Sunday, December 31, 2023

Toadman Champion Names

Hypothetically, if one were building a dungeon level full of vicious toadmen, one would need names for the champions of their gangs (the ones who speak Common).  Many of these are combinable into compound names. 


  1. Ribbert
  2. Hroat
  3. Hreeen
  4. Brep
  5. Hrerm
  6. Hrup
  7. Peep
  8. Hreet
  9. Chup
  10. Kermax the Eviscerator

Friday, December 29, 2023

Soulflayer Canyon and One-Way Doors

I picked up Dragon's Dogma on the steam holiday sale and it's been a refreshing reminder of something one of my old bosses told me - "You can innovate on the technology, or you can innovate on the business model, but as a small company you probably don't have the resources to do both."  Dragon's Dogma declines to innovate in its setting.  The enemy roster is pretty much all classic D&D monsters played straight - they even have a beholder with the serial numbers filed off.  There's a quest where you rescue a princess from imprisonment in a tower and literally carry her across bridges and over gaps.  There are plot holes large enough to ride a griffon through and plot-agency is pretty negligible (why am I working for this asshole duke anyway?).

But the core combat gameplay!  Maybe for Capcom the combat gameplay isn't really innovative.  But taking the language of fighting games, of grabs and throws and parries and knockdowns, and applying them to giant fantasy monsters, played totally straight rather than FromSoft-style "everything is corrupted and weird", is just...  a lot of fun to fiddle with.  To say nothing of the pawns.

But I'm really here to talk about the design of one particular dungeon in Dragon's Dogma - Soulflayer Canyon.  It's a real piece of work.  "Criminally vicious", as the Tucker's Kobolds guy would say.

Spoilers beyond this point.

Soulflayer Canyon has a number of pretty vicious encounters - a cockatrice, ghosts who possess your henchmen (backed up by camouflaged lizardmen), a cyclops on a narrow bridge whose club will absolutely fling you and your hirelings down a long fall to your deaths, harpies who try to grab you and pull you off ledges...  But the thing that makes it really nasty is that it's full of one-way waterslides, rock slopes with water running down them that you can descend but not ascend, where you can't see what's at the bottom until you go for it.  Topologically, the dungeon is mostly a loop of one-way slides (with ladders in between to make up the height losses) with a couple of branches (one is to the treasure, another is to an exit from the dungeon).  I'm not sure it's possible to exit the dungeon by the door I came in through once you've entered the main loop.

There came a point where I'd basically cleared the dungeon and was faced with a choice between three slides.  One went to the treasure, one back into the loop, and one I think to a terminal fall.  I chose the loop and had to re-run the dungeon, some of which had restocked.  Dragon's Dogma has enough mundane resource management of healing items and lantern oil for this to be a really worrying twist if you were already running low.  The cockatrice's lair is also at the bottom of a one-way slide, and while there is a climbable rock wall that you can use to get back out, you probably have to go through the cockatrice to get to the exit.

So anyway, it's a wild dungeon.  The other dungeons in the game aren't like this (mostly).  It's like they took all their most vicious ideas and put them into this one zone that only sidequests point you to.

I had been thinking about using one-way doors in gauntlet dungeons, so it's been interesting to see them in action here.  One thing I like about these waterslides is that they're pretty telegraphed.  They're not a literal door that closes behind you but is indistinguishable from a two-way door until crossed.  It's probably worth thinking up more types of clearly-one-way "doors".  It was also interesting to see a dungeon with a single main loop composed primarily of one-way doors; I had been thinking about one-way doors used sparingly in the context of dungeons composed of multiple intersecting loops, where there are almost always multiple paths to any point.  But Soulflayer Canyon goes all-in on them and it certainly makes for a memorable "level".

