2020-01-13

Subterrane Surveys: Gygax's Castle Greyhawk

We're incredibly fortunate that on August 16th, 2007, Matt "Eridanis" Bogen managed to take a few photos of Gary Gygax running his Castle Greyhawk adventure (for ENWorld moderators at Gen Con), catching glimpses of at least two maps and a key. The very fuzzy first-level key was mostly decoded with a great deal of effort from people like Allan Grohe (grodog), Zenopus Archives, and others. The evidence to my eye seems to be that was likely the true, original map of Castle Greyhawk (e.g., he wrote on the Pied Pier boards in 2004: "Fact is that I have run OD&D games every year at several cons for the last five or so years. I start them at 2nd level and use the old dungeon levels"; and Ernie Gygax has also recollected these match the earliest games). Let's assess the first-level key in detail.

Design: The dungeon design is of the fully-packed, use-every-space style, similar to what we see in the early Dungeon Geomorphs product, AD&D Module S3 (Expedition to the Barrier Peaks), etc. Many rooms are entirely empty, with no key code whatsoever. The Red Baron estimates there are about 160 rooms on the map; my own estimate is that maybe 40 have a key-code in them. That would be about 25% with any content at all -- similar to the guidelines in OD&D that about 33% have content (per 2-in-6 die roll; Vol-3, p. 7), the earliest Monster & Treasure Assortment suggestion that only 20% of rooms have monsters, etc.


Characters: Gygax's brief key doesn't state an expected party size. Recall from last week that neither do the core OD&D rulebooks. If we assume standard modern convention group size, then that could be maybe 6-8 people -- but recall also that in the earliest days a group might commonly be 12 or more people at once (e.g., see Jon Peterson, Playing at the World, p. 563, re: Gygax at Origins II). The question of what might be reasonable is, in fact, the whole point of our current investigation.

Monsters: The key for the level has 18 numbered codes; several are repeated in the map in clusters of up to a half-dozen located together (e.g., the kobolds, goblins, orc entries). All 10 of the monsters from the OD&D Monster Level Table 1 appear in this key. There are 3 from the Level 2 list (hobgoblins, berserkers, gnoll), and one each from Level 3 (giant snake) and Level 4 (giant beetle). One is novel (giant bats), and the last two keys have evil NPC spellcasters (2nd level magic-user and cleric, which could arguably belong on any of the 1st-3rd level lists).

Wandering monsters are not shown here, but if Gygax used them at all, it would make sense to simply use the book tables from OD&D Vol-3, because (as noted) they inherently represent the Castle Greyhawk ecology, and are a perfect match for the monsters present in lairs (see above).

Monster lair numbers are given in ranges of 3-7 orcs, 3-12 giant rats, etc.; very similar to what was suggested in Gygax's Monster & Treasure Assortment, and Holmes Basic D&D products. Using my "Equivalent Hit Dice" (EHD) analysis, I find that the average (mean and median) danger level of these encounters is 6 EHD, i.e., a fair fight for around 6 1st-level PCs. (There are reports that Gygax adjusted numbers to match the strength of PC parties present. Did he stay in the given ranges, I wonder?) The expected total strength for all the set encounters (ignoring duplications) is about 90 EHD.

Treasure: Almost every room with a monster has a treasure listed -- only 2 of 15 do not. This contrasts with the OD&D Vol-3 rule (p. 7) that half of monsters will have treasure; but it closely matches the Monster & Treasure Assortment rule that "about 20% of the monsters should have no treasure whatsoever". There is one room with no monster and a hidden treasure -- the largest reward on the level (room 8, with about 1/3 total value on the level). Note that with some key codes being duplicated, it's questionable whether every such room has the same treasure or not (unlikely that every kobold room has the same dusty silver mirror?), which would skew the treasure-to-monster ratio lower. The objects are clearly not from the monster-keyed Treasure Type tables, because coin values here are usually in the 10's (not 1000's as on that table; with one exception in room 4). Coin treasures are in the same basic order of magnitude as the OD&D dungeon treasure table (Vol-3, 7), without being exact matches. The gem values are usually in the 100's (as per book median gem type).

