Tuesday, 23 March 2021

A first stab at the megadungeon campaign

 


A couple of weeks ago, some of our regular players couldn't make our daily D&D session. I decided that this was an opportunity to start up the episodic megadungeon campaign. This is my long-planned set-up for spontaneous play on just such occasions. There was just one problem: I hadn't yet designed the dungeon. 

That, I decided, need not detain us. I procured a highly rated free adventure from the web (I won't give its name here as we aren't yet done with it, but it might concern a sepulchre of sibilant segmented slitherers ...). And off we went.

This gave me the opportunity to test out some of the mechanics I've plotted out for episodic games - notably the last lantern-bearer. I had every player roll up three characters each beforehand, but instructed them to equip only one with the conventional 3d6 x 10 gp. 

That was the only time that a gold piece was a gold piece. Taking a tip from Little Odo, I've rebased the currency to silver for the purposes of treasure and XP - so that equipment is much more expensive after the initial kitting out of a character. I don't mind a starting character showing up in plate armour - but I want it to be worth prying from his corpse if he comes to harm. 

Players also have to choose between spending silver pieces looted from the dungeon or using them for XP. I've made the same switch in our regular D&D campaign (still on the gold standard), and I like it. The advantage for higher-level characters is that all living and travel expenses are hand-waved as "in the XP" - but the players are still left hungry for loot. 

Anyway, the first session went nicely. We were using the Swords and Wizardry rules, which I really like. We had four players, and the party consisted of an assassin, an elf fighter, an elf thief and a magic-user. They decided to hire some extra muscle, so recruited two additional fighters - one first level and the other second. That meant that treasure would be split seven ways, with Carla the Clumsy (the second-level fighter) taking two of those shares. 

For this adventure, I decided that the mules and porters/lantern-bearers would remain outside the 'dungeon'. My megadungeon will chiefly consist of spacious hallways and wide subterranean roads, but this outing involved caves and a tomb. The map, however, allowed me the conceit of having the porters within earshot at first, so that potential reinforcements were at hand. 

The players advanced cautiously and overheard some orcs talking. The elf thief realised that the party had been overheard. During a period of indecision, one of the orcs crept round to the party's rear using other tunnels and began shooting arrows. They drove him off with missiles and managed to make short work of his two comrades. But when they were beginning a rope descent into a lower level, the orc archer returned and shot dead one of the NPC fighters holding the rope. That led to chaos, with some of the party getting into trouble on the lower level while one was left on the upper level to deal with the orc. A well-placed arrow eventually did for him, but two of the party had died below and Carla the Clumsy was badly wounded (0 HP). The lantern-bearers were summoned from outside, the wounded Carla was carried out, and two new PCs replaced the fallen from the ranks of the hired help. These two were unequipped, of course, so they had to scavenge gear from the fallen - less than ideal given the mismatch of character classes. 

Two players had to go at that point, so one original PC left with the wounded, but one of the original characters and a replacement made their way a little deeper into the lower level. Only one came back - but with a fair bit of XP and a little loot for her pains. Next time, the two surviving starting PCs and their new comrades can aim to get further in and deeper down. 

All in all, the game went as I'd hoped. After the higher-level escapades of the regular campaign (approaching its first anniversary this weekend after what must be around 340 sessions), it was nice to be cause fear with a single orc arrow out of the darkness. And I was happy with the lethality level (three survivors out of a total of seven adventurers, starting and recruited during the session). 

What I liked most was the feeling that the adventure was a raid on a dangerous locale - and a fairly unrewarding one at that, especially when the recovered Carla claimed her share. That sets the stage for a second expedition next time - and perhaps this time, the players will recruit all the NPCs they can!


Tuesday, 2 March 2021

Hooooo! A kitbashed 1/72 troll

 


Progress on my megadungeon project has stalled in recent weeks (thanks to that toad), but I've been beavering away on miniatures for the game. Some of these will be used in our regular D&D campaign, which will be switching to 1/72 scale when more epic encounters demand it.




This fellow is a troll of the Poul Anderson/D&D sort, quickly kitbashed from a Reaper Bones ghast and a spare Otherworld troll head. He'll be one of the bigger sorts; I have others underway kitbashed from Games Workshop ghouls. There are plenty of trolls around that work in 1/72 scale, but the D&D/Anderson sort should be armed only with their teeth and claws, which is more of a challenge.