Monday, December 18, 2023

Classic Traveller Jump Tech Levels

Omer of Stellagama made an interesting remark about Classic Traveller on discord the other day:

Ship hulls, computers, and drives interacted very differently with tech levels. Most Traveller players know the Book 5 (High Guard) jump tech levels, which are rigid (Jump 1 at TL9-10, Jump 2 at TL11, etc.) but permit big and fast ships even on low TLs. Book 2, on the other hand, has drives gradually appearing as technology progresses, so that lower-tech ships must be smaller and slower. On the other hand, its computer and jump rules permit Jump 3 at TL9. So that big ships, especially fast big ships with far jump drives, are locked behind higher TL, all while permitting longer-range jumps for smaller ships at lower TLs. This is conducive to adventures, as even lower-tech worlds can manufacture player-size ships capable of traversing longer interstellar gulfs, and as smaller ships make mass planetary invasions a very costly affair - making diplomacy, guile, and subterfuge much more important than simply bringing in your 100,000 ton battlecruiser with a Spinal Mount that can shatter moons and threatening everyone, all while your whole Marine armies land from a few 50,000 ton troop carriers.

I hadn't made particular note of when different jump techs became available in CT (and I haven't read CT's High Guard), so I went and did a bit of digging into the details.

CT Book 3's tech level table on page 15 notes that TL 9 can manufacture drives A-D and jump drives, as well as computers up to level 3.  In Book 2 on page 58, class C jump drives and powerplants are enough to give a 100-ton ship jump 6, while class D drives give 200 tons jump 4, 400 tons jump 2, and up to 800 tons jump 1.  The limitation to jump 3 actually emerges from computers, not powerplants or jump drives, since a computer/3 can only navigate/perform/drive 3-parsec jumps.

At TL 10, jump-capable 1000-ton ships become possible; at TL 11 2000-ton starships.  3000-ton starships don't arrive until TL 13, skipping 12, and 4000- and 5000-ton ships are at TL 15.

Looking at 400-ton ships, at TL 10 they go from max jump 2 to 4; at TL 11, jump 5, and at TL 12, jump 6.  800-tonners get jump 2 at TL 10, jump 3 at TL 12, jump 4 at TL 14.  So an 800-ton jump-3 mercenary cruiser is TL 12 just off of the drives (and TL 11 for the computer/5).  The patrol cruiser looks doable at TL 10.

I was curious to see what Mongoose Traveller 1e had to say on jump drives and tech levels, but the MgT 1e core book didn't really say anything about the tech levels at which different ratings of fusion plant and jump drive become available.  There are some subtle differences at the top end of the powerplant table, and the more relevant difference is in the computers table, with computers of each model available about two TLs later than in CT Book 3.  They are also, however, much less expensive.  So MgT is closer to the Book 5 paradigm, of higher jump numbers arriving across all ship sizes as a given TL.

I was curious just how common worlds of various TLs were under Classic Traveller's world generation assumptions, so I scripted it up, generated 100,000 worlds, and counted.  The numbers below assume an 80-hex subsector with a typical density of 0.5 stars per hex for 40 inhabited systems.

  • TL <9: 66446 (66.4%) - about 2 in 3, 26.5 per subsector
  • TL 9: 10,364 (10.4%) - about 1 in 10, 4 per subsector
  • TL 10: 8,244 (8.2%) - about 1 in 12, ~3.3 per subsector
  • TL 11: 6,075 (6.1%) - about 1 in 16, ~2.5 per subsector
  • TL 12: 4078 (4.1%) - about 1 in 25, ~1.5 per subsector
  • TL 13: 2463 (2.5%) - about 1 in 40, 1 per subsector
  • TL 14: 1371 (1.4%) - about 1 in 70, ~0.5 per subsector
  • TL 15: 632 (0.6%) - about 1 in 170, ~0.25 per subsector
  • TL >15: 327 (0.3%) - about 1 in 300, ~0.12 per subsector 

The highest TL code I saw was a single J, which is...  TL 19?

A-100A56-J    R I  Hi Ht In Na Va

Surprisingly mellow government and law level for that population...

So what does this mean under CT Book 3's rules for powerplants and jump drives by TL?

