AD&D Demographic Data Part II

Serves me right for wanting to get something out there before I really let it settle. While watching the World Cup matches, I figured out a reasonably quick and easy way to properly adjust the percentage chances for each of the ID# categories in the last post. So, here is the adjusted data (after the cut). All that needs to be done now is to divide each ID#'s percentage among the classes included in that group. This could be done in various ways, which is why I am leaving it as the raw data. One could simply divide the percentage evenly among the classes, or one could weight the classes and divide the percentage according to that weight. One could even make a priority list and give the full percentage to whichever class comes highest in priority. The details will vary by setting.

Note that when that division of percentages is finished, there should be a bit left over. That covers the "hopeless characters", who don't qualify for any class.

Without further ado, here's the corrected data, now just listed as ID#, classes, raw stat minimums of the category, and then the odds of that category occurring (multiply by 100 to get actual % chance):

AD&D Demographic Data

I like simulation. I like it when the game rules imply things about the setting, or vice versa. One of the D&D assumptions is that random people are "rolled up" by the 3d6 in order method. PCs might be different, but that doesn't mean that the setting as a whole is.

When working on figuring out what the rules imply about the setting, one thing I noticed is that there isn't a good, reliable accounting of how many people in a given setting are members of which character classes. Gary Gygax seems to have just made something up for his table in the DMG, and no one has bothered to improve much on that intuitive method.

But we know some things. According to the DMG, 1 person in 100 has a chance of having a character class. Of those 1%, though, there will be different numbers of randomly-rolled characters who qualify for each of the various classes.

Now, sure, we can just make something up and deal with that. That works, and is fine. But if we had a better way available… why not use it?

All of the numbers involved have a mathematical probability. It just takes a lot of effort to work out if done manually (trust me, that's how I started to approach this). A computer program, though, can work through all of the possibilities in just minutes or seconds. So I wrote one. Here's the raw data regarding the basic AD&D 1E classes (excepting the Bard, which is more a matter of choice; suffice it to say that Bards are extraordinarily rare in AD&D 1E). This set of data does not include the UA classes at all, though if someone requests it I could probably put that together. It also doesn't include anything about demihumans. I haven't consolidated that information yet, and there are some characteristic maximums to deal with that I would have to write into the program, not to mention the characteristic modifiers that change probabilities around.

The data is in the following format: an ID#, the name of the classes that qualify under that ID#'s set of characteristics, and the characteristics themselves, then on the next line it gives the probability of a random roll of 3d6 in order qualifying for that set of statistics, and finally on the last lines a list of the ID#s that qualify for the ID#, but include one or more stats that are higher (all of the probabilities of those ID#s are included in the probability listed, so they have to be subtracted out). My purpose for all of this is to work out what classes a given set of stats can qualify for, subtracting the probabilities of the higher sets from that of the lower sets that they also qualify for. Then, the probability of each set of stats will be divided among the set of classes. I'm also thinking about ways to modify the software to do that work for me, but I don't know if I'll be able to do that with my level of programming skill.

I don't plan to do this myself for the raw AD&D classes (unless I can work out how best to code it), so I'm giving that data out for someone who might be interested in doing so. There are only 76 sets that qualify for different groups of classes in AD&D, so it probably wouldn't take all that long to work through them. The data I'm working with includes over 500 distinct ID#s, divided into three main groups to cover the three main culture groups so far in the Middle Sea world, so I have my work cut out for me. It's nowhere near as imposing as it was before I wrote this program, though.

Data after the cut. Remember that it is the raw data as described above. I am also willing to take any set of classes and stat requirements and do the same for others, since the software is easily modified for just about anything. It won't easily handle very complex things, though, like requirements that one of several stats be above a value, but the other stats of that set only need to be of a lower value (like the Psionicist, which requires a 10 in Int, Wis, and Cha, but at least one must be 16, which particular case I did work out for my own use, but requires more modification of the basic program than just plugging in a set of classes). Drop me a line if you want this kind of data.

Quick Random NPCs For D&D-like Games

Found this by searching for "random fantasy
characters".
By dawnsio-ar-y-dibyn, licensed under a Creative Commons
Attribution-Noncommercial-No Derivative Works 3.0 License

Sometimes, you know what sort of NPC you want, what class they should be and so forth, but you don't want to either pick the stats or roll until you finally get a set of stats that is high enough to join the class. There's a method in the DMG, but it isn't really close in terms of normal spread of stat numbers. For those situations, I have come up with the following sets of tables. The assumption is that only characters with appropriate stats will join the class (obviously), and that you don't want to keep rolling sets of stats until you finally get an appropriate set. Most of the charts use d1000, which is rolling three distinctive 10-sided dice and using one for the hundreds place, one for the tens place, and one for the ones place, just like rolling a d%/d100. The higher charts could be replaced by a d20, a d10, or even a d4 roll, but I stuck with d% as the smallest die I'd use here. You could probably round the d1000 tables to d%, but I'm not interested in doing that.

Note that higher leveled characters should probably have slightly higher stats, due to the increased likelihood of survival. Delta discusses the matter here, which gives the following adjustments (do with the information as you will):

Level	STR	CON	DEX
2	+3	+2	+1
3-4	+4	+2	+1
5-7	+4	+3	+2
8+	+4	+4	+3

The charts were pretty easy to make, actually, just requiring some time with a calculator and some tables of the properties of 3d6 and 4d6s3 ("4d6, keep the best three") rolls. I just reduced the possibilities to those that were required, and divided the number of results that can result in each number by that value. The fraction, rounded off to the nearest thousandth, provided the width of each result. In a few cases, the rounding resulted in a variation of 1 or 2 points, which I worked into the table as the feeling took me.

The following charts are organized by the required stat (obviously, if the required stat is 3, you would just roll the stat as normal), and by the method used for the character, either 3d6 as normal, or 4d6 discarding the low die (so that the character can be either normally-distributed or on par with the PCs if you use that method). I'll put in a jump cut for your sanity.