Earlier, I made a couple of smaller ones out of GW ur-ghuls and gnoblar/goblin heads. These are less canonical in appearance, but I want plenty of variety among my trolls - not least so I can tell them apart on the tabletop when tracking hit points and regeneration.


The important thing is that they're unarmed. I'm planning to create reusable sheets with troll stats on them, assigning those with higher hit points to the bigger ones and noting special features (extra arms, heads, etc.) as I go. The idea is that I can print out the troll sheet whenever there's a troll encounter and have extra details (names, quirks of personality, possessions, etc.) and specific hit points ready to go. If the party kill a troll whose name they know, that stat block will be renamed on the sheet. It's a simple trick, but it'll make tracking combat easier and lend the encounters more character and precision from the get-go. 

I'll do the same for all the major monster types in the megadungeon - so that the two orcs you meet wandering down the corridor aren't just faceless 1HD drones but Sharguz and Bagtuth, two cowardly grifters on the fringers of the tribe with some information they think the PCs might be willing to pay for ...




Monday, 4 January 2021

Megadungeon mechanics part five: the joys of one hit die and the implications of what has gone before


I mentioned in the previous post in this series that my forthcoming megadungeon campaign will be more like a wargame than our main D&D campaign. That could be  misleading, though, as I mean it in only one very specific sense: that combats will mostly be of the one-hit variety. This is going to be - for most players - a low-level campaign. A character that gets to level two will be cherished. And one that gets to level three will be prized. 

In part, this limited advancement will come down to player decisions. As I've noted before, I'm using gold for XP - but only if that gold is spent on unspecified advancement-related stuff: training, study, sacrifices, etc. I'm also going to make equipment - notably armour - much more expensive than it is in the common iterations of D&D. I want a mail hauberk to be a significant treasure item in its own right. Full plate harness might cost as much as levelling up. And then there's the process of hiring mercenaries that I've set out: the common way to do this is to have them fight for an equal (or double or triple) share of the expedition's loot, which will cut down the amount of gold available to PC. When they can be hired for a day-rate, that day-rate will be high. All of this means that the players will be constantly choosing between advancement, equipment and assistance. 

It also means that most PCs will be on a solitary hit die for much of the time. I'm toying with the idea of having PCs roll their hit dice for each expedition, although that might prompt a mutiny. But as monsters will be doing fighter weapon damage (d8/d10) for the most part, even second-level fighters will be vulnerable to a lucky stroke from a hobgoblin poleaxe (d10).

So characters will be more like Balin (taken out by one orc-arrow) than Boromir (pin-cushioned by countless arrows and still fighting to the last). This has three main implications.

1. Big battles can be handled easily. If the players hire 20 mercenaries to raid a goblin lair, we might conceivably have 50 or more miniatures on the table. But if we know that most of the goblins will be killed in a single attack and that many of the mercenaries (and PCs) will be too, we can look forward to an initial frenzy of d20 rolls and then a rapid clearing of the battlefield. Mass combat won't involve much more book-keeping than a typical skirmish wargame, and the deadliness should keep it exciting. I expect a lot of traditional PC activity (sneaking, stealing, trickery) to take place during large-scale combats as well as as an alternative to them.

2. Shields shan't be splintered. Because I want highly dangerous encounters, I won't be using Trollsmyth's excellent "shields shall be splintered" rule. We use that in our regular D&D campaign, and it's great. But in the megadungeon game of fully equipped expeditions, it will just lead to mules being burdened with umpteen spare shields. What I'll probably do to compensate is have larger shields modify AC by 2 or 3 rather than 1, which will be reserved for bucklers. So a tower/kite/buckler hierarchy of 3/2/1 modifiers is probably the way to go - although I'll be thinking about this a bit more. 

3. Tactical retreats will be common. When running our main D&D group through the Caves of Chaos last year, I though that one of the most satisfying sessions was when a raid on the hobgoblin caves utterly failed. It ended with some PCs badly wounded and the hobgoblins hunting for them with hounds in the forest. And of course, the PCs swore revenge - which was all the more satisfying when it was eventually achieved. If your characters have only one hit die apiece, you're going to be very sensitive to the turning of the battle's tide. And that leads on to the next point.