2/3 of planets basically can't manufacture starship parts at all.  However, model 1 computers are available at TL 5.  45.6% of worlds are in that TL5-8 range and can plausibly provide replacement computer parts, which is nice I guess.

Depending on how you feel about the Industrial trade code, it might not even be possible to build ships at scale at most of these TL 9+ planets.  They can probably nanolathe you a replacement flux combobulator for your drives, but building ships there is a harder sell.  I count 1797 worlds of TL9+ with the Industrial trade code, or 1.8%, 1 in 55, about 0.73 per subsector.  A quarter of typical subsectors can't meaningfully manufacture starships in quantity (to say nothing of low-density subsectors with fewer planets).  There were 563 TL 12+ industrial worlds, 0.6%, 1 in 160, about one per four subsectors.  If you want to have a mercenary cruiser built, it's going to take weeks for an X-boat to even deliver the order to a major TL 12 shipyard.

(TL 7 Industrial worlds may be required to mass-produce non-starship spacecraft; only about 1000 of the 66000 TL <9 worlds are also Industrial.  So probably system defense boats are also mostly imported, under the math I did here about how heavily-defended small worlds in Traveller are?)

This distribution of TLs is all roughly true in Mongoose too (there is one small point of divergence, where in MgT balkanized governments get +2 TL?), but being computer-constrained opens up an interesting angle.  Computers are small but complicated and value-dense.  Since the technical details are under-specified, upgrading a computer/3 to a computer/4 on a TL10 world might not require a heavy industrial base; maybe just a few high-TL pieces that can be manufactured in small batches to glue a bunch of lower-TL hardware together.  Adding one more dTon of computers is like, one extra server rack.  That seems less daunting to retrofit than adding 8+ dTons of fusion reactors and technobabble jump drive bits (sadly it's not quite that simple for either the scout ship or the free trader, whose jump numbers are constrained by their drives as well as their computers.  Still, it's possible to build over-reactored low-TL ships and then just need a higher-TL computer upgrade).

(This does also raise the possibility of extended supply chains for starship construction - if you have a TL 9 industrial world and a TL 12 non-industrial world, maybe getting a hand-crafted TL 12 computer in your starship is an optional upgrade package that the shipyard offers.  Maybe they need high-value computers delivered from across the subsector and it would be a shame if such a shipment were waylaid...)

The same value-density also makes upgraded computers great treasure / salvage.  If you find a high-TL derelict, you're probably not taking the whole engineering section.  But recovering the computer sounds pretty doable and could be a big increase in your capabilities.  And starship programs are also super value-dense (also, you need different jump control programs to jump different distances - you can recover a big computer from an old ship but if they were using it mainly for combat programs and didn't have that big a jump drive, like the Mercenary Cruiser, it may not help you immediately).  I could see collecting starship programs in Classic Traveller almost being like collecting spells as an MU in D&D.

Anyway.  The implied industrial base for shipbuilding in Traveller is much smaller than I thought.  It's not quite Stars Without Number level but it's kinda borderline; there might be extended supply chains and it could be brittle and prone to disruption.  And having jump on small ships at typically-available tech levels be computer-constrained rather than reactor-constrained does seem like it would open up some gameplay in addition to Omer's points about creating space for player-party size groups.

To a certain extent this also addresses the bulk cargo problem Traveller has had, where it's hard to justify tramp freighters making a living when you could just build enormous heavy freighters.  MgT 1e has a 1000 dTon bulk freighter design with Jump 2; it would be TL 11 under CT Book 3 rules, vs TL 9 in Mongoose, which more than halves the number of Industrial worlds where they can be built.  550 tons of cargo is about 7x the cargo volume of a free trader; enough to disrupt a route maybe, but not orders of magnitude more.  Its description notes, "this thousand-ton vessel is still tiny compared to the mammoth corporate vessels that also ply the trade routes", but under CT Book 3, a 4000-5000 ton vessel isn't jump capable at all until TL 15, and jump 2 at 5000 tons and jump 3 at 4000 tons are just barely achievable.  TL 15+ Industrial worlds to build these enormous type Z powerplants and drives are exceedingly rare - 80 in 100,000 worlds, or 0.08%.  A full sector of 16 subsectors has a 50% chance of having such a world.  And then you need jump-2 routes connected to that TL 15 Industrial world for them to actually get where they're going!  So there are going to be places these big ships just can't go that smaller ships locally-build ships can get to, or places where smaller high-jump-number craft can get to much more quickly.  And 5000 tons is certainly not small, but it's a far cry from the 100,000 dTon superfreighter in MgT High Guard with 50k dTons of cargo capacity.

