The Lost Dungeon Crawls of 2nd Edition

2nd Ed AD&D gets shat on a lot for being too railroady, overly story-focused, full of broken options, and a total departure from earlier versions of the game. I really don't think that's true. First, the overarching narrative plot can be found in many of the classic older D&D modules. Second, here's a list of modules as evidence.

Unlike some of the other 2nd ed adventures that I like (Swamplight, Tales of Enchantment, Fighter's Challenge) these are pretty much for a dungeon crawl style of play. Yes, they do include some extra elements of backstory that you won't find in earlier modules. Are they all quality? I've only run part of one of these. But I'd love to add Shattered Circle to my repertoire of beginner adventures, and both Dragon Mountain and Axe of the Dwarvish Lords look pretty epic. Firestorm Peak does as well, but the dungeon didn't seem as big/loopy at first glance. Baba Yaga's hut has been redone a few times but could also be exciting, just haven't been running things in the right setting lately.

Dragon Mountain (1993)
Temple, Tower, & Tomb (1994)
Dancing Hut of Baba Yaga (1995)
Labyrinth of Madness (1995)
Undermountain: The Lost Level (1996, Dungeon Crawl series)
Undermountain: Maddgoth’s Castle (1996, Dungeon Crawl series)
The Gates of Firestorm Peak (1996)
Undermountain: Stardock (1996, Dungeon Crawl series)
Hellgate Keep (1998, Dungeon Crawl series)
Lost Shrine of Bundushatur (1998, Dungeon Crawl series)
Dungeon of Death (2000, Dungeon Crawl series)
The Shattered Circle (1999)
Axe of the Dwarvish Lords (1999, Tomes series)

This list doesn't even mention the rehashes:

House of Strahd (1993)
Return to the Tomb of Horrors (1998)
Return to the Keep on the Borderlands (1999)
Return to White Plume Mountain (1999)
Against the Giants: The Liberation of Geoff (1999)

Basically, when you look at the old B/X, BECMI, and 1st ed modules you see a few mentioned over and over as either classics actually good: a lot of others ignored. I think some of these modules I listed above get ignored just because they're second edition, and probably deserve a closer look (nothing from 2nd ed showed up in Tales from the Yawning Portal!). This doesn't mean that some of these might be stinkers: I'm looking at you, Stardock! Just that perhaps some of the baby has been thrown out with the bathwater of 2nd Edition, and while the 90s are known for heavy-handed plot and adventures on rails, there was a good amount of the old-timey dungeon crawl as well.

5e Sorcerer Variant?

One of my beefs with 5e is the uneven distribution of long rest and short rest abilities among classes (i.e. encounter and daily powers). The biggest offender, imho, is the sorcerer who gets only long-rest resources and is encouraged to use them up quicker because the sorcerer can spend more resources for a nova round. The quick brainstorm with a friend is to make sorcery points short-rest resources somehow. There's two issues with it.

Number of Sorcery Points. First, the number. 5e sort of assumes something like 2-4 short rests per long rest, so if we simply divide the sorcery points between half and a quarter, that should come close. Say half for now. That'd mean a level 5 sorcerer gets 3 points, which equates to one heightened spell per short rest or a bonus 2nd level slot. That seems pretty strong since warlocks get two 3rd-level slots per short rest, and the sorcerer would also have 3rd level daily slots. If it's a quarter, that means the level 5 caster has 2 points per short rest, which prohibits most expensive metamagic on spells other than cantrips and first level spells, but one level 1 slot per short-rest seems about right but weak. A third is probably about right.

Creating spell slots. The second issue is sorcery points being used to make new spell slots. Short rests could essentially be used to recharge, meaning why not take 3 short rests in a row to recover more slots? This is easily solved by capping created slots with a duration. I'd say end of next round, so you could create a slot in preparation for using it the next round. 1 minute might also be reasonable, or even 10 minutes though. But since they're easy to create, there's not a lot of reason to make them so long-lasting. And if/when they expire? They could automatically convert back to spell points. There's a slight loss here, as it's more expensive to create a spell slot (5 sorc points to make a level 3 spell) than you get from spell slots (3 sorc points for a level 3 slot). But that's probably fine.

Level 20. The sorcerer capstone ability is finally the short-rest sorcery points we want. Nice, but super late. They'd need a new capstone.

