Monday, May 30, 2016

A Database of Traits

Hm.  Well, I didn't know it was Memorial Day weekend.  But Trump is still a problem, no matter what war ended 151 years ago.  Still, I apologize for that.

I was thinking about ship captains yesterday and today.  When last I played with my party on the south coast of Morocco, they had discovered some things about the ship's captain they hired to manage the caravel they bought.  At the time, I had thrown together a three prospects, each NPC with a bit of information with the party and a story behind them, that would play out as the party continued with their quest.  The captain they settled on was Genoan; he knew the Canary Islands, where the party wanted to go, and the party learned he had been at sea for most of his life and that he had little desire to go home.

The party has since discovered that he's a wanted man in Genoa - but that's how it goes sometimes.  They're nowhere near Genoa or a Genoese ship, being in Spanish and Portuguese waters now (neither of those with much love for Genoans), so they're satisfied they can control the situation.  These are 10-11th level characters; they can handle themselves.

But as I was working on my trade tables last night, taking a break from trying to work out some troubles in Chapter 17 of the book I'm writing, I found some notes surrounding the ship prices regarding the payment for captains.  I had made these on the fly for the party when they first hired the captain - so I decided to add the price for a captain to my tables.

That got me to thinking.  There are captains and there are captains.  Some will know navigation and others not; some will be excellent pilots and others not.  Some will have naval training; some will have only worked on merchant ships.  Some will be prepared to look the other way when the party does something questionable and some will not - and there will be different degrees of discipline, risk-taking, knowledge of sea lanes, knowledge of odd areas, knowledge of big oceans vs. little seas, better skill in storms, better skill in getting the most out of selling cargo and so on.  Not to mention that some will steal or stab an owner in the back, while others won't.

For most, the thinking only goes so far as these being interesting possibilities for characters that party may have to deal with.  But I think it is more than that.  Unlike other professions, there aren't that many captains able to manage a ship; and the number is small enough that some captains acquire a reputation - and that reputation affects how much they will ask for at the barrel head.  It isn't just a matter of captains saying they'll take 200 g.p. per month: some will ask for more because they can get it - and some will ask for less because they have to.  When some captain has been four months in some backwater, just this side of quitting straight work and becoming a pirate, there's a chance that the hiring price could be less than that of a good crewman.  Which the party may accept, because they're in the same backwater and there's no one else.

I'd like some sort of system for working out prices based on those issues raised above - simplified, of course.  Much of it can be piled together into "reputation."  Much of the rest can be applied to "competence" and the remainder to "experience."  We can't use the word experience, of course, because in game terms its always easier to come up with a new word that refers specifically to that thing - I like "seasoning" because it has the suggestion of weather and the sea in it.  A young captain may be competent and have a good reputation, but without a lot of years on the sea, it's not the same as an old sea dog.

Three qualifiers is enough.  Just now, I'm not sure how to reflect them in the costs - or how much each should balance.  The idea is new and it deserves further consideration.  But I think (and this could be important) that this sort of framework could be important for determining the value of non-player characters as a whole.

What's wanted is not just a standard that applies to ship captains, but to everyone.  Some universal framework that would establish (preferably in a random fashion) the likelihood of a particular individual's "personality" - and from there a set of rules for the likelihood of a given personality to respond in a particular way.

Let's say we start with a simple cubic diagram, with an x, y and z axis.  We could roll 2d6 upon each axis, defining a result of 7 being equal to 0.  Therefore, 12 = +5, 11 = +4, 10 = +3 and so on down to 2 = -5.  All negatives upon our axis would suggest bad personalities and positives good personalities.  Someone with poor seasoning would be either young or sheltered; someone with a high reputation would be stable and perhaps of high status.

With three axes, we would define 1331 different personality types, with 116 of those personalities having no deviations more than 1 from center (in none of their three traits); 394 would have personalities that would not deviate more than 2 from center; and 770 would not deviate more than 3 from center.  The number of total psychotics, where all three traits were at their maximum deviation, would only be 1 in 1331 personalities, or slightly more than 2 per 100,000 population.

I think there's a remarkable potential here: but a huge amount of leg work defining just what those 33 deviations are for each axis (+1 for no deviation at all) - and then further work defining pairs between deviations.  Such as, what does it mean when someone has a -2 reputation and a +1 competency?  A person could go crazy just figuring out all 363 of those pairings.

What a database it would be, though?


