Some Statistical Analysis, or why is this goat herder so tough?

[Evilschemer]

I'm writing up NPC stats for my Swords of Cydoria game and I wanted to be balanced as well as well integrated into the other BRP games.

So I was looking at other NPCs and thinking "Hey, none of these NPCs have stats less than 10. Everyone's such a bad-ass compared to PCs!"

I had some time yesterday afternoon, and was curious, so I did some statistical analysis (the kind of thing I do for my job) on the characteristics of various NPCs from several sourcebooks.

For the purposes of this analysis, I am looking at human NPCs from the following books:

• BRP - sample NPCs in the back of the book - I exempted the superhero because it's an outlier
  
• Rome - several good "representative" NPCs
  
• Devil's Gulch - a lot of named individuals and a few "representative"
  
• Fractured Hopes - two named individuals
  
• Dragon Lines - Martial Artist, Foreign Devil, Foreign Devil officer
  
• Chronicles of a Future Earth - a few named NPCs who could be considered "representative"
  
• Runequest GM's Guide - has a great chapter on representative NPCs
  

BRP books I own but I did not include in this analysis for various reasons:

• Call of Cthulhu - I just have way too many modules to capture all those NPCs, I didn't have the time. However, based on my analysis below I will go back and do them next.
  
• Stormbringer - the NPCs are just too idiosyncratic and powerful so I felt they were outliers
  
• Elfquest - the stats are for elves, not for humans
  
• Laundry - same reason as CoC
  
• Ringworld - doesn't have a lot of human NPC stats
  
• Superworld - too superpowered, not enough normal humans, so I considered them outliers
  
• original Worlds of Wonder - there's only a handful of human NPCs, but I might include them later.
  

So, here are my findings.

Standard Normal PC (baseline)

AVG STR 3D6 10.5

AVG CON 3D6 10.5

AVG SIZ 2D6+6 13

AVG INT 2D6+6 13

AVG POW 3D6 10.5

AVG DEX 3D6 10.5

AVG APP 3D6 10.5

AVG EDU 3D6 10.5

BRP

AVG STR 14

AVG CON 14

AVG SIZ 13

AVG INT 14

AVG POW 14

AVG DEX 15

AVG APP 13

AVG EDU 15

CoFE

AVG STR 15

AVG CON 15

AVG SIZ 12

AVG INT 12

AVG POW 11

AVG DEX 14

AVG APP 11

Devil's Gulch

AVG STR 14

AVG CON 13

AVG SIZ 13

AVG INT 12

AVG POW 12

AVG DEX 12

AVG APP 12

Dragon Lines

AVG STR 16

AVG CON 15

AVG SIZ 13

AVG INT 12

AVG POW 11

AVG DEX 15

AVG APP 11

Fractured Hopes

AVG STR 13

AVG CON 13

AVG SIZ 13

AVG INT 15

AVG POW 15

AVG DEX 15

AVG APP 14

Rome

AVG STR 12

AVG CON 12

AVG SIZ 12

AVG INT 14

AVG POW 12

AVG DEX 12

AVG APP 11

Rune Quest

AVG STR 13

AVG CON 12

AVG SIZ 13

AVG INT 13

AVG POW 13

AVG DEX 13

AVG CHA 12

Average of All NPCs

AVG STR 13.2

AVG CON 12.8

AVG SIZ 12.9

AVG INT 12.9

AVG POW 12.7

AVG DEX 13.1

AVG APP/CHA 12.2

AVG EDU 15.0

Clearly, those who create stats for NPCs for BRP games tend to score APP/CHA lower in favor of higher STR and DEX. Of course, as I said earlier, my sample does not include CoC, and my sample games tend towards more action-oriented and, thus, are biased towards STR and DEX and away from APP/CHA or POW. Were I to include CoC, I think it would average out more.