Done. What other consequences would this have? Some metamagics aren't going to work on your big guns. You won't have enough points to heighten, twin, or whatever your highest level spells. There's also a question of limitations, can you have more sorcery points that your max? I say sure, convert one big slot to points to power metamagic on your other big slots. Finally, it limits some metamagic options flat out, like Heighten Spell. We might be able to address this by either explicitly mentioning the option is best when you have more sorc points, or by fiddling with the number a bit, more like 1/4 plus 1 or half minus 1 something to ensure the number of sorc points at each level is about right. This still feels a little clunky, but I like the idea. Combine it with 1-2 bonus low-level utility spells for each origin, and it might do well.

The lost perception skills of older D&D

Older D&D was weird. Though I don't think I ever used the rules in the 90s, dwarves and gnomes (and some other races) have these strange detection abilities. And they're subtly different between the races. I'm going to focus on dwarves, elves, and gnomes here because the others just cherry pick some of these detection powers.

In AD&D 2nd Ed, dwarves can spend one round to detect things if they're within 10 feet of that thing (except depth underground, since you're always within 10 feet of your depth I guess): sliding and shifting walls or rooms (4 in 6), grade or slope (5 in 6), stone traps/pits/deadfalls (3 in 6), new construction (3 in 6) as well as depth underground (3 in 6). Gnomes can detect depth better than dwarves (4 in 6), grade/slope (5 in 6), but also detect direction underground (3 in 6) as well as unsafe walls and ceilings (7 in 10).

That's right, while 2nd Ed raised a few of these detections so they were all using a d6 (dwarves went from grade/slope and new construction of 3 in 4—75%—to 5 in 6—83.3%) except the gnome's 70% chance to detect unsafe walls and ceilings couldn't be lowered to 4 in 6 or raised to 5 in 6 to make it standard. Stout halflings and half orcs likewise ended up using a d4 for their detect grade/slope abilities (3 in 4 and 1 in 4 respectively), and later races in Skills & Powers just get percentile scores assigned. So while the system mostly uses a d6, it doesn't use it consistently.

In the basic Rules Cyclopedia D&D (which only has dwarves), dwarves get all of these stone skills at a 2 in 6 chance, whereas Swords and Wizardry (my ersatz oldest-school D&D) says dwarves get this but the rolls are up to the DM.

This is fascinating in part because these skills were/are probably almost wholly ignored by most D&D players, but also how specific they were to dungeon crawling. Knowing how deep you were (approximate dungeon level) would be hugely helpful in figuring out how tough monsters might be, as well as if the 5% grade in the 100-foot corridor (nigh imperceptible) was taking you further down or not. New construction seems like an ersatz measure of things being walled off, but it also gives you a sense of history within the dungeon, while shifting walls, traps/pits, and unsafe walls/ceilings are obvious hazards (of apparently differing difficulty). Also interesting to note is that dwarves and gnomes got slightly different sets, meaning it was useful to have both types of characters in your party. It also conjures some strange image of dwarves keeping their eyes down and watching for pits and new construction at the base of walls while gnomes are eyes-in-the-sky noticing the ceilings of the dungeon plus a different sense of direction.

Elves get a different ability, which is passive secret/concealed door detection. They get this at a 1 in 6 chance, so just by walking through the dungeon elves will find one secret door in 6. Not great, but totally nice if the DM remembers it (I also just roll this for random noticing facts in my AD&D Al-Qadim Church game). Elves who are actively searching find 1 in 3 secret doors, and 1 in 2 concealed doors.

It's also interesting to note that there are no rules for how these interact with thief abilities (does a dwarf searching for a trap get his roll plus the thief roll? does it take extra time to get both rolls?) nor do these detection abilities ever increase with level: you have them or you don't, and they never improve.

In 5e, we get a little of these as ersatz fixes. Dwarves get their stonecunning and elves get free proficiency in perception, but I can't help but wonder if the sheet blinders work in reverse here. Often with skills, new players are wondering which one they can roll for a certain thing to get more info or move the story along, but there's nothing specific enough about those skills in 5e to warrant the utterly specific uses (or expectations?) that the old AD&D options provide.

I've been thinking about this more and more because, in my 2nd edition game, many of the rules are buried in paragraph form, when bullet points would bring them to light easier. The oft-maligned Skills & Powers book does a nice job of presenting each as a clear option since you needed to spend points to purchase each one, but it can also be used to nicely gather each option together so you can really see what powers you have. So in my vague dream to do some cut-and-paste work to compile and solidify some 2nd ed material and make it more accessible at the table, I'm wondering if these strange powers shouldn't be made broader (i.e. dwarves and gnomes might passively notice sloping corridors rather than needing to actively do it) or otherwise wrapped into a coherent system which includes general thief/rogue skills.