Total expected treasure (again, ignoring duplicated rooms) is about 4,400 gp. Given that, there are about 50 gp of treasure per monster EHD.

Magic: Magic is also relatively abundant. There are 6 of 14 treasures that have magic items (43% rate), for a total of 7 useful magic treasures; this includes 2 potions, 2 magic weapons, a magic shield, ring, and a staff (with 6 charges left). Contrast with the book rule that says only 5% of treasures at this level should have magic included (from which we might expect there to be none at all, given the total number of treasures here) .

Experience: Reports say that Gygax indeed used something like the Greyhawk revised XP system, as shared by B/X , AD&D, etc., instead of the original Vol-1 system (or maybe something simpler, like a flat 10 XP per monster HD?). For example, a play report from January 2008 with Gygax says, regarding area 18 above, "they got '20 xp from 1 gnoll' and '30 xp for the evil cleric'" -- which would be correct values per the Greyhawk chart for a 2 HD gnoll and a 2nd level cleric with double bonus for special abilities (suggested on Sup-I, p. 13).

So let's assume Gygax used the Greyhawk table, and each of the keyed areas only appears once. Then the total monster XP is about 2,000 and the total treasure XP available is roughly 4,400, for a grand total of about 6,400. This gives a monster: treasure XP ratio of about 1:2 (31% monsters, 69% treasure). Note that the total would only be enough to graduate 3 fighters to 2nd level.

On the other hand, if we use the original Vol-1 XP system, then the total of about 90 EHD would give monster XP of 9,000 or so, treasure the same 4,400, for a grand total of 13,400. Ratio would be inverted to about 2:1 (67% monsters to 33% treasure), and this would be enough to raise 6 fighters to 2nd level -- a good match for the 6 EHD average encounter here.

A few commenters have helpfully pointed out that there's at least a question about whether we should be adding XP from picking up and keeping magic items or not. There is a short line in OD&D suggesting as much (Vol-1, p. 18: "when they obtain various forms of treasure (money, gems, jewelry, magical items, etc.), they gain 'experience'."), but there are no rules or suggestions for what values to give them. Gygax gave a few examples in The Strategic Review Vol. 1, #2 (e.g., potions 250-500 XP, scrolls at 100/level, a sword +1 at 1000 maximum, etc.). Later in the AD&D DMG Gygax gave full XP valuations for all items in the game; they broadly match the Strategic Review values for potions & scrolls, with other items  mostly devalued (on the order of about a half).

As a rough estimate, the first level of Castle Greyhawk would give about 5,000 XP for magic items under either system. This would make a very significant difference; roughly adding 2/3 more XP under the Greyhawk system (constituting 44% of the total, and then very nearly promoting 6 fighters to 2nd level), or about 1/3 more under the original Vol-1 system (then 27% of the total, and promoting as many as 9 fighters to 2nd level). If we look at other adventures in the Basic D&D line, then this won't be a consideration, because XP-for-magic-items was entirely dropped in those rulesets (even after Gygax personally edited the Holmes Basic D&D set).

44 comments:

  1. Interesting stuff. Regarding the encounter lines that are keyed multiple times, do you think it's possible that the monsters could be encountered multiple times but the special treasure would come up only once? Just a shot in the dark guess since I've done something like that before.

    25% is dead on the rate of rooms with monsters from 1e (rolls of 13-17 on d20), although also including tricks, traps, and special rooms as "rooms with content" would bring the expected rate up to 40%. Your text and the provided table seem to indicate you were looking at monster encounters only as "rooms with content", but I wanted to mentioned both interpretations for clarity.

    Magic being on the high side isn't too surprising, given both that magic items in Gygax's modules seem to be more common than the book rates (e.g. James Raggi's 2011 review of B2 counted 69 magic items even with bundling stacks [arrows, multi-spell scrolls, etc.] as single items and discounting the turn-resistance stuff in cave K) and that Gygax was at least reputed to run a very "easy come, easy go" style with respect to treasures (as evidenced by monsters like aurumvorax, rust monsters, oozes, etc.).