4. Game sessions may involve multiple expeditions. I really like this concept: that three or four hours of play might not be confined to a single raid, but could feature two or three. In running our 250+-session lockdown campaign, I've observed that the joy of the game comes as much from the story that emerges between the sessions as from the sessions themselves. I want to reproduce that even in an episodic and occasional campaign, so that grudges can form, vengeance can be taken and best-laid plans can gang aft aglae. And the best way of doing that is to have more story in a single session. For that reason, downtime will be almost instant. Once the PCs are out of the dungeon, we move straight on to preparing the next expedition: "Two months pass, and your wounds are healed. Assemble your new party! What's the plan now?". As the megadungeon will have many entrances, there should always be a different approach worth trying - whether that's launching a diplomatic mission to attempt to recruit another faction to your current cause, attacking from a different direction or hiring a squadron of stone-masons to block up some of the entrances of the dungeon altogether.

With the lights out, it's more dangerous ...

I've also been considering some other implications of my previous posts. Most obviously, the continuity conceit of the last lantern-bearer won't work if many of the PCs have infravision. So perhaps all PCs should be humans. That frees up dwarfs and elves to be villains, which is just as it should be

Then there's the presence of pack animals. I want these to be essential for expeditions to the megadungeon, for realism (food, blankets, water), to simplify encumbrance considerations and to allow for the removal of large items of treasure. They do, of course, make the party more visible, noisy and exposed - but I think that can be turned to the party's - and the game's - advantage too. If you're pursued by angry dire wolves, you might think about hobbling your mules and abandoning them to their awful fate so that you can get away. That's one option that the last lantern-bearer often has - and it's one reason that he always gets away.

The danger to mules also means that it's not only monsters that move around the dungeon, but treasure too. That provides a great boost to emergent narrative, I think. If the dire wolves dragged down your mules outside the hobgoblin stronghold, what has happened to the golden idol you looted from the troglodytes? The hobgoblins probably have it - but what have they done with it? Could you somehow play the vengeful trogs off against them? Or have they returned it to the trogs and forged an alliance with them as a result? Does that mean that their lair will now be guarded by chameleon-like stink lizards that you won't see until they're upon you? Or has a drooling ogre rescued the statue from the dire wolves' leavings and dragged it to his cave where he's now attempting to woo it by piling severed heads in front of it each day? 

That's for the players to find out ...



Sunday, 3 January 2021

Megadungeon mechanics part four: a universal table of wandering monsters

I started writing this a few months back, in response to some interesting posts on wandering monsters from Lich Van Wrinkle and noisms. As so often, I failed to finish the post off - but I'm now incorporating its substance into my series on the megadungeon.

In their respective posts, Lich describes mixed results in his first experience of using wandering monsters 'properly' while noisms takes against them somewhat, especially in the dungeon:

In the artificial and enclosed 'dungeon' environment, though, I've always thought that something stinks about random encounters. Unless the encounter is with a being that already exists within the dungeon key, and is assumed to be moving around (that is, if the encounter is with a being that is extraneous to what is already plotted), then one is forced to simply put out of one's mind the question of where it came from. Why is this giant slug, which the random encounter table just threw up, suddenly here? Where was it before? And why is it that it it does not appear to have had a material effect on its dungeon surroundings prior to this point? 

Like Lich, I didn't use wandering monsters much or at all in my first incarnation as a GM. In part, that was because I started out with RuneQuest, which lacked the concept of the procedural dungeon crawl. And in part, it was because my adventures tended to be sketchy affairs in which a complete map was a bonus and there was never enough prep time for extras such as wandering-monster tables. 

That's not to say that I didn't use dynamic encounters - but they tended to be of the "some guards come round the corner" variety when things started to lag. 

I've still got plenty of time for that approach. But since I've been using Basic D&D for the nine months and running of our daily lockdown campaign, I've used wandering monsters a lot, and I like them. So, I think, do the players. Wandering monsters formed the basis of whole sessions on the Isle of Dread; as noisms notes, "the wilderness is all random encounters". 

A good table of fauna offers lots of instant adventure, especially when two or more rolls are used together. One highlight in our campaign involved the PCs glimpsing a roc on the wing from atop the Isle of Dread's volcano; when they descended onto the plateau and encountered a mastodon, I had the big bird swoop down and snatch it - much to the players' horror and delight.