Another indication is that NPC stats are routinely higher than default normal PC stats. Rounding off to the nearest integer, the average stat is 13 (12 for APP and 15 for EDU). Based on this, I changed the default characteristic rolls for PCs to 2D6+6 for all stats (Dragon Lines does the same).

Also, I am making sure that all of my Swords of Cydoria NPCs will average to 13.

[MrJealousy]

I quite like a bit of statistical analysis now and then and those stats look quite interesting.

For human NPCs in Aces High i wrote an excel spreadsheet that generates random stats, works out the points values and tells me whether they were Normal or Heroic. But even using random stats and hand picking the results still produces some very odd outcomes.

Noteably, a group of zombies (where I didn't check the results at all) almost all of them sprang up with high STR or SIZ (or both) makes them a bit formidable.

I prefer this random method though. I means I don't get to say 'right I need an NPC fighter, so thats a high STR, CON, and SIZ'. I just look for the set of stats that suggest a competent fighter type, discarding those that wouldn't last 5 minutes. But for townsfolk (like the zombies above) I just accept what the formula spits out. Still produces some weird results though...

I think what your statistics show is what stats are most important for that particular setting... that and I'm fairly certain that none of them were goat herders...

[Thalaba]

Interesting exercise. Why do you think that the average of NPC stats is important?

It's a little difficult to determine how to interpret your results without knowing exactly the spectrum of what you sampled.

One thing you don't seem to have factored into place is the number of foes of any given type. Let's say a scenario assumes there will be from 4-6 players with average stats. The writer would provide one stat block for the zombies that will be faced (say, below human average) even though 20 of them will be faced. He'll also provide one stat block for each major NPC, which might be designed so that they are challenging when fought as a singleton. This would tend to skew the 'average' numbers upward. If the scenario is mainly designed so the PCs face one oponent or a few oponents at a time, then the average stat numbers will be higher.

Personally, I think that balancing each encounter so that the PCs are challenged is more important than making sure the average of all NPC stats in the book are the same as human average.

[rleduc]

Can I play?

Rubble and Ruin has (by my count) 13 sets of human attributes:

Str: 10.46153846

Con: 11.53846154

Siz: 10.53846154

Dex: 13.61538462

Int: 13.07692308

Pow: 10.46153846

App: 11

Not bad... Dex is high, but there is some "self-selection" there, NPCs that are not somewhat above average will be less likely to bother the PCs and will then not need to be stat'ed.

[Evilschemer]

My hypothesis is that designers tend to think "what's important to this NPC?" and assign high stats to what's important. Conversely, what ISN'T important tend to be left at the average of 10-11.

So an NPC fighter might have a STR, CON, SIZ, and DEX of 14-17, while his INT, POW, APP remain at 10-11. Or the NPC Wizard might have a INT, POW, DEX at 14-17 while his STR, CON, and SIZ are 10-11.

This is unconscious on the part of the designers, but it will bias NPCs towards higher stats while normal PCs will have more stats less than 10.

[Evilschemer]

I think what your statistics show is what stats are most important for that particular setting... that and I'm fairly certain that none of them were goat herders...

Runequest GM's Handbook

Herder

STR 10

CON 14

SIZ 11

INT 12

POW 16

DEX 15

CHA 17

AVG 13.6

[Evilschemer]

Interesting exercise. Why do you think that the average of NPC stats is important?

Primarily, I was curious. I was looking at other equivalent NPC stats to give my NPCs appropriate stats, and I was noticing a tendency for published NPC stats to be higher than PCs, on average. So I decided to see if the data backed up my hypothesis, which it seems to have done.

I didn't want my NPCs, for example, to all have average 15-17. I wanted representative NPCs who were supposed to be roughly equivalent to PCs to actually be roughly equivalent to PCs. Named NPCs can vary wildly, of course, but representative NPCs should represent some sort of average.

[Thalaba]

I'm willing to bet actual PCs have higher stats than theiretical PCs, too. I should do an analysis of my current group and see. All of there were rolled using the 3D6 / 2D6+6 method, but they tended to discard bad sets of rolls.