  1. It strikes me that it might not be quite so ridiculous...

    Since seasoning is a representation of career length pretty much, it comes down more to the competency and reputation, then you can use seasoning to narrow down whether or not it was one spectacular incident (lowest seasoning, or sub roll for higher seasoning) or a lifetime's representation of steady competence and reputation.

    If you wish you could chop up both reputation and competence for the one incident vs lifetime.

    This results in two real dimensions here competence and reputation.

    If you want to simplify those you can break those down into positive positive, positive neutral, positive negative, neutral positive, neutral neutral, neutral negative, negative negative. And each of those with a descriptor of reliable history of this or a single noteworthy (or maybe not so noteworthy) defining event.

    This still results in way more captains than any one adventure may need, which presumably would be applicable for general use. Especially if you already have some random rolls for origins and so on...

  2. You presume that seasoning makes competence. Speaking for myself and my tremendous ability to know something is a stupid thing to do but not being able to keep myself from doing it, I wouldn't say that seasoning automatically improves a person.

    Think of it more in terms of stubbornness, if you will. A bad person is stubbornly bad; an incompentent person is stubbornly incompetent. Just because they've done something for years doesn't mean they don't insist stubbornly on doing it the same way year after year. Think of the way that some people drive.

    Not thinking in the least way of simplifying this further. In fact, it's pretty simple if you figure you only need 11 traits for each subdivision of personality. That's only 33 total traits that are needed initially to use as guidelines. That's no so many!

    Imagine, instead, that you set each axis at 4d6 instead of 2d6, giving 21 different traits for each subdivision - only doubles the work and gives 9,261 personalities overall. Wow!

    Your suggestion of only 7 combinations would very, very quickly become tiresome over a campaign. It may be enough for captains, but what about all the other NPCs in the universe? I create 7 NPCs for a single night's running.

  3. I see potential in this system for defining personality and/or some other trait or aspect of a character.

  4. I want to go have a head-scratch about how the personality rating would manifest in certain responses, as you touched on. That sounds like good fun.

    Please allow me to take a stab at some traits for the Seasoning quality:

    -5 sheltered or badly trained (perhaps only has a "theoretical" understanding of the work)
    -4 overly young (for their given profession)
    -3 not completely trained
    -2 still an amateur
    -1 greenhorn (officially professional but w/ no real experience)
    0 has seen 1 battle, manufactured 1 quality item, wooed 1 important client, etc (depending on profession)
    +1 the local go-to guy
    +2 the go-to guy for this profession in a 20-mile radius
    +3 has seen many battles, has received recommendation from minor noble, been held on retainer by an esteemed community figure, etc
    +4 has achieved high rank in the appropriate professional organization, has won an prize/award/commendation, etc.
    +5 best-in-kingdom status? has been made a knight? controls a professional guild? is renowned as a tastemaker for (certain type of art/fashion)? has the ear of (admiral/king/bishop/etc)?

  5. A few points, Maxwell,

    Your list has clearly missed that the chance of getting a +5 in any personality category is only 1 in 36 (boxcars on 2d6). This would make ten 'best-in-kingdom' knights/guildmasters/admirals and so on for every village of 360 adults. A bit much.

    Secondly, a label is ineffectual. We need a list of concrete game mechanics for each trait. For simplicity, let's not define seasoned as "stubborn" as I did above but leave it as strictly aware of the world/not aware.

    So, someone with a -5 seasoning level might be one or more of the following: A) untrained; B) never encountered combat before; C) lives with parents; D) has never left village or neighborhood; E) fears strangers; F) xenophobic. Morale: very poor. May not be leveled, will certainly be low level. Useful: for local information only, may be unfriendly, may resist communication.

    Shrink your expectations to things that anyone might be; this is a medieval world so think in those terms. Someone in the middle is probably not the go-to guy; the middle means been out of the village a few times, knows enough of the surrounding hexes to use them on occasion, knows most common people in the village or neighborhood, this isn't their first rodeo but on the whole, AVERAGE.

    The go-to guy is the +5, who has met everyone in the neighborhood, has been in combat before, could maybe be 3-6th level, makes a good guide and knows what a really big city looks like. Got to think of the knowledge they can offer like circles that get wider or smaller as we get away from center.

    Looking for more universal traits than the jobs they hold: what are they like as PEOPLE. That is the main thing.