But what about the dungeon? Wandering monsters are the quintessence of the dungeon crawl, I think, because they communicate to the players that dungeon-crawling is a dangerous business and that they should be quick, quiet and careful. And they also bring the setting to life. 

So where do these monsters come from? Noisms argues that the "only reasonable answer as to where randomly encountered monsters in the dungeon come from is: outside, or from further down". I don't think that's quite right - or at least, I think that these and other answers are easy to rationalise even in an "naturalistic" dungeon.

Now, I agree that wandering monsters should fit the environment. Each dungeon - and probably each dungeon level - should have its own tailored table. But I'm thinking about universally applicable categories that can be intuitively filled in for a given area of the megadungeon. Let's consider eight of these.

Patrols

This is the easiest and most obvious sort: patrols of humanoids (or similar) that live in the dungeon. You can divide these into 'home' and 'hostile'. So, if a dungeon level has mutually antagonistic orcs and gnolls in it, a home roll in orc territory brings orcs while a hostile roll brings gnolls - and vice versa. How do the hostiles get there? Most likely by following the PCs, hoping to exploit the carnage that they've created - and perhaps to pick off the depleted adventurers. Rationalising their route should be easy, presuming that the PCs have come through no-man's land to get to the orc section. The gnolls have simply followed in their (bloody) footsteps.

Stalkers

While patrols probably want to prey on weakened foes and parties alike, other creatures might be specifically interested in the party. They might be predatory, or they might be looking for security, company or even friendship. So, this category can include scavengers hungry for fresh corpses, beings that want protection in the dungeon (and what better escort than a bloody-handed bunch of murder-hobos?) or those interested in something the party has: past owners of treasure, perhaps, or their minions. If Moria is the ur-dungeon, then Gollum is the ur-stalker. 

Vermin (a.k.a dwellers in the wainscotting or the Wolves in the Walls)

One obvious way to have dungeon vermin suddenly appear is simply to have them come out of the walls. I've often used kobolds in this way: their tunnels riddle the dungeon but are too small for (most) adventurers to get down - and too dangerous even if they could. Yes, your halfling might squeeze into a kobold crawlspace, but the excursion probably wouldn't end well. 

This gives a nice rationale for why weak creatures like kobolds survive and even thrive in the underworld. Rather than inhabiting large and easily invaded caverns, they live in twisting tunnels; their communal spaces are almost impossible to access. And it also explains why they can crop up at almost any time or place in the dungeon. They tunnel and burrow constantly, the entrances to their crawlways are cunningly concealed, and they're always creating new ones. 

Kobolds are just one example of creatures in this class, of course. Giant rats, giant beetles and many other smallish, burrowing beasts fit here nicely, as do stirges that roost in the recesses of the dungeon but fly through its corridors when on the hunt. 

Squeezers (through cracks)

Monstrous oozes and jellies are pretty much a D&D invention (drawing on shoggoths and the Blob, of course), and they provide another easily explainable source of monsters. Where did that thing come from? It dripped through the ceiling. Oozes gonna ooze.

Raiders

Now, these are typically from the outside (or above or below). I'd differentiate these from patrols, above, in that they're not on regular duty or simply being opportunistic, but have mounted an attack against the area of the dungeon in which they're encountered. So, while these could be the orcs/gnolls from across no-man's land on the same level, they'd appear in far greater numbers than a patrol. But they could be hobgoblins from a nearby stronghold or bugbears from the surrounding hills. They might signal their attack with drums or horns, and the party might initially just encounter their scouts or vanguard - giving the PCs the opportunity to make something of the situation. Repeated rolls in a single session would typically indicate encounters with different elements of the raiding party. Raiders could also be other adventurers, of course.

Rampagers

Another dangerous category, this: big monsters on the hunt, whether from nearby caverns or from deep within the earth. This might include trolls, landsharks and umber hulks. If these creatures exist on the dungeon level, their coming will prompt other inhabitants to bar their doors and retreat to their fastnesses. And it could be signalled by howls, heavy footsteps or trembling in the earth.

Haunters

Given that most dungeons have witnessed many scenes of slaughter, you'd expect a few ghosts to be clanking their chains. Different bestiaries assign specific roles to synonyms for ghost, but we can usefully divide this class into apparitions: harmless but disturbing visions that merely show something that happened in the past - from a march of doomed troops to a murder; unquiet spirits: ghosts that want the party to do something for them - bury their bones, avenge their death or return some object to their grave; and evil spirits: hostile ethereal undead of the D&D spectre or wraith sort.