    ReplyDelete
    Replies
    1. Right, totally agreed that's my best guess; probably you've got a whole bunch of monster rooms in a community and only one of them have the treasure (which itself is pretty similar to B2, as you say).

      In this particular case my 25% with content is counting anything that isn't literally just blank (incl. traps and even "rustling corridor" and "water seepage"). With later adventures that will become less of a distinction to make (completely blank rooms becoming less used over time).

      And I also fully agree that Gygax's magic distribution was consistently generous throughout all of his adventures that I've seen -- very dissonant with the book rule percentages (at least for low levels). Great observations!

      Delete
    2. I almost wonder if the passage in the AD&D DMG about how rare and special magic items should be is one of the sections actually written by someone other than EGG (in the DMG Designer's Notes in The Dragon at the time of its release he mentions that in order to get the DMG finished ASAP he assigned TSR's newly-hired "design department" - Lawrence Schick and Jean Wells - to ghost-write some sections as a trial-by-fire). Not only were his modules generous with magic, the Monster & Treasure Assortments were as well - something like 20% of the treasures were magic items - and this key backs up both of them. The random treasure stocking tables in D&D Vol. III and Strategic Review #1 both have a much lower occurrence, but I think that can be explained that in both cases those tables are for use in generating "filler" treasures after the most important and valuable ones have been deliberately placed. So only 5% of random/filler treasures include magic items, but if 50% of deliberately-placed treasures do, then you might see an overall rate of something more like 20-25%.

      Delete
    3. That's a fair theory. Part of the ongoing goal for my project here is to try to tease out what was intended for OD&D DM-fiat "special" treasures for which we just didn't get any guidance at all. I think it's actually a really good point (for me to keep in mind) that the Vol-3 treasure table could, quite likely, not having any bearing on the main DM-placed treasures at all.

      Now that I think of, I'll observe that the OD&D treasure type tables also have magic percentages averaging about 30% (range 10-60%). Of course, that's not for standard dungeon play, but is suggestive.

      On the other hand, the DMG random dungeon appendix has a treasure table with only 3% occurrence of magic (throughout the dungeon), so I think it's also possible Gygax was maybe a little bipolar on the issue (or being coy and liking to pleasantly surprise players, or something).

      Delete
    4. The possibility of the book treasure tables being meant as completely separate from DM-fiat treasures isn't something I'd considered before, but it sounds entirely plausible. My intuitive counterpoint would be that an awful lot of special-sounding stuff comes up in those tables for them to be generic, but that might be an issue of recycling content (e.g. magic item A is placed specially in adventure 1, magic item B is placed specially in adventure 2, etc., then while compiling tables for publication, magic item A and magic item B are both added as rare results to fill out the lists). Interesting idea!

      Unrelated side note: I've always found it curious how magic weapons get special details all the time, yet magic armors and shields are often just a "plus" rating (at a quick glance, 0e has no special magic armors and 1e has only 4 [2 if I discount the shields that are just different against missile attacks] ). Granted, some of that is rolled up in secondary equipment like gauntlets of ogre power, cloak/boots of elvenkind, etc., and 2e AD&D at least had a ton of special cases for making armors out of monster parts to get bonuses, but it just seems like a strange area of omission.

      Delete
    5. When it comes to the DMG random dungeon appendix, it's worth remembering that it's pretty much just an expanded and prettied-up version of The Strategic Review No.1's Solo Dungeon Adventures. There you see a 3% chance for magic, but if there's a monster it's by monster type "with pro rata adjustment for relative numbers". The DMG simplified this to just being two rolls (+10%) on the treasure table instead, bringing monster magic up to a whopping 24.31%. (Albeit complicated by a 25% chance of monsters, 60% chance of monster treasure, and 1/15th chance of monsterless treasure.)

