Another point on the nature of stats is what is meant by "average."

A 10 might be average human strength, but that doesn't mean any warrior has a strength that low. The average includes women, elderly and the infirm.

If you rolled for every adult human NPC about 1-in-4 would have a Strength of 13+. About 1-in-10 would have a 15+ and about 1-in-50 would have a 17+.

The typical ratio of full-time warriors in the population of medieval societies is about 1-in-100. The maximum you could conscript between planting and harvest without wrecking your economy was about 1-in-10 (and 1-in-20 is safer).

So that means with the standard attribute distributions for humans on 3d6 in order, there are enough humans with a Strength of 17+ to completely fill the ranks of the professional warrior class twice over and enough with Strength 15+ to completely fill the ranks of conscripted forces. And that's on 3d6 in order.

Realistically, every human fighter should have a Strength of 17+ if they're any type of professional warrior.

How common are wizards in your setting? If it's fewer than 1-in-1000 there's no reason for any wizard to ever take on an apprentice with less than an 18 Intelligence as they'd have 4+ candidates from the general 3d6-in-order human population to choose from... why settle for someone with only a 17?

And on down the line it goes; all the classes are rare enough in society that instructors and mentors could afford to be extremely picky in their choice of students. Even if the bulk of human society used 3d6 in order for their stats, the only way that makes sense for PCs relative to average humans is if you're literally playing average humans (i.e. peasants or serfs) or if you're letting the player pick "best of 100 sets" for their scores.

Quote from: Chris24601 on June 22, 2021, 03:47:29 PM
Another point on the nature of stats is what is meant by "average."

A 10 might be average human strength, but that doesn't mean any warrior has a strength that low. The average includes women, elderly and the infirm.

If you rolled for every adult human NPC about 1-in-4 would have a Strength of 13+. About 1-in-10 would have a 15+ and about 1-in-50 would have a 17+.

The typical ratio of full-time warriors in the population of medieval societies is about 1-in-100. The maximum you could conscript between planting and harvest without wrecking your economy was about 1-in-10 (and 1-in-20 is safer).

So that means with the standard attribute distributions for humans on 3d6 in order, there are enough humans with a Strength of 17+ to completely fill the ranks of the professional warrior class twice over and enough with Strength 15+ to completely fill the ranks of conscripted forces. And that's on 3d6 in order.

Realistically, every human fighter should have a Strength of 17+ if they're any type of professional warrior.

That assumes half the strongest people in the planet become warriors, with only the remainder spread across things like farmers tilling the fields, millers grinding flour, blacksmiths beating on iron, and on and on. Which seems extraordinarily unlikely, because there are a lot of pre-modern professions where great physical strength matters. Not to mention it ignores class -- the social kind, not the D&D kind -- a lot of people become warriors or are prohibited from becoming warriors because of their birth. And that's assuming 3d6 represents the overall population distribution not, say, the distribution among adventurers.

A more realistic approach is to look at the standard array. For 3d6 in order, it's roughly 13, 12, 11, 10, 9, 8. The optimal choice for a fighter is to put that 13 into Strength. Won't always happen of course, but it's reasonable to assume it will occur more often than other placements. So while it's not a precise tool, using the optimal allocation of the standard array can work as an upper limit of a baseline for NPCs in a particular profession.

Quote from: Tristan on June 21, 2021, 11:24:33 PM
I know later editions have moved to arrays, and I can see the appeal in it honestly. Just old habits are hard to break.

I agree.  Most people prefer rolling the dice, even if it turns out badly.  It's more exciting than arrays.

Quote from: DocJones on June 22, 2021, 06:23:01 PM

Quote from: Tristan on June 21, 2021, 11:24:33 PM
I know later editions have moved to arrays, and I can see the appeal in it honestly. Just old habits are hard to break.

I agree.  Most people prefer rolling the dice, even if it turns out badly.  It's more exciting than arrays.

See, and this just goes to show the value of anecdotal evidence, since in my circles the majority of players prefer to not play dice with something as important and lasting as ability scores in a campaign of any length. Save the dice rolling for the actual game, not chargen.

In RPGs there are a lot of regional preferences that get presumed as applying across the board when they're not.

Quote from: DocJones on June 22, 2021, 06:23:01 PM

Quote from: Tristan on June 21, 2021, 11:24:33 PM
I know later editions have moved to arrays, and I can see the appeal in it honestly. Just old habits are hard to break.