Apparitions are perhaps the most interesting class. In a science-fantasy setting, they might be fragmented holograms (as in Serenity or Prometheus). In any setting, they can be used to hint at some tragedy from the past. Perhaps the PCs see grim-faced warriors in bronze armour march down the corridor - and then encounter animated skeletons wearing that same armour elsewhere. And perhaps these apparitions aren't visible to everyone. You can have the whole party make a save vs spells; if everyone passes, they'll still be spooked.

Personalities

An easy way to bring the dungeon to life is to ensure that its most celebrated inhabitants don't simply sit in their room waiting for adventurers to arrive. There are two obvious categories here - the 'home' encounter, which might be the orc chief visiting his lieutenants or interrogating prisoners; and the rarer 'hostile' encounter, in which an eminence from another part of the dungeon has arrived. A 'hostile' result could actually be a welcomed guest - what matters is that it's someone or something who's not on their home turf. Previous random encounters might give clues to this. If you've had a 'hostile' raiders encounter previously, a subsequent personality might be the hobgoblin commander striding in like Darth Vader in the Hoth base. If you haven't, he might be here with his bodyguards to plot mischief in alliance with the orc chief. 

So, those eight categories might give us a table like this:

D8                    Category                        Encounter (D4)

1                        Patrols                           1-2. Home (2d3)   3. Hostile (2d4)   4. Outsiders (2d6)

2                        Stalkers                         1-2. Ghouls    3. Protection-seekers 4. Treasure-seekers

3                        Vermin                          1-2. Kobolds (2d6)  3. Giant rats (2d6)   4. Stirges (2d8)

4                        Squeezers                      1. Gelatinous cube (1)  2-3 Black pudding (1-2) 4 Grey ooze (1-3)

5                        Raiders                          1 Home (4d6)  2. Hostile (4d6)  3. Outsiders (4d6) 4. NPCs (2d4)

6                        Rampagers                    1. Umber hulk  2. Trolls (d3) 3. Landshark 4. Purple worm

7                        Haunters                        1-2. Apparitions 3. Unquiet spirits 4. Evil spirits

8.                       Personalities                   1-3. Home        4. Hostile

Over a session or two, rolls on this table should illustrate the dungeon's history, politics and ecology. 

And while they should challenge the players, they should also provide them with opportunities. A hostile patrol offers the prospect of a temporary alliance. A treasure-seeker might obtain its goal in return for local knowledge. If a 'home raid' takes 4d6 of the bugbear warriors out of the complex, PCs who have snuffed their torches and concealed their pack animals may have the chance to fall upon a lightly defended bugbear lair. If the hobgoblin chief is talking to the orcs, then he and his bodyguards won't be at home if the PCs come knocking. 

Saturday, 2 January 2021

A third round of megadungeon mechanics: mercenaries

 

I've been thinking a bit more about how adventuring parties will be structured in my episodic megadungeon campaign (see here and here for previous thoughts). 

In some ways, this campaign is going to be much more of a wargame than most of the games I run, though there will be ample opportunities for roleplaying. But the game will take the form of successive armed raids on a (largely) subterranean stronghold of great size, and so tactics and organisation will play a large part. 

Each party will consist of the PCs plus an equal number of non-combatant lantern-bearers (torch-bearers, porters, etc.) and the same number of pack animals (mules, donkeys, camels, whatever). These are 'free' in that they're assumed to have been hired or bought before the game starts. So a party of three adventurers will be accompanied by three lantern-bearers and three pack animals. The fixed ratio makes it simple, and there's no 'on-screen' cost to the players. 

A party with just three combatants will represent easy pickings for monsters, however. So PCs would be advised to bolster their numbers with mercenaries. This decision is down to the party's leader - the PC with the highest CHA - but individual PCs can also choose to hire their own bodyguards.

So, how does hiring work? Well, it's simple: hired fighters get a share of the loot: an equal share (with certain exceptions - see below). 

That introduces a nice risk/reward consideration from the get-go. Do you hire 10 archers to rain arrows on the orcs on the first level? It might work well - but then your party of three will have to split any loot 13 ways, assuming that everyone survives. That assumption might not be safe, of course, but most parties will balk at getting such a small fraction of any takings.