There's also a strange element of DM determinism in these skills, which is the information these detection powers provide or the paths these open up need to be a bonus in the adventure. For traps, their detection is generally simple since the failure result is merely that the trap is set. For sliding walls/rooms though, a passageway could be blocked if not detected. Similarly, if a secret door is the only exit and it isn't found, the party is just stuck. Just because this stuff is on a character sheet, doesn't mean it will work when you need it. Also, while there's been a lot made about dungeons haveing interesting design (loops, multiple entrances, etc.) there seems to be little attention paid to some of these other issues here: sliding doors or cave-ins which trap a party and force them to continue onwards, subtle changes in elevation leading to distinct levels (or over and underpasses), the history of a dungeon's construction and determining which areas are original and which are recent editions, natural construction problems and hazards vs consciously-made traps. It's a small laundry list of good elements to put into a dungeon that I hope to keep in mind for the future.

It's easy to see how these old skills got consolidated in third edition into a version of notice and search (along with the rogue skills), what you're noticing or searching for has perhaps been lost a little. Also what's lost is the race-specific nature of these skills. Dwarves and gnomes have an uncanny ability to detect a grade or slope in a corridor, others could attempt it with weights or marbles or water, but these races just are that attuned to their world that they get these automatically and no one else will ever be as good as them. And I think that's one thing that's been a little lost in modern D&D: some races were just better at some things than others, which actually made the race/class choice matter a bit more.

The Evil Eye: Actual Play 5 – Finale

Well, that's that. We just did chapter 6 and finished on-time. So at least I can get some of that pacing right.

I was pretty disappointed by this one. I think it was a combination of the group composition and adventure though. We had three and a half hours today and didn't need any of the extra material I had prepped and that was with cutting down on the hit-and-run from the villains. It didn't help that one of the players was super chatty. I know we lost 20-30 min to leveling up to level 5, but I feel like we lost another 30 minutes at least to tangential stories (I know someone ordered sweet potato fries with dinner, but a couple minute bit on the difference between sweet potatoes and yams... that's what I'm talking about).

Ultimately, it was also disappointing when the two most on-top-of-it players both were lost during the final battle. I know it's designed to be confusing, but the group didn't seem to grab onto the goal of escape and get un-cursed. They seemed lost in the sandbox I gave them as I removed a couple rails in the adventure.

So the adventure itself, I like because it's got some nice NPCs at odds with one another, but that was also the downfall. The players didn't really track who they had met, and there are maybe two too many NPCs. My attempt to identify important NPCs by giving them notable eye colors was seen as a red herring, and having two new-ish players (one brand new, one returning after 20 years from AD&D 2nd Ed) slowed things down. So it was perhaps a bit confusing for them.

There was another problem, namely that some of the classes didn't come with much investigation power built in, which is a bit of a problem with 5e in general: fighters just fight. I'm going to curate my list of class options better if I run another Ravenloft.

Ultimately, while I was a bit disappointed, I think some of it was the set-up (new players not knowing what Ravenloft meant) and group composition (chatty player). The adventure wasn't as solid as I had hoped, but I picked it because I thought it had some good NPCs and the finale of U1-3 would be lackluster and it was close to the right size for 4-5 sessions (which it was perfect for 4-5 sessions). I could potentially run this again, but it's about a 20-hour adventure which is too long for short things and too short for long things, so not sure when I might get to it.

As done is better than perfect, I'll just move on to what comes next. Also, story recap in another post soon.

The Evil Eye: Actual Play 4

A little late on this update. We did a Friday session, but it was pretty slow because the group is fairly chatty and I didn't really push to keep people on task as we were eating and drinking. I'm a little irked by the group composition and might not assemble the same people for another game. We'll see.

As before, no real spoilers. I'll post a big summary at the end.

In session 4, we finished up everything except the final chapter. I cut down chapter 4 as it was a lot of mood and storytelling plus we didn't hit it in the railroad order. I also had to do some impromptu fleshing out of the Midnight Slasher, as they ended up following up on that thread with a clever use of a locate item spell. I also Curse-of-Strahd'ed up the villains in this (making them more interactive), so the party did get to encounter the Dukkar and do a quick combat with him, which was somewhat satisfying (and will be more so for the final battle).

Overall, I'm a bit dissatisfied with this adventure. I really liked it when I first read it, except it was a bit light on combat. Now I feel like it has very little concrete goal for the PCs to achieve. Ravenloft is often a "stop the evil / escape" situation, but the players were definitely floundering a bit in terms of what the heck they're supposed to be doing. The presentation in the adventure is more railroady than the adventure itself is, but even adding in a second path to defeat the Dukkar, the party ended up going mostly by-the-book on this one. Also, there are some really long bits of exposition in the book to flesh out the background that could probably be given in other ways.