I agree.  Most people prefer rolling the dice, even if it turns out badly.  It's more exciting than arrays.

Greetings!

Yep, I agree, DocJones! Every gamer I have played with in my groups have loved rolling dice for abilities. They so often love the excitement and unpredictability of seeing how the dice come up for everyone, not just their own characters.

Semper Fidelis,

SHARK

Quote from: Chris24601 on June 22, 2021, 03:47:29 PMRealistically, every human fighter should have a Strength of 17+ if they're any type of professional warrior.

The good ones were conscripted. The PCs are the ones who didn't get conscripted, and thus are free to adventure.

3d6 down the line. Anything else and you may as well go join the People's Republic of Lake Woebegone.

I really liked and prefer the wider distribution of bonuses in B/X compared to 1st and 2nd edition AD&D. So much that I use it instead for my homebrew houseruled 2nd edition.

Quote from: Ratman_tf on June 23, 2021, 04:08:31 AM
I really liked and prefer the wider distribution of bonuses in B/X compared to 1st and 2nd edition AD&D. So much that I use it instead for my homebrew houseruled 2nd edition.

Interesting. Not being a B/X player, I hadn't realized how B/X was a precursor to 3rd edition in that respect, by standardizing and smoothing out the attribute bonuses. Just for anyone else who didn't realize, I looked this up and the bonuses standardized to this for most attributes:

3: -3
4-5: -2
6-8: -1
9-12: +0
13-15: +1
16-17: +2
18: +3

I think it's a subtle enough change that I wouldn't necessarily want to port it to 5E, but I like the design. In general, I have liked attributes being less of a focus in play. In D&D, I preferred the mechanics side to focus most on class and skills and special abilities -- and less on attributes and magic items.

My experience of AD&D1 was that getting good attribute rolls was a big deal. Later editions moderated this by removing the experience bonus and attribute minimums. They also smoothed out the bonuses which benefit of lucky 18s, but overall size of modifiers was just as important.

One of the advantages of not trying to be backwards/sideways compatible is that you can do things like this.  It's what I'm using with the 3d6 in order, swap any 2:

Score	Modifier
1	-4
2-3	-3
4-5	-2
6-7	-1
8-10	+0
11-13	+1
14-16	+2
17-19	+3
20-23	+4
24-28	+5

There is a small chance that a starting character can get a level 1 bump that will turn a naturally rolled 16-18 into the 20-23 territory and get a +4 (tiny for 16 to 20, roughly 3%).  But otherwise starting characters are kept in the -3 to +3 range, with minuses less common than plusses simply due to the skewed distribution on the 3d6.  It does have the side effect of making the players a heck of a lot more open to rolling 3d6 in order.  However, the main reasons that I did it were:

A. I wanted the full range of possibilities defined in the mechanics from the beginning even though I knew I'd be keeping most players in that -3 to +3 range, and even the outliers in the -4 to +4 range.  That range is one thing that I do want to be compatible with earlier D&D versions.  You'll note that the standard 13, 16, 18 scores have the compatible modifiers, too.

B. I wanted an easy, intuitive way to show that it's a lot easier to get rid of minuses with a bump than improve a positive.  Since most bumps require the character to roll over their current stat on a d20 to get the full possible benefit, it's even more skewed than the chart would indicate.  (Fail that roll, increase score by 1 point.  Succeed, get 2-4 points).

Yeah, I know.  You can do all of that and keep the traditional chart with some clever math while most preserving compatibility with D&D.  Once compatibility became a low priority, however, this lets me ditch the clever math.  Point being for this topic that it was actually easier for me to keep the old original style of rolling (mostly) and change the results of the roll than it was to change the rolling method.

Quote from: jhkim on June 23, 2021, 01:46:57 PM
Interesting. Not being a B/X player, I hadn't realized how B/X was a precursor to 3rd edition in that respect, by standardizing and smoothing out the attribute bonuses. Just for anyone else who didn't realize, I looked this up and the bonuses standardized to this for most attributes:

3: -3
4-5: -2
6-8: -1
9-12: +0
13-15: +1
16-17: +2
18: +3

It approximates the first 3 standard deviations. This would be closer:

3-4: -2
5-7: -1
8-13: 0
14-16: +1
17-18: +2

But doesn't get to +/-3, so it's fudged a point.