Wednesday, December 6, 2023

Curated Outdoor Spaces

I visited the Chicago Botanical Gardens the other week and had a sort of a weird experience.

I'm used to parks and gardens which strive for a veneer of plausible wildness, if you will - sure, there's a paved trail cutting through it, and yes, they're a bit organized, with a grove of cherry trees (in a variety of labeled types) here and a grove of a variety of magnolias there, but there's still a fair bit of undergrowth and the groupings of plants could have occurred naturally, and you stay on the trail mostly because you don't want to get into the weeds (and the "beware of hornets' nest" signs).  Obviously these environments wouldn't fool an experienced outdoorsman used to bushwhacking, but I don't think they're intended to.  Maybe a better description of what they're aiming for is idealized, Edenic - the woods you used to play in as a kid, where your parents told you the names of the trees and where the hornets' nest was.  Not wild wild, but an environment kinda doing its own thing with limited human intervention to make it suitable for human enjoyment.

The Chicago gardens weren't like that - there were lots of "stay on the trail" signs and art pieces in among the plants.  Parts of it (like the aquatic plants exhibit) reminded me of The Witness - I kept expecting to find puzzles to block my progress rather than just art objects.  There was no escaping the artificiality; a veneer of nature was not a design goal.  Chicago's garden felt like a heavy-handed exercise in power, in control, in making appearances just so.  The outdoor environment as a canvas, something to be written to before it is read from.  When this succeeds, as in their Japanese-style garden exhibit or the desert greenhouse full of crazy cacti, it can be quite beautiful.  But parts definitely fell flat or felt forced.

I think there are a couple of points of relation to D&D and the design of artificial environments.

I don't think I've ever attempted to design an outdoor adventure site that was heavily-modified by its inhabitants (beyond, say, fortresses), but Gardens of the Elf King does sound like a TSR module title.

I think on the other end, it's probably worth distinguishing between attempts to provide a simulated wilderness as it would actually be in a fantasy setting (with eg simulated migration of monsters) and aiming for an idealized wilderness of myth and fiction, full of dragons and treasure and talking birds.  The latter is more compatible with the Tiki Style of Early D&D - the players are tourists into this mysterious place and we don't need to have everything that happens in the "backstage" worked out because it's more important that it feel right than that it be right.

Maybe part of my interest in the wilderness game has been that wilderness is almost by definition unmanaged - it does not require simulating minds/agents to build a plausible wilderness, whereas a dungeon wants for explanation and justification for its construction.  And then the gauntlet funhouse dungeon appeals to me in part because it is a rejection of in-world explanation, design, purpose.  Rather than making a dungeon of the wilderness, it makes a wilderness of the dungeon.

Saturday, December 2, 2023

Starship Geomorphs

I was poking through GDW's Classic Traveller catalog in DriveThru the other day and made a remarkable discovery: this Starship Geomorphs listing.  It's a free pdf of about 200 pages of deckplan geomorphs for building large starships, space stations, and just general sci-fi environments out of.  Most of the geomorphs are 20x20 5' squares, or 200 dtons.  There are a couple of sample ship deckplans demonstrating how some geomorphs might be put together and they're in the 700-1000 dton range.

It seems like it would be really useful for HOSTILE, since HOSTILE has a focus on really big, industrial ships and stations, and they're often adventure sites when something gets loose (so you might want a high-fidelity map).  And the price is certainly right...