  6. I think you'd be doing us all a great disservice if you didn't refer to seasoned captains as salty dogs.

  7. I laughed at the idea of Admiral Village. Thanks for the reality check.

    Of course there would be mechanical associations for each result -- but no, I didn't write any.

    This is fun to think about. I expect you'll probably beat me to a working first-draft generator but I will keep musing. Everyone's always wanted a NPC generator that spat out something more USEFUL than "Fighter 2nd level, with crossbow, leather armor, eyepatch", right? And any crappy random generator can assign personality traits, but you're right, you've absolutely got to have actual game mechanic considerations to be worth anything, or else it's just something you have to remember to mention whenever the party talks to that one guy. Which is no help at all. It's got to limit (or expand) the actions the NPC can take, or explain how he takes them, or make concrete what he can provide to the party, etc.

  8. I was actually thinking Seadog, Tim. Jack London, I remember, preferred sea wolf.

  9. As far as applying the scales to your trade tables, a multiplier for each axis ranging from 0.5 for the -5 trait up to 1.5 for the +5 trait might be easily managed. So the sought-after 1331st could command a fee of 337.5 GP where the absolute average Josef could only get 100 GP and Mr. bottom-of -the-barrel only asking for 12.5 GP with his three 0.5 price multipliers. You might prefer smaller increments (or bigger on the positive end), Or to weight one trait more than the others you could increase the steps for one axis.

    At the very least, this method leaped out at me for how to apply it to the costs.

  10. This is a great foundation for building interesting non-player characters on. There was a thought that popped up for me when you pointed out the statistics on a village level: isn't the level of rarity going to differ based on how unusual the profession being rated is? You're going to come across a fellow with straight +5s across the board once in ~46,000 times either way, but 1 in 50,000 sea captains is less common than, say, 1 in 50,000 farmers.

  11. I was thinking along those lines as well, Shelby - with adjusted weights to each axis depending on the profession being hired. A competent labourer isn't nearly as important as a competent naval captain would be.

    Quite right, Jonathon, and thank you for correcting my hastily done math. Mathematically, we're unlikely to come up with a sea captain that has either extreme - there are perhaps ~46,000 sea captains in all of my world - and those with the highest ratings (4s & 5s) will be in the employ of governments, we can be sure.

    We may want to assume that anyone with a rating of, say, 3 or up are more interested in working for themselves, for god or king, rather than getting hired in a random port. Of necessity, parties should recognize that ANYONE they can hire is probably the dregs - and then reserve the higher numbers for followers that are gained through reaching name class status. That makes a big difference between the 20-200 men-at-arms that a fighter gains at 9th level as compared to what the 8th level fighter can hire at the market. Hm. Hadn't thought of that before just now, but it makes sense.

    Lots of different ways to build this - but I do see it as a really big project, something more than just a set of tables. Deserves programming to handle all the variables.

  12. Alexis,
    Like this a lot. Complexity from Simplicity
    From the player's/pc's view, how could they potentially perceive the differences in seasoning or competence? Same question for perceptions of levelled NPCs. What way in your game can pcs perceive lying or potential danger from npc?.. is it dm roleplay body language clues, some sort of wis roll "she seems to be lying"...

  13. Probably not, kimbo. It is more a basis for how an NPC would act over time - information to be discovered later, rather than clues to help players avoid picking bad people up front. The only 'clue' would probably be the cost per month - but since players don't know what that should be anyway . . . they always just pay what they're told.

  14. Thanks for this idea, it helped me overcome a major hurdle I had faced. I took this idea to do the same with merchants, though I shortened the range to -2 to 2 as I only needed ~100 NPCs, and I distributed the values via bell curve.

    Some of the programming can be done in Excel through use of arrays. On the roleplaying end, I found it helpful to consider that a poor Reputation doesn't necessarily mean the NPC is a cheat; just that people think he is.

    As for mechanics, as mine were with merchants, I tied each axis to a major variable (Reputation -> Traffic, Connections -> Stock, Bartering -> Sales), and had other factors (type, quality, size and location of the shop, plus modifiers for local crime rate and the merchant's personal affiliations) tie into each of those variables, as well as a bit of random chance, to derive the merchant's income.

    It at least helped create the skeleton of what I'll be working with going forward.


If you wish to leave a comment on this blog, contact with a direct message. Comments, agreed upon by reader and author, are published every Saturday.

Note: Only a member of this blog may post a comment.