      While I'm honestly not sure, I suspect that Solo Dungeon Adventures might have low-balled treasure a bit since it would only have been shared by a single player rather than a group? Still, it remains the only thing close to a dungeon design guide in the DMG.

      Delete
    6. Ash, re: weapon special abilities: I'm sure that's a reflection of the literature where usually the weapons, esp. swords, get special names and abilities and armor generally doesn't. E.g.: Thor, Arthur, Elric, the Broken Sword, Fafhrd & Gray Mouser, etc., etc. Most notable of course Elric which basically spawned the all swords (and only swords) being intelligent in OD&D, etc.

      Some have pointed out that it's effectively a special class feature of fighters, since only they can use swords, and thus unlock all of the intelligent/special abilities. I doubt that was fully intentional, but interesting effect nonetheless.

      Delete
    7. Neveron: I totally overlooked the DMG +10% monster rolls, that's a good point.

      Delete
    8. I agree completely that special weapons are far more prominent in the literature than armor (it might be an enlightening sociological study to see how that varies when considering non-American/European sources, but I digress). I will say, though, that the first exception coming to mind is the Chaos shield that Elric quested for in Moorcock's stories. I could've seen including some leather armor made from the flesh of Achilles (or Siegfried of Die Nibelungen, for a more obscure reference), but such was not to be and probably too dark for the general tone of the D&D core rules.

      Delete
  2. As Gary went along he found that the game was more fun for him when he made it a little less deadly for the players. I think he wrote his revised rules down in a few posts on ENWorld in 2004-2007 and had guys start at level 3 and gave out more stuff.

    It would make sense that his creature load out was similar to the books, since he wrote them based on his experience at the time and not the other way around.

    So from what I have learned this all seems perfectly normal for a late-Gygaxian ecology

    ReplyDelete
    Replies
    1. Totally agreed, and following that I'm also in the zone of starting PCs at around 3rd level myself these days (at least in my recent campaign). If you had like 4 PCs of 3rd level (total 12 levels), I think they could pretty reliably beat up most of the monsters here (average 6 EHD encounter). The creature population here certainly bolsters my prior understanding that the standard rules basically express his Castle Greyhawk ecology.

      Delete
  3. Great post - I agree that the 2008 version likely represents the pre-1974 dungeon.

    It would be interesting to compare it with Paul Stormberg's Origins II version, some day...

    ReplyDelete
    Replies
    1. I’m pretty sure someone already did, and determined that the maps at least were the same.

      Delete
    2. Good question and cool point, glad to know that!

      Delete
    3. Even better, thanks for the link!

      Delete
    4. That's an epic thread I somehow missed while preparing my "Greyhawk Castle" post - answers a lot of questions that I had. Thanks for the link!

      Delete
    5. I should also point out that (thanks to Zenopus for linking to it) when Gygax talked on the Pied Paper boards in 2004 about running games at conventions in that era, he said, "I start them at 2nd level and use the old dungeon levels."

      Delete
  4. Don’t forget that magic items are also worth XP as treasure in OD&D. The FAQ article in Strategic Review #2 gives guidelines that line up closely enough with the values in the AD&D DMG that I always just use those. Adding those will significantly alter the monster:treasure XP ratio

    ReplyDelete
    Replies
    1. Well that's a fair point. Admittedly I skip that in both my games and analyses because the values are unstated/ambiguous (unless we just wholesale back-port in the DMG values). Possibly his reduced Greyhawk XP system counterbalances my non-use of XP awards for magic items (as a 0th-level approximation).

      Delete
  5. Also, when calculating the total XP value of the level, and how many characters can gain how many levels out of it, you really should factor in the repeated encounters, some of which look to appear as many as a half-dozen or so times.

    ReplyDelete
    Replies
    1. You're right there as well. And again I confess I skipped that for a few reasons: (1) the map key numbers aren't legible enough to be sure which is which, (2) it's a bit ambiguous how the monsters are distributed, and (3) it's unlike most later adventures, so it didn't fit with my overall process well. Hopefully this can be a sketchy suggestion, and future posts will be less ambiguous.