Almost all of my NPCs have higher than average stats, and yet the players usually have no trouble dealing with them. Both of these things lead me to wonder if it's really important whether published adventures have higher than normal stats.

[USAFguy]

It's obvious. That "goat herder" is Odysseus who refused to drink from the river Styx and thus did not forget his past but chose a quieter life when he was reborn. (I don't recall the source, but it is part of Greek myth of what happens to souls once they pass from this plane of existence)

[Guest Vile Traveller]

I think there is a definite tendency to beef up NPCs when you "design" them as a referee, a trait which a lot of scenario writers seem to share. I don't like this tendency, because it skews all the probabilities and results in a power race between PCs and NPCs. When I've made up NPCs for the Spica Publishing Traveller books, I always used the RAW. It makes for a much more varied cast of thousands.

[MrJealousy]

Runequest GM's Handbook Herder

...

POW 16

DEX 15

CHA 17

...

Now that looks like a HOT herder, and would probably be better suited as a Bard or a catwalk model

[Chaot]

Fun.

Elric!

STR 11.5

CON 12.3

SIZ 13.6

INT 12.4

POW 12.2

DEX 12.7

APP 11.4

average 12.3

Here's the breakdown.

STR 12 13 13 12 13 12 13 10 10 10 13 13 14 10 11 12 9 10 10 11 11 11.5

CON 12 14 12 12 14 13 15 12 10 11 13 15 15 10 10 14 10 11 11 13 13 12.3

SIZ 13 13 12 12 13 12 13 16 9 14 13 12 14 11 13 11 9 10 10 13 13 13.6

INT 10 13 12 13 10 12 15 10 12 12 15 12 9 10 14 15 15 13 17 11 11 12.4

POW 10 16 11 12 9 10 16 9 11 11 14 10 9 10 13 16 12 11 15 16 16 12.2

DEX 12 14 13 13 11 13 14 10 11 10 16 13 10 11 10 16 17 13 12 13 12.7

CHA 9 16 11 12 11 10 15 9 16 9 17 10 8 9 12 13 12 10 11 10 10 11.4

I think it may be appropriate to keep in mind that there's a difference in making a random NPC and making an NPC that's supposed to be broadly indicative of a profession. As you noted with the CoC comment, different genres are going to stress different professions, and if you have a list of professions where strength is important, your numbers may be skewed in that direction.

[soltakss]

Now that looks like a HOT herder, and would probably be better suited as a Bard or a catwalk model

But he is a herder, so has herder skills. His father and grandfather were herders before him. Just because he is good looking doesn't mean he should be a bard or a catwalk model - where would he put the sheep?

[rust]

Now that looks like a HOT herder ...

In German a tryst is called a "Schäferstündchen" (= "shepherd's hour"), borrowed from a French word of

the same meaning, because the aristocrats of France's Ancien Regime loved to "play shepherds". I guess

that "hot herder" would have been most welcome to take part.

[MrJealousy]

But he is a herder, so has herder skills.

And thats what it all boils down to in BRP. Your stats may say "You should have been a Bard..." but it is the skills that define you. With this guy, the son of the son of the son of a family of herders, he is going to get an excellent education concerning all things herding and will probably be a good herder himself. But he would have made an excellent Catwalk model. On the flip side someone with ideal Herder stats (whatever they may be) would probably made a mediocre Catwalk model even if he was the son of the son of a family of catwalk models, but he would have been an excellent herder... if only he had been given the opportunity

[Evilschemer]

Keep in mind, the stats for the herder represents what Mongoose thinks should be the stats for the AVERAGE herder, not a specific herder.

[rust]

Keep in mind, the stats for the herder represents what Mongoose thinks should be the stats for the AVERAGE herder, not a specific herder.

