 12, 12, 11, 11, 10

No negatives to worry about, and start off "typical". Friendlier for Bounded Accuracy design too.


As long as everyone in the party has the same array, it doesnt really matter which one it is.

 

[Rabbitbait]

It was in the 1e Unearthed Arcana.

Ah. That was a good book, but there was some stuff from there that was wonky. That said, 1st edition was wonky throughout. That was part of the fun of it.

 

[dave2008]

It is an universal aspect of risk assessment and valuation.

maybe, but sometimes the reward for being risky is just the fun of being risky!