Classes As My Balls

I was pondering as I often do about classes in RPGs - or was I self-analyzing my tendency to like contradictory game mechanics? maybe both - when I had a bit of an epiphany.

I've often argued that you "don't need" certain classes in your elf games. Barbarians are just a type of warrior, all casters might as well be wizards, etc. On the total flip side, I'm as much entrapped by the goblin-fueled black hole as so many others in the RPG blogosphere, where there are hundreds of classes ranging from 30 different kinds of "Fighter" (by my count) all the way to person whose dog is a ghost and brings them money.

Liz over at Magnolia Keep has two great blog posts (about combat and player skill respectively) deconstructing the reasons why different groups call the same thing different things for different purposes, and why the way we tend to categorize them is mostly wrong. I happened across these posts at around the same time I had these ponderings, and it fit the format, so here we are. Here is my own take on As My Balls.

Let's talk about what classes are, and why they are actually several different things.

Classes as Classification (Type 1)

This is the one I have argued in favor of in the past. The classic Warrior, Mage, Thief trichotomy is so ingrained into the zeitgeist of RPGs both tabletop and computer that I think it's fair to argue that almost all other classes are subsets or combinations of the three. There is a whole (ironically, classless) TTRPG by almost exactly that name. The original RPG had only fighters, mages, and clerics that were explicitly somewhere between the other two (thieves weren't until Greyhawk, despite their status as an RPG icon).

When using classes like this, the reason they were called "classes" to begin with is clear. They are broad classifications of types of characters. Stepping further back in time, the wargames that RPGs grew from had "classes" of troops, but one Light Foot and another Light Foot weren't the same. I don't think the basic classes were ever meant to be "the" options for characters, but rather the real characterization was to develop within the broader classification.

This is classes in their most literal sense and I like it for the narrative and thematic freedom it grants through its lack of mechanical specificity. Important mechanics are based on a broad category and the rest is worked out in person through play. The monk, archer, pirate, and dwarf in the party are of course different, but the class they all belong to only tells you that they do big hits and don't do magic.

The caster, martial and half-caster groupings essentially function as this kind of class in 5e, if you stopped there.

Classes as Archetypes (Type 2)

Give someone a fighter and it isn't long before they ask for a paladin. It only took a year from the birth of the hobby (from OD&D to Greyhawk) for this to happen in an official capacity. Granted, adding new classes was an obvious move to make and in part its status as a subclass was an attempt at giving Fighting-Men something exciting, but the point stands. With only three or four broad options, people just wanted more.

By now, a culture had begun to develop around these classes. There was some idea of what a Cleric or a Fighting-Man was like. In turn, there would have been players who wanted to play something a bit different, and for some a logical leap was to create a new class. To be fair, it's what Volume 1: Men & Magic tells you to do. A player can play as virtually anything, even a dragon, as long as the referee determines the progression.

For some, spreading out within the bounds of the broader "Magic-User" class either does not come naturally or does not feel as appealing as having a true mechanical differentiation. A Wizard, a Warlock, and a Sorcerer might be the same to me, but to other players this distinction really matters. (Sidebar: I actually think the level titles of OD&D lessened the freedom players felt here, because Wizard, Warlock, Sorcerer and even Necromancer were defined game terms that were only differentiated by level). Take this a few steps further and you get to recent editions of D&D, which provide about ten-ish classes for players to build a character within. These are more specific than "does magic, hits things, or has skills" but still fairly general archetypes.

Everyone knows what the Bard is supposed to be. The class comes with not only mechanical expectations but roleplaying ones too. There is a whole default personality to fall back on if necessary - and I really don't mean this as a dig against modern D&D. If D&D's good at anything, that thing is being the standard, and this approach to classes makes it pretty easy for all players to get something they want out of it without expecting a specific kind of effort in return. With most bases mostly covered, players are free to put their effort where they want to.

This middling number of classes is enough to create a "best of" of well-known fantasy character archetypes while not being expansive enough to induce high levels of analysis paralysis.

Classes as Options (Type 3)

Multiclassing in 5e comes close to bridging the gap to this next one, but I'd argue that it doesn't quite cross it because the classes are still designed with clear and separate individual identities and multiclassing is sort of an afterthought. 5.5e even more so, with the sweeping structural changes to classes making multiclassing less mechanically appealing. Pathfinder on the other hand very much seems to expect multiclassing and has more specific classes in greater numbers.

This is also where you are likely to stop seeing race-as-class entirely. Games that favor options for the players will prefer having races as another option to combine with your class rather than the defining characteristic of your character.

A great example of classes as options is Shadow of the Demon Lord, whose classes are in essence a menu of options to choose from when you level up. Games like Dragonbane or other Basic Roleplaying-based games also fit into this, being "classless" but with your profession and combination of skill selections primarily making up your character identity where your class normally would.