Some of my dissatisfaction is maybe with the group, as I mentioned above. I'm just not as excited to run this as I had been with Strahd. Then there's the problem with the melee-only classes. I'm finding that skills alone aren't enough to keep players interested for investigation. I ended up giving the Totem Barbarian the full druid ritual caster feat, so he could swap out the initial two selections and possibly learn a few other spells, I think that's a change I might keep in any other game I run. But honestly, the ranger has some nice features to help with travel, and spellcasters can do a bit with investigation and interaction that others just can't. The Barbarian, Fighter, and Monk just don't get that many skills that are grand, and many of their subclasses are 100% focused on combat. This didn't feel like as much of an issue with Curse of Strahd, but I'm sure some of that was the players and party composition.

I still do like the adventure, and it may fit well into a series of Ravenloft adventures with a Sliders style plot. You could even string it along before Curse of Strahd if you wanted to throw something like Night of the Walking Dead, this, and another one or two before they get to Barovia and ultimately deal with the Amber Temple. I just don't feel like this adventure quite lives up to its potential in terms of the adventure presentation, ease of use, and playability. It's got some great NPCs, and good background on the domain, but just lacks a bit in terms of the structure.

Powerful NPCs destroy verisimilitude

My current DM really likes throwing high-level NPCs at us now. I'm not super happy with this because it really brings us out of the game. Here's how it came about, as far as I can tell:

The DM has decided the CR system is ridiculous since we could easily handle many battles in the past campaign. In the last campaign, we did well with a lot of battles, but he's also remembering the past year of high-level play when we really could handle a lot of stuff. I think the earlier parts of Princes of the Apocalypse (before level 7 or so) were fairly well balanced modulo travel nonsense (see below).

As part of his play style, the DM tends to keep combats small, where I think the math of 5e makes a battle with fewer enemies than the number of players easy. Players can gang up on people, and with a few clever uses of some battlefield control can really keep one to three of the enemies out of combat to focus fire on the half of the group that is still a threat. Still, it's easier to keep track of 3 enemies' stats than 6 enemies' stats.

Another aspect of his play style is he tends to say yes, even to ridiculous requests at times. He was very willing to give surprise rounds and advantage right away, and still is at times. That makes stealth much more valuable, and I'm a little surprised we didn't make an all stealth party this time because of that. We've always got time to take short and long rests, and he doesn't pressure us for time usually, so we can really handle many battles with ease. The short rest mechanic also means many of us get some nice resources back after a rest (particularly when I played a warlock), so you can really handle battle after battle when he makes them easy (or easy for us).

On a related note, overland travel with random encounters is awful in 5e. Not only does our ranger ensure we'll never get lost or need to forage for food while traveling, but since all your resources come back after a long rest, there's really no point to doing these in 5e. I like random encounters to help set the mood or tone of a story, or to siphon off a few resources from the party, but travel in 5e (without the Adventures in Middle Earth fix where you can't get a long rest in the wilderness) makes the resource drain pointless.

Finally, I'm not sure if he nerfs the enemies intentionally or not, but he tends to ignore many aspects of the by-the-book stats. It may be a combo of not reading things close, being tired at the time, or lack of system mastery for the spells (He used crown of madness totally ineffectually tonight and didn't use the enemy's second winds). For the spells, I suspect some of it is by not really playing 5e much, you don't know off the top of your head what most of the spells do. That's easy to mess up, or having an archmage up-cast lightning bolt as a 9th-level spell is just not super useful. So he's not always using his enemies effectively. Some of the theater of the mind aspects of play also might hinder this, there's really not interesting terrain in the game for stunts, and we usually have trouble tracking where enemies are and which one is which (probably once or twice a night at least).

How does this lead to high-level NPCs? Since The DM has thrown out the CR system, we end up having three 6th-level PCs fighting things like two CR9 champions plus a CR6 warlock (this past week's tomfoolery). Or the end of Rise of Tiamat where there were just tons of archmages wandering around. Those archmates were functionally something like 14th level NPCs we just fought, it took everything we had to defeat them, and we got a short rest in the middle because we had been beaten (non-lethal damage on taking down the paladin and fighter/rogue). Obviously, we hunted the survivor down to take our money back, because D&D.

So we're left wondering why the noble was able to hire these very powerful mercenaries to chase us down (The druid made a deal to reincarnate his dead nephew for a lot of cash, but he wasn't going to pay us for it when the mother wasn't 100% convinced it was her son). So why are we the ones going to deal with the giant problems when there are these high-level mercenaries able to be hired?