Now, if the party's leader decides not to hire mercenaries, party members can still opt to split their shares with hired hands. So, a cautious dwarf might not like the idea of going in with just three fighting adventurers and might opt to split his share with three of his kin. Those dwarves will answer to him, of course.

There's also potential to play on the idea of the Doppelsöldner and allow the PCs to hire higher-level fighters for a higher share of the loot. Dungeons & Dragon's level system provides an obvious way to do this. But there will be scarcity limits on high-level fighters: no more than one third-level or two second-level hirelings per PC. So, if our three-strong first-level party hires three third-level fighters, the loot will be divided into 12 shares, with each of the hirelings taking three of those and the PCs getting only a quarter of the spoils between them.

Non-divisible treasures (magical swords, amulets, etc.) will be chosen by combatants in turn, starting with the highest-level character, whether PC or NPC, and then rolling dice for priority. So if our first-level party hires three third-level fighters and then finds the Drinker in the dungeon, one of those hirelings is likely to make off with it (which might be just as well).

Note that hirelings are not potential PCs in the way that lantern-bearers are. If your character gets killed, you can roll up a new character and embody him or her in one of the lantern-bearers, who, when offered the gear of the slain adventurer, has just acquired a taste for adventure. But hired fighters don't become PCs in the same way. They can, of course, be run temporarily by players who've lost their characters if the party is down to its last lantern-bearer.

This means that players can't leapfrog to higher-level characters after PC deaths. It also means that hiring high-level fighters is always costly unless they're slain on the job. Your party doesn't get to keep the Drinker because you took over Ulav the Fearless mid-game; instead, you roll up a new PC when the expedition ends.

That's the main system for hiring fighters. But there will be an alternative available once the party has amassed a bit of cash. They will be able to hire fighters for a day rate,  paid in advance. This rate will be much higher than that in the Rules Cyclopedia, so that it forces the players to make tough choices about how they spend their loot: hire help, progress in experience or buy better equipment. And fighters hired in this way will also be subject to the classic Rules Cyclopedia morale rules. They've already been paid, so if things go badly, they're more likely to break and run. After all, if you think that the orcs are about to finish off the party, you're not going to be too concerned about breaking your contract with a corpse. 

I also like the idea of unscrupulous PCs hiring equally unscrupulous orcs. As in Rules Cyclopedia, this will be significantly cheaper than hiring human fighters. But it will also be riskier. First, orcs will be morale 6 rather than the morale 8 for human mercenaries. Second, their loyalty will be vulnerable to reaction rolls in encounters with certain creatures (powerful ones that can speak orcish, such as ogres). And third, you always have to worry about situations in which your orcish hirelings outnumber you and are not afraid of you. I might, though, allow orcs to fight for half-shares under the share system, in which case they would still be subject to morale rolls (unlike human mercenaries fighting for a share)  but with morale 8 rather than 6 (i.e. the share incentive acts as the equivalent of a strong leader as described in the Rules Cyclopedia).

The net effect of all this is to make orcish hirelings a high-risk proposition that could be tempting in certain circumstances. If you're going to storm the hobgoblin stronghold, a few dozen orcish mercenaries could be a cheap way to achieve overwhelming numbers. But you don't want things to go too badly for them (prompting morale rolls) - or too well (so that there are lots of them inside and in control of the stronghold before the party establishes control). 



Scratch-built brain collector

 


The party stumbled across a brain collector in the depths of Castle Amber as our last session of 2020 came to a close. For the first session of 2021, I knocked this fellow up out of Fimo, tin foil and mince-pie trays.




Friday, 1 January 2021

More megadungeon mechanisms

Happy New Year!

After yesterday's post on the last lantern-bearer (the one who always gets away), I've been thinking of a few more mechanisms for the episodic megadungeon campaign I plan to run this year. These are mostly fairly arbitrary-seeming rules that will - I hope - add tension, excitement and a bit of grit to the game.

I'll probably use some variant of the original D&D rules for this (possibly Whitehack). I want to strip things right back, so that hit points are low and deadliness is high. In our regular D&D campaign, I use STR in the Into the Odd style: with characters losing STR points after their hit points are gone and having to make a STR save each time they take STR losses. That works really well - but for this campaign, I might go for maximum deadliness, so that starting PCs may well just have 1 or 2 hit points each. That should amp up the risks nicely.