This is for the players who want to have lots of mechanical choices at their fingertips. When I say mechanical choices, I mean the ones that are written into the rulebook rather than presented by the GM. It's not necessarily just for powergamers, though this type of classes does create the kind of game most conducive to powergaming. Even many classless games, which I once thought of as the cureall to powergaming, lend themselves to optimization for those who don't specifically choose to play suboptimally. Having options kind of just does that.

This approach to classes is great for those who value the progression of their character being written into the game for them to strategize their way into rather than abilities being handed out at the GM's judgement. I will grumpily admit there is merit to that, despite my own preferences for diegetic progression.

Classes as Specific Fantasies (Type 4)

When all of these seem like reasonable classes, you are in Type 4.

The complete opposite end of the spectrum from Classification is when you are no longer grouping characters into classes, but creating bespoke classes for character concepts. This is what GLoG classes do. You come up with what you want to do and a class can be made for that. And that character fantasy can just as well be Conan or Aragorn as it can be a werewolf unionist or the classic "weird little guy".

In Type 4, the more classes available to the players the better, and these games are usually designed such that creating a class is quick and easy, maybe even doable in ten minutes before the start of your first session.

Classes as Nonexistent (Type 0)

The philosophy of classless systems counterintuitively shares a lot with the countless-classes philosophy of the extreme end of Type 4. Both of these kinds of systems want the players to have game mechanics ready to go for exactly the character they want to play. While I was writing the section for Specific Fantasies classes, I remembered an interview with someone from Chaosium (I apologize for not remembering the actual source!) where they discussed having played a game with a player whose character was a dolphin in a mechsuit.

That is GLoGgy as hell.

Follow the arrows as classes get more specific.

And I think this hits the crux of why I am drawn to the seemingly contradictory ends of the spectrum. Type 1 and Type 4 aren't actually so far apart when what you want is something that allows the player to do what they want to do without too much bloat. I consider classes a useful and intuitive player-facing tool for this, which explains why I love the two ends of the spectrum and have a soft-spot for classless systems. I consider an excess of mechanical options to be bloated, gameable, and come with a lot of baggage, which explains why I don't Types 2 and 3. But those are just, like, my opinions, man.

DCC across the 2 Axes of Abstraction

Murkdice wrote about what games are about based on what they have rules for and what they don't have rules for. It's an interesting ludological article, though I'm somewhat skeptical of the suggestion that anyone seriously claims 5e isn't about combat specifically because most of its rules are for combat, though. There were some cool four-sector charts in there so I wanted to do one for DCC, and in the process I had a few thoughts of my own about what the axes of abstraction actually tell us.

Go check out the original article. There's more charts!

It's a bit more cluttered than Murkdice's examples but the point of that post wasn't to deep dive into any one game, it was to demonstrate an overall point about what games tell us they are about.

When you illustrate the point like this, it becomes clear that the things the game is about are the things it expects to become complex. This is true across the entire coded-free spectrum. DCC goes to great lengths to tell you about all the things it wants you to do but doesn't have specific rules for, so it's really a perfect demonstration of this.

DCC is about unpredictable magic because it has a lot of rules and charts to follow (making it complex), and DCC is about Questing For It because it leaves the quest-and-reward process open and free (making it complex). You can entirely throw away the axis of codedness and find that the number of rules a game has about a particular thing actually has no correlation to what the game is about.

It turns out a game is actually about what you spend time focusing on while you play it.

A d6 Skill System: 7-in-6 odds

This is a skill system based on old school x-in-6 checks which allows for effectively endless scaling. Obviously, 7-in-6 odds doesn't make much sense, but you're here reading this now so I owe you an explanation. It'll make sense soon.

This is intended for the sort of skills a Thief uses. Climb Sheer Surfaces sort of things - abilities that can't necessarily be attempted by everyone, or if they can, the skilled character gets an extra chance to do it beyond the normal level. I listed several more examples under the Specialist.

The original Thief skills table, a la Greyhawk (1975).

The basic idea here is to have a skill rating be a number from 1 to 6 that represents your odds of success. There are two problems* with this. The first is that 6-in-6 makes failure impossible, and the second is that 83.3% is hugely different from 100% so to stop it there feels like there is a lot more room for progression.

*Whether or not these are problems is subjective.

The twist is to not cap progression, but to soft-cap it instead. Rolling a 6 is always a failure, so with a skill of 6 you still only have 5-in-6 odds, but your leftover 1 rolls over to another die, so you need to fail 5-in-6 then 1-in-6 to truly fail. Thus, at a skill rating of 7, 2 points roll over. At skill 8, 3 points roll over, and so on. By skill rating 11, you hit 6 on your second die and carry the 1 again.

So that's how we do "7-in-6". 5-in-6, carry 2.