In the last campaign, where he started out using obviously 20th-level NPCs as patrons for each faction, we also were asking ourselves multiple times why we were being sent to investigate things while the big powerful spellcasters (able to cast wish spells and the like) were doing nothing themselves. I have no explanation for that.

I think this is often a criticism of some of the published D&D worlds: when there are these massively powerful NPCs walking around (Elminster, the Seven Sisters, etc.) why are the PCs really needed to be heroes? It's now exacerbated by the inclusion of these NPC stat blocks in the Monster Manual and Volo's Guide to Monsters, and incongruent with the idea of tiers of play.

In by-the-book 5e, characters of levels 1-4 (Apprentice) are local heroes, levels 5-10 are regional heroes, levels 11-16 (paragon) are global heroes and levels 17-20 (Epic) are cosmic heroes. This means that those CR 6 NPCs that function as 14th level spellcasters are really high up there, and the ones beyond that are just basically god-like. In theory, the CR for one creature is equal to average party level for a 4-5 person party. So that CR 6 warlock is an average encounter for a 6th level party. Give the warlock some support henches and it's a good time. But what's the level of that CR6 warlock? The spell-casting says the warlock is 14th level. Clearly paragon. We could instead try to break monsters down into their tiers, but a proficiency bonus diagnostic doesn't quite map to the tier levels exactly. Nonetheless, it may be a decent stand-in: things with a proficiency bonus of +3 are regional-level threats. +4 spans the end of regional and beginning of global. +5 is solidly global, with +6 being epic. That might not work out properly though because the CR 6 warlock was a 14th level caster with a +3 proficiency bonus. I'm sure with some more time and math I could reverse-engineer something somewhat close though.

What it really means, is by using these high CR NPCs, the DM is telling us there are many powerful people in the world that could be doing our job instead of us. So we need something more of a real hook for the game, especially since the others poo-poo'd my idea of all being good-aligned.

What's the fix? I guess telling the DM. He should know that the one warlock with a couple veterans (CR 3) is much more believable, and something like four veterans, a gladiator, and the warlock might have been a much more believable set of enemies to face. The mage might have an apprentice and two hired bodyguards rather than being an archmage, or might have a bound elemental or demon servitor to defend him. It's difficult to point out NPC powers that I don't know about, but I do try to point out some of the stuff most of the time if it'd make an easy fight harder. I could also try to point out that doing 4 easier combats will drain resources so it's useful to utilize that in the right context, and that random encounters on the road in 5e just serve to bog the game down.

The Evil Eye: Actual Play 3

Session three. One of the players was absent, and we got a new novice to join. We mostly finished chapter 3, skipped 4, and are halfway through 5. So the pacing seems about fine that the game can end after chapter 6 or perhaps with a custom 6.5 in the next two sessions.

This session was totally investigation again. Perhaps a flaw with the adventure (depending on your point of view). I was resigned to keep it a little slow because we had a new player who had never played before, but she caught on pretty quickly. I also wish I had not told players we'd advance in level. Curse of Strahd was designed for 3-9 basically, and this isn't really designed to level up midway through. I had them hit level 4 now, and they'll hit 5 at the beginning of the final session I think.

I feel like I've done a bit better at NPC voices, but I find I need to give them explicit labels. The voice isn't a description (like deep voice, British accent) but a stereotype (1920s gangster). That's helped me give the characters a bit of personality.

I had to do a bit of improvisation with the House of Mists, just because repeating the encounter seemed ridiculously boring at the time. So I drew random cards and tried to deliver some NPC hints based on the card.

I also did some improvisation by deciding to use theme/mood encounters in addition to random interactive encounters. This means I'd totally want to revise my random encounter tables, but it also means I can have them spot an NPC in the crowd or find a body in a dark alley, encounter the darklord's spies, or whatnot. Basically a lot of the old Ravenloft material suggests no random encounters, and I understand that. They generally give railroady set-piece encounters. But you can deliver some of those here and there, and curate your randomness. Basically, I want to revise the tables I made so the generic encounters that can happen anywhere are mostly theme/mood encounters rather than possible combats. Or perhaps so that there's a separate chance for theme/mood builders than possible combats and strong clues or whatnot.

I've also noticed that I wish the text were re-organized a bit more to be encounter-specific. The encounters with Malocchio, Matton, and Gabrielle plus the House of Mists in Karina are really the heart of chapters 3-5, so it was really good that I re-read it today and organize things in my mind if not in my notes. I need to re-organize my notes if I want to put some notes on the DM's guild.

I realize these posts are pretty details-light, I should do something more plot-oriented at the end.