Which may be why the range feels fairly natural to a lot of people. Humans seem to have an intuitive grasp of the normal distribution.

Quote from: Shasarak on June 21, 2021, 07:09:58 PM

Quote from: Omega on June 21, 2021, 05:54:44 PM
Its been all over the place with D&D. Early game had little limits on how far you could take stats. IF you could get the things needed. The usual scarcity if those boosters was the limiter.

I miss the ADnD stats and being able to unlock things like regeneration if you could get over an 18.

Oh and hilariously enough. A hidden quirk of SSI's gold box games was that if you played the series through, had a dwarf with max con 19 without editing, and then got a manual that boosted CON, they gained regeneration since the Gold Box series used as much of the D&D rules as could code in.

Characters can even die of old age. I found this out the hard way since we had to sometimes rely on camping extended periods to heal.

Quote from: Omega on June 23, 2021, 05:01:55 PM

Quote from: Shasarak on June 21, 2021, 07:09:58 PM

Quote from: Omega on June 21, 2021, 05:54:44 PM
Its been all over the place with D&D. Early game had little limits on how far you could take stats. IF you could get the things needed. The usual scarcity if those boosters was the limiter.

I miss the ADnD stats and being able to unlock things like regeneration if you could get over an 18.

Oh and hilariously enough. A hidden quirk of SSI's gold box games was that if you played the series through, had a dwarf with max con 19 without editing, and then got a manual that boosted CON, they gained regeneration since the Gold Box series used as much of the D&D rules as could code in.

Characters can even die of old age. I found this out the hard way since we had to sometimes rely on camping extended periods to heal.

The Con boost was pretty well hidden. To get it, you have to read the manual and then rest for 30 days straight. That's completely abnormal behavior, and nothing in the game tells you that's how it works. So unless you remembered that particular detail from the DMG, you'll never figure out how to use it. And it didn't work across the entire gold box series. The manual is in the Pool of Radiance, but attempting to import a 20 Con dwarf to Curse of the Azure Bonds results in Con dropping down to 19.

But yes, it's remarkable how faithful the gold box series was.

Speaking of.

What is the stat array spread for 3d6?

Quote from: Pat on June 23, 2021, 05:47:42 PM
The Con boost was pretty well hidden. To get it, you have to read the manual and then rest for 30 days straight. That's completely abnormal behavior, and nothing in the game tells you that's how it works. So unless you remembered that particular detail from the DMG, you'll never figure out how to use it. And it didn't work across the entire gold box series. The manual is in the Pool of Radiance, but attempting to import a 20 Con dwarf to Curse of the Azure Bonds results in Con dropping down to 19.

But yes, it's remarkable how faithful the gold box series was.

Your CON dropped back to 19, but for whatever reason the regen persisted into Azure Bonds. Not sure if it persisted into Silver Blades and on.

I hit on it by a combination of accident and being a DM. It was like, huh, will this really work on this dwarf with 19 CON? wow? It did. Though took me a while to actually realize it was working.

Another thing the SSI games showed aplenty was just how ruthlessly good Fighters could be over Magic Users due to spell interruption and MU's being a bit fragile usually. That and how easy it was to mis-guess where that darn Fireball or Cloudkill was going to land. ow... lots of ow. On the other hand Stinking Cloud and Cloudkill became vital for defending the frontline fighters by making a deadly barricade.

Interestingly the SSI  games were r4h3. But in order. You did not get to assign those rolls.

Quote from: Omega on June 23, 2021, 06:01:40 PM
Speaking of.

What is the stat array spread for 3d6?

There are a couple ways to generate the normal array for 3d6. The obvious method is to generate all possible results of 3d6 in order, sort each array from highest to lowest, then find the most common, or the median, or the mean array. The problem is these are fairly computationally intensive (6^18 is hundreds of trillions arrays), so doing a Monte Carlo simulation based on the mean is a quick easy alternative. The average of 100,000 randomly generated 3d6 in order characters, with the numbers in each array sorted from highest to lowest:

14.2, 12.4, 11.2, 9.9, 8.6, 6.8

Converted to integers, that works out to:

14, 12, 11, 10, 9, 7

Different methods of finding the normal array will generate minor variations. 13, 12, 11, 10, 9, 8 is popular because it's easy to remember, though it doesn't come up on its own. Looking for the mode, arrays that start with 13, 13 are common as well, though I don't remember the rest of the array. Picking any of them is probably fine.