Skill Level	d6 Odds	Percentile Equiv.
1	1 in 6	17%
2	2 in 6	33%
3	3 in 6	50%
4	4 in 6	67%
5	5 in 6	83%
6	5 in 6 + 1 in 6	86%
7	5 in 6 + 2 in 6	89%
8	5 in 6 + 3 in 6	92%
9	5 in 6 + 4 in 6	94%
10	5 in 6 + 5 in 6	97%
11	5 in 6 + 5 in 6 + 1 in 6	98%

This works out neatly because, in essence, every 5th skill rating you get one extra attempt and the probabilities become more granular. Though it's absolutely worth noting here that a 5-in-6 is already high-level (equivalent in probability to a Master Thief!), so getting multiple roll-overs is ultra-high-level. Over 10, we're talking about less-than-halves of a percent. It scales infinitely but never hits 100%, though a skill rating of 25 means you need to roll a Yahtzee with all sixes in order to fail (that's 1 in 7,776).

Another really cool thing about this is that it maintains the existence of suspenseful moments at higher levels. A 5-in-6 doesn't feel suspenseful because you know you'll probably win. If anything, it falls flat when you fail an expected success. On the flip side, in this system, if you fail your first roll you now feel suspense on the second roll because this time it feels like you really can fail. In fact, for the first few levels above 5, the odds are really high that you're going to fail the second roll, but you might just scrape through - that brings back some of the suspense from early levels (albeit less frequently, which makes it more memorable when it happens).

Notably, skill rating 14 is the point where your failure rate is approximately 1% (actually 1-in-108). Each 5th skill point is a notable breakpoint where your failure rate halves between levels.

The elephant in the room here is that I haven't defined the actual progression rate for the skills. Oddly, I don't think this actually matters as much, but using the table above you can very easily convert odds up to 98% to their nearest equivalent. For my own games, I'm yet to nail down the exact progression I want.

Hexes with Depth: Overcomplicating the hexcrawl

Today's post is about something I tried that didn't work. It was a cool idea in theory, but it didn't produce the results I wanted.

My basic idea was this: add a second dimension to hexcrawl terrain types. Rather than simply having plains, hills, forest, deserts, swamps and mountains you now have differing intensities of each. Forests become sparse forests, medium forests and dense forests.

I was hoping this would provide a bunch of extra descriptive variety while traversing the hexmap without adding any real noodly details to keep track of. One end of the scale is the most featureless, least difficult to traverse version of a terrain type and the other end of the scale is busy, intense, and difficult. You have three times as many types of hex but still only have to think about the same number of types of hex you were already using.

The problem ended up being that there were three times as many types of hex.

You see, I like a hexcrawl where the players can get lost. If you aren't actually exploring, aren't figuring things out on the fly then there isn't really much point in doing a hexcrawl. Being lost is a part of the experience that I want to create. Traversal itself being a challenge and the possibility of ending up somewhere totally unexpected are awesome and fun.

With a lot of different types of hex, it became too easy to pinpoint one's exact location on the map simply based on the visible surroundings. There was no being lost in the woods, because there are some sparse woods this way and some dense woods this way. There's now only one hex anywhere in the vicinity you could reasonably be in. It became completely trivial, which made the whole process feel like a waste of time.

I ended up throwing the whole idea out in favour of fairly bland hexes with spread-out points of interest to navigate between.

I guess the lesson to be learned here is somewhere between "keep it simple" and "make sure you playtest". If you don't, it's easy to get lost.

Which die is the most honest?

There's a great Goblin Punch article about core resolution mechanics and the pros and cons of each. If you're even the slightest bit interested in RPG mechanics or game design I'd recommend giving it a read, but the general idea can be gleaned from the table at the bottom directly comparing a bunch of core mechanics qualitatively. I won't copy-paste it here, just go look.

What I want to talk about today is one specific aspect of this: transparency. Which die do people understand the odds of most accurately? Which one acts the most like people think it does?

Arnold lists the d100 roll-under mechanic as "perfect" visibility and the d20 roll-under mechanic as "high" visibility. I'm going to make the admittedly weird argument that the d20 in its 5% steps is actually clearer to the average player than the d100.

People do not intuit probability well at all. 65% doesn't feel very high despite the fact that you have almost double the chance of succeeding than you do of failing. Yet somehow, a coin flip feels like pretty good odds to most people. I mean, it's mathematically fair isn't it? Well, until you do flip it and it doesn't go your way. Or you've already flipped once and lost - surely it'll come up your way this time.

Famously, the video game XCOM tried to remedy this by having the displayed hit percentages lie (in a good way!) to the players, telling you that you have a 95% chance to hit but then applying a bunch of modifiers which give you a much greater true chance of hitting. Basically, they wanted people not to feel like the game was cheating them, so they made it cheat in their favour instead. The result? There were still memes about always missing at 95%. The odds are better than what's displayed and people still feel like the game is cheating them.