      Delete
    2. It's all good. Your calculation that the average EHD per encounter is 6 is valuable, and I think the repeated encounter areas are balanced out by the likely assumption that a single party wouldn't attempt to clear the entire level before moving deeper. Meeting each type of encounter once is probably as fair an assumption of the typical successful party's activity as anything.

      Delete
  6. Very interesting analysis. That is a lot of empty rooms. I wonder how mapping was done at his table. Even with a map it must take a while to find your way out.
    I think there might be an error for room 4 as it seems to contain 1000gp but none is counted in the total

    ReplyDelete
    Replies
    1. Yeah, the people who played with him in the early days (thinking M. Mornard here) were pretty adamant that exploring/mapping was THE core experience of the gameplay, and that the whole party was bent over the map in silence sweating bullets on the lookout for error and wandering monsters. Probably an amazing revelation for particular people at a particular time.

      Thanks for pointing out the spreadsheet error, I just fixed it. (Kind of embarrassing; that was about one-quarter of all the treasure in the dungeon.)

      Delete
  7. I feel like I've read in the past that Castle Greyhawk magically restocked itself. Is it possible that the idea in campaign play would be to have an unkeyed master map, and to create a photocopy with a brand new key each week? That would ensure that the 1st dungeon level wouldn't run out of treasures for 1st level characters to find, but also make it take a few trips into the dungeon before they would hit 2nd level. That would resolve the discrepancy between total treasure and a party size of six or so.

    ReplyDelete
    Replies
    1. That could be possible. I think my top hypothesis would be that Gygax just kept the same key and didn't make any notes or mark changes at all. (So: effectively reset simply out of non-action.) I could be crazy but getting access to a photocopier was nontrivial in those days.

      So , good guess about the XP issue. Also Trent above points out the OD&D rules briefly assert that magic items should get XP, too, which I don't account for here; although that might be counter-balanced by reducing XP for PC level ratio, etc.

      Delete
    2. I feel like I've seen some other maps of that vintage (maybe scans in Rob Kuntz's DVD) that did show some evidence of play-based tweaking and restocking: encounter areas with lines through them, evidence of erased areas on the map and later additions (i.e. encounter areas 1-12 in pencil, areas 13-16 in pen, etc.). But I really doubt it happened systemically or very often.

      Delete
    3. Right -- Gygax even explicitly says that in Vol-3, p. 8 ("Maintaining Freshness"). That said, I could easily imagine him taking the "lazy DM" route and just not bothering at some point.

      Delete
    4. I wasn't alive back then, but I did check on Wikipedia real quick and it seems like it was common to have coin-operated photocopiers in libraries in the 60's and 70's. I know that at least by the 90's most hardware stores did as well; I always went to the TrueValue to copy more character sheets when I was a kid because it was a couple miles closer than the library. But I think there's a good chance EGG had access to photocopying if he wanted it.

      The magic items are a good point, though, you're right. I don't have the values handy, but off the top of my head potions would be a few hundred each and a +1 piece of equipment might be 1000 XP. So that might increase your calculated total to be enough for a party of 4-6 members to reach 2nd level.

      Delete
    5. FWIW, per the AD&D DMG the XP values of the magic items on the matrix are as follows: potion of strength (500), +1 shield (250), potion of undead control (700), +1 battle axe (400), ring of protection (2,000), +1 dagger (75), staff of healing with 6 charges (~1,500). So, 5,425 if all of it is recovered. So the total XP value from monetary treasure and the total XP value from magic items are almost exactly the same, and even with Greyhawk monster XP values that's enough to get a full party of 6 up to 2nd level, with treasure having accounted for about 85% of the total XP.

      Delete
    6. In Gary's mid-2000s home OD&D game, which used an abbreviated version of the dungeon (including this 1st level map), he indicated that he was not restocking the dungeon: "As they are active in the dungeon complex, only wandering monsters will be likely to be encountered on the depopulated 1st level" (2/22/05, Enworld post). But he did allow surviving stocked monsters, particularly the Old Guard Kobolds, to become more dangerous: "PCs have enabled the monsters inhabiting the levels to become better armed and most astute in their tactics" (11/19/04, EnWorld).