Well, Fantasy is a different world ... - for example, according to Chaosium an average Orc

(BRP page 346) with its INT of 3D6 / 10 - 11 is normally slightly more intelligent than an

average Dragon (BRP page 341) with its INT of 10 and almost as intelligent as the generic

Knight (BRP page 363/363) with is INT of 11 ... Tolkien (and others) obviously got it com-

pletely wrong, Orcs are clever and Dragons dumb ...

[MrJealousy]

Well, Fantasy is a different world ...

No kidding! Looking at those herder stats I forgot that this thread was about averages...

Thats really crazy, it's like they had a bunch of stat points left over and decided that to make him not like a fighter they dumped them anywhere but STR, CON and SIZ.

If thats an average, the really excellent herders must be seriously good looking. It all seems like such a waste...

I had noticed the Orcs INT in BRP, I don't remember them being that smart in (GW)RQ... (quick check up)... but they are! I accept that Knights aren't especially renowned for the intelligence but to be comparable with an Orc? Thats just insulting!

[deleriad]

This is one of my bug bears. It's only a minor one but it is irritating nonetheless. When I did the stats for Blood of Orlanth for MRQII I went over the whole lot with a spreadsheet and redid them so that more than 90% of the printed NPCs had stats that you would expect to get if you used the official points-based cha-gen to create them. Meant that the human NPCs were all built on 80-90 points depending on their expertise equivalent. There were some deliberate outliers, both under- and over-pointed but only for specific reasons.

I suspect that nearly everyone who has looked at it hasn't noticed this, but it made me happy.

[rust]

I had noticed the Orcs INT in BRP, I don't remember them being that smart in (GW)RQ... (quick check up)... but they are! I accept that Knights aren't especially renowned for the intelligence but to be comparable with an Orc? Thats just insulting!

It is even worse for the poor Primitive (BRP page 365), with his INT 8 he is only one

single point more intelligent than a "grotesque blob of slimy protoplasmic matter", the

Blob (BRP page 358) with an INT of 7 ...

[PhilHibbs]

NPCs are not random people. That guard at the door, he's been picked for the role. The evil overlord that the PCs are trying to defeat killed the weak and pressed the strong into his service. So PCs are seldom up against crappy people.

And I gave the players 85 points instead of 80, which is still fewer than would be needed to create the sample PC in the core MRQII rules.

I find it easier to put the party of 5 up against 3 or 4 tough opponents than more, weaker opponents, because it involves tracking fewer sets of stats and actions. In the dragon fight on the ship that I posted on the MRQ forum a while back, there were 4 Orlanthi rune levels who all were tough enough to hold off 4 or 5 of the ship's soldiers each. The PCs had to get involved in order to tip the balance.

Edited August 21, 2011 by PhilHibbs

[deleriad]

And I gave the players 85 points instead of 80, which is still fewer than would be needed to create the sample PC in the core MRQII rules.

Well published stats should be based on the rules as written, that way GMs and players know what the baseline is and can adjust accordingly. Same ought to be true of examples in rulebooks. This goes all the way back to the original RQ2 rulebook where Rurik had significantly better than average dice rolls but was actually presented as, if anything, a little sub-par. When I first started playing RQ I used to wonder why none of my PCs where anywhere near as good as he was.

This was particularly important in RQII because it was so expensive to improve characteristics. I know in the campaigns I've run, I've basically halved the increase cost and still the PCs have no more than 1 or 2 points more than when they started 3 years ago.

[rust]

NPCs are not random people. That guard at the door, he's been picked for the role. ...

Since the average INT is 2D6+6 = 13 and BRP's generic Knight has an INT of only 11,

do you assume that noble lords intentionally pick followers with below average intelli-

gence to knight and train them ?

In my view both the Herder's high CHA and the Knight's low INT are caused by the

game mechanics, but have nothing at all to do with the internal logic of a setting -

in any remotely plausible setting the herders are not a bunch of supermodels who

happen to like goats and sheep, and the noble warriors do not all suffer from mild

cases of inherited debility.

Edited August 21, 2011 by rust