Human comprehension of probabilities is atrocious.

Finally getting to the point:

Let's map this to the initial examples: 65% and a coin-flip, but in the roll-under d20 system.

• To succeed you need to roll below 10.
• To succeed you need to roll below 13.

How do those statements feel? I don't know if I am an outlier or in the majority here, but the discrete steps of the d20 seem more intuitively comprehensible compared to the mathematical honesty of the d100. I have even seen convincing arguments from some real grognards that the "x-in-d6" of yore is clearer again, and I find it hard to disagree. Who knows, six might actually be a sweet spot between granularity and comprehension, even though it seems backwards that 16.67% steps would be intuitive.

The Ten Foot Pole Store

Read this post as if it's a goofy merchant character trying to sell you something. Prices are given in ten-foot-pole-equivalents.

Ten-foot poles:

Ten-foot pole - 1 - it's a pole, about ten feet long. Cumbersome but useful.

Decempeda - 1.5 - it's a pole, exactly ten feet long. Perfect for precise measuring, still cumbersome.

Jointed ten-foot pole - 2 - Unscrews at three points so it can fit in a pack or something. Can be separated into four 2.5 foot poles. Takes a minute or so to set up and pack up. If you buy multiple you can construct a longer pole with the parts! But the ends only have one joint, so you only get 5 extra feet per additional pole.

Folding ten-foot pole - 2 - Folds up in a zig-zaggy pattern with joints that lock. Quicker to fold and unfold than the standard jointed model, but can't be separated into four 2.5 foot poles.

Telescoping ten-foot pole - 4 - Very quick and easy to extend, but a bit flimsy. You can swing it to open it and you will look very cool.

Elastic ten-foot pole - 10 - Can be bent all the way round so that it touches end-to-end and still spring back into shape. Could be considered a minor magic item, as the material it is made of is fantastical in nature.

Other ten-foot products:

Ten-foot ladder - 1.5 - Two ten-foot poles connected by a bunch of rungs. Probably useless, to be honest.

Ten-foot rope - 2 - An amazing pole that can twist and bend in any direction!

Ten-foot chain - 20 - Like the ten-foot rope but metal!

Ten feet - 1 - From rabbits. Very lucky. Enough to share.

Magical ten-foot items:

He probably doesn't have all of these on hand - maybe the ones he doesn't have are legends he is seeking.

Staff of Ten-Footness - 100 - Once per day, a successful DC 15 spell check makes anything shorter than ten feet expand to ten feet long. Now I know what you're thinking, and I will not be liable for any damages!

Immoveable ten-foot pole - 250 - The infamous immoveable rod but ten feet long. Very dangerous.

Sun's Pole - 3000 - Extends and contracts at the will of the holder, but is very heavy. Has seemingly no upper limit on length.

Hitting a hydra

This is a mechanic for determining which head of a hydra gets hit. I swear this is simpler than it looks. If the logic doesn't immediately make sense from the chart, read on rather than scratching your head over it.

Heads	Dice rolled
1	-
2	-
3	2d2-1
4	1d2+1d3-1
5	2d3-1
6	1d3+1d4-1
7	2d4-1
8	1d4+1d5-1
9	2d5-1

In some games (e.g. DCC) you are already tracking the HP of each head of a hydra. In some (e.g. 5e) you are not, instead tracking only a, HP total and the number of heads. If you're already keeping track, this isn't a ton of extra overhead.

If a player chooses to strike at the hydra without aiming, roll the listed dice to determine which head is hit. The dice are simply a pair of dice whose maximums add up to the number of heads, plus one. For instance, 9 heads, plus 1 is 10, half of that is 5 so you roll two d5s. The reason you need to subtract 1 is because you can't roll a 1 on two dice, so you end up with as many possibilities as heads.

If there is only 1 head to hit there's obviously no need to roll, and if there's only 2 then just let the player hit whichever one they want. I considered a 3-in-4 roll to hit the desired head but it's getting too complex at that point.

Optional extension: Aiming

If the player aims for a specific head, treat that head as the "center" roll: the most likely roll on the pair of dice (the smaller die total plus one (yes, with different die sizes there will be two equal results, that won't matter)). Lower numbers go down the line, higher numbers go up the line, and the ends wrap around (aiming for the first head and rolling low can hit the last head).

This lets players be strategic if certain heads have certain abilities, rather than the center ones probably dying first. It also allows for players to try to kill the heads all at the same time.

Alternate version

Roll the two dice together. The leftmost die represents left, the rightmost die represents right. Literally just use where they land on the table relative to you. The difference between the two dice determines how many heads the player misses by, and which one is highest determines the direction.

This version came from a hexcrawl "lost" mechanic where two dice determined how far left or right travelers drifted.