      Delete
    7. He even posted his rules for leveling up the Old Guard Kobolds on Kuntz's Pied Piper Forums in Feb 2004. See this K&KA post, which quotes/links the original but is a bit more easily readable.

      Delete
    8. Note, however, that the above is not necessarily how he ran the original GC games in the early '70s.

      Delete
    9. Oh geez, that's super lovely! Personally for simplicity purposes I also kind of prefer _not_ repopulating dungeon level when they've been cleared. Once the lair of a monster is wiped out, I strike them from the wandering chart, too. So I'm pretty comfortable with those choices.

      Delete
  8. I thought experience was divided by player characters so if 8 did it all and lived at 6400 experience that's 800 each. What am I confusing here.

    ReplyDelete
    Replies
    1. In order for 3 fighters to get 2,000 XP apiece, the other 3 fighters are assumed not to have survived...

      Delete
    2. Yeah I think I'm just pointing out how many total fighter-levels it makes as a way of saying, "this would not be enough XP to graduate a party 6 to 2nd level".

      Delete
  9. That "someone" was me! I took that picture the evening of August 16th, 2007 - the day before D&D 4e was announced, and Gary's 25th wedding anniversary. He ran a game for the ENworld moderators that were lucky to be able to attend GenCon 2007. I'm glad it's still being talked about 13 years later.

    - Matt "Eridanis" Bogen

    ReplyDelete
    Replies
    1. OMG, what a spectacular service you did all of us with that photo. It's really a historical treasure!

      Delete
    2. I added those details into the main post to make it look more like I know what I'm talking about, thanks! :-)

      Delete
  10. Greetings. I've been reading your series and other various people delving into the minutiae of dungeon filling and I had a question for you. Considering that in the olden days (it seems to me, I wasn't there), megadungeon levels should align +- with the character levels, we can expect that, for example, a 3rd level dungeon should give about 20k-25k of GP worth. However, if I randomly assign monsters as per AD&D formula, I end up with about 187 HD for 39 rooms. Using the proposed 50 GP per EHD, it gives a total of 9350. Therefore, using that rule, it seems we are half-short of what a third level should generate. My question is therefore: which is "broken"? AD&D monster HD per dungeon level? 50 GP per EHD? The fact that a level should give enough XP for a party to level up? It seems that some of the module you reviewed didn't have enough GP for party to level-up. However, others, such as Keep, does have it. For more info on where I come from, I previously watched this very informative video: https://www.youtube.com/watch?v=VEpdWgGgcrg&ab_channel=Jon%27sD%26DVlog Also, even if we add up XP value of the monsters, it doesn't add up since the majority of XP is supposed to come from treasure and not from monster XP if I recall correctly. Again, thanks for the hard work and cheers!

    ReplyDelete
    Replies
    1. David, thanks for reading this stuff and the great (really, the crucial) question. I'll confess that as 2020 started I was writing these articles and trying to get to a place where I'd have a concluding thesis about precisely that -- and then got derailed a bit.

      I agree with the assumption that PC level should generally track with dungeon level (it's certainly the simplest assumption).

      As I write this, I'm not super confident that I have a solid thesis. Maybe there wasn't any thoughtful balance to it, and just no one ever noticed. But one thing I always cycle on is the switch in XP awards from Vol-1 to Sup-I and after; a huge downgrade in monster XP (about 1/10th at lower levels) -- as you say, so making most XP from treasure -- but no corresponding change to the treasure tables or recommendations. If it was balanced at the time of Vol-1, then it wasn't later, and vice-versa. AS we see here, using the Vol-1 monster XP awards, and then also maybe adding in fairly big bonuses for magic items, you do get enough XP to level up reasonably. Still working on the grand unified theory for that, honestly.

      Delete