So here are some of these mechanisms.

1. Adventuring parties are always accompanied by an equal number of non-combatant porters/lantern-bearers and the same number of pack animals. There's no point going on a treasure-seeking expedition to the Underworld if you don't have the means of carrying off the loot. And - in meta-game terms - that gives us the pool of reserve PCs, the encumbrance and lighting solution, and the continuity mechanism that I outlined yesterday. The pack animals also carry plenty of food, water and fuel.

2. Party casualties are always replaced from the lantern-bearer pool, with a new character rolled up on the spot. Such replacements can be of any class (hitherto hidden talents) and are equipped from what can be salvaged from the dead or provided by the rest of the party. The pack animals are likely to carry a few spare spears and bow. I'll probably have hand-to-hand weapons do damage by class rather than type (fighters d8 or d10 for two-handed; clerics, elves, dwarves and thieves d6 or d8; magic-users d4 or d6), to avoid any class restrictions on weapons.

3. Expeditions always end when the session ends. This isn't an original idea, of course, but I think it'll provide a nice rhythm and aid continuity, especially as the players aren't likely to be exactly the same each time. I might provide some narrative cover for this ("You need to get out by nightfall - the place is crawling with monsters after dark"), or I might just leave it entirely arbitrary.

4. Escaping lantern-bearers don't bring party treasures or equipment with them. They are assumed to have fled as fast as their legs can carry them. But they do know where the loot was last seen - and are able to guide a new party back to that point. It's unlikely, of course, that the loot will still be lying around. But if Baldros the Bold and his men met their end at the hands of the Iron Hand hobgoblins, it's a fair bet that the hobgoblins now have the golden idol that Baldros stole from the troglodyte tribe.

5. Gold provides XP - but only when it's spent (on training, better equipment, magical research, etc.) Living costs are included in this it's assumed that the PCs pay for their upkeep during the course of their training or studies. I'll probably seriously revise whatever armour and weapon list I use so that better weapons (two-handed swords, etc.) and - especially - better armour is seriously expensive. That way, there should be a nice dilemma for expedition survivors: better gear or character advancement? And of course, this mechanism keeps the focus squarely on loot - which incentivises dungeon-crawling.

6. All parties have a nominal leader - the PC with the highest CHA. This is largely for "historical" atmosphere in the dungeon: "That's where they got Captain Juras and his lot. That's him there - turned to stone!" or "See the tarred head on the spike up there? That's Chardro. 'Prince of Thieves,' he called himself. But the orcs are no respecters of royalty". The leader also has total control of the lantern-bearers/porters and pack animals, however. They go where he tells them to (except when they flee, of course), which should help to keep the party together - especially in light of the following.

7. There is safety in numbers. In the dungeon, large groups are much safer than small ones. You might think that a couple of thieves could sneak off on their own, relying on stealth and guile to get in and out with choice items of loot. But the dungeon ecology is against you. Wandering monsters - dangerous vermin of all sorts, including kobolds (or equivalent) - are always ready to pick off the isolated. And these creatures can see in the dark and are ever watchful from the safety of their burrows. Stragglers or break-off groups incite regular wandering-monster checks - and those monsters are numerous and emboldened. This makes plenty of sense: small creatures like kobolds are unlikely to risk frontal attacks on large, well-equipped groups (the kobolds don't know that the lantern-bearers won't fight). But ambushing a couple of thieves in the darkness? It's a near certainty. 

The point of all of this is to create a game in which the dungeon, rather than the party, is the focus, and in which 'victories' over the dungeon are to be prized. A total-party kill should come as an exhilarating last stand - and a successful expedition should feel like a real achievement. So too should taking a character up to second or third level - gaining the chance of surviving a hit or two in combat.

I'll be running this with 1/72 miniatures, which offers more space for all those pack animals and lantern-bearers on the table - and especially over Zoom. That scale is also somewhat "depersonalised", with the miniatures being more generic and interchangeable - which I see as a good thing for RPGs. 

The key, though, will be creating a megadungeon with plenty of interesting features and perils beside its inhabitants. I started work on this last year, and I hope to make it a major focus in 2021.