Book of War: Weather Control

Book of War includes the following simple, optional rule for weather, rolled once at the start of a game (Open Game Content between the rules):

2d6 Weather Effect
2-7 Sunny Orcs/goblins at -1 morale
8-9 Cloudy (No effect)
10-11 Rainy Missiles at -1 to hit
12 Stormy No missiles, move cost x2,
cavalry attack 1 die/figure

Now, most of the time this won't make much difference -- until the ability to Control Weather gets put into play, as by any Storm Giant or Wizard in the advanced game. Then it becomes highly likely that the special figure can severely disadvantage opposing goblinoids, or more generally any missile troops or cavalry (recall that modifiers are in terms of a d6 roll). We've found that when playing by advanced rules, both sides almost always pick at least one Storm Giant by default so as to access this powerful ability.

I've previously written a column on "Spells Through Ages -- Control Weather", as Book of War was being finalized last year. Among the observations there is that the OD&D spell is immensely powerful, permitting instant conjurations of tornadoes, heavy rain, heat waves, etc., without limit. AD&D did seem to have a clever modification of allowing only a "one step" modification to weather, which I included in BOW. You can't automatically switch the weather to "Stormy" and lock down all the opponent missiles; the natural weather already has to be at least "Rainy" before that's possible. (And multiple spells or giants won't help; only a one-step change from natural, maximum, is ever allowed.)

Even so, the ability has proven to be very powerful. Partly this is due to the way we've adjudicated the gray areas of the spell in-game: First, there isn't any 10-minute delay to the spell as specified in AD&D, so the effect is immediately usable by the caster (otherwise, you'd need a rule for casting it prior to a battle, unlike any other spell, which is the only time it would be useful). Second, you're not stuck with just one change; if desired, you can switch the weather up and down by one step per turn (common usage: make the weather "Rainy" and cripple the opponent's missiles on one turn; then switch back to "Cloudy" and let your own missiles fire effectively, etc.). Third, opposing weather-controllers don't simply negate each other; they're allowed to ping-pong the weather back and forth on opposing turns, according to taste.

Now on the one hand, you might find this overly fiddly (weather changing dramatically every minute?). But we've actually found this to make for a lot of exciting and dynamic play situations (fun, you might say). Control Weather is a top-level ability that makes an immediate and high-impact effect on the game; it's in your face every turn as the casters battle for the sky overhead turn-by-turn. The alternative option, perhaps saying that preferred weather is picked only once before the battle and that opposing casters simply cancel each other invisibly, seems not nearly so dramatic or interesting.

One other thing that we need to adjudicate, not being written in the book: What happens to weather when the controller is killed? Options here seem to be (1) Nothing, weather stays at current level for the entire battle; (2) Gradual reversion back to natural weather, by one step per turn; or (3) Immediate switch back to natural weather, cancelling the magic. (Consider also how this interacts with an opposing caster who was being partly cancelled out; does it switch multiple steps to their preference instantly?) We've been playing by option #3, which is perhaps mangling normal D&D rules a bit (spell effects generally don't end with the death of a caster), but it makes for interesting in-game situations, where opposing weather-controllers are critical targets if the opposition can chase them down and eliminate them.

So: what's your preference? Do you like dynamic weather changing turn-to-turn while a battle progresses, or do you prefer the AD&D-style delay, where one choice for weather would be maintained throughout the whole battle? Do you like the idea of opposing casters ping-ponging the weather every turn, or should they simply neutralize each other without visible effect?

[Picture courtesy smiteme under CC2.]


Thief in Action

Over at Playing at the World this summer, Jon posted pages from the earliest version of Gygax's "Thief Addition" (used at GenCon VII prior to publication of Sup-I Greyhawk). I think one of the great things in there is a half-page "Example of a Thief in Action" which seems to clarify some expected usage of the thief class at that time. (Not unheard of rulings, but clarifying nonetheless.) Here we see:
  1. Automatic detection of traps prior to removal attempts.
  2. Explicit note that failure to remove a trap sets it off on the thief.
  3. Play wherein the thief splits from the party, hides from a pursuing monster, and then slips into its lair to lift its treasure.
One thing that always bothers me about these scenes -- surely I'm not alone -- is that the number of successes on display for the sample thief is wildly unlikely (probabilistically) for the chances given for the class abilities. There's definitely a tension in early D&D between sort of expecting a thief to succeed on several burgling-type tests in a row, but then having the probabilities set so low that they will fail most of the time. Some people will want to revise those chances upwards, whereas my response is to mostly remove any penalties for failure (like #2 above), and to try and give very generous benefits to success.

See Jon's article here (esp. the last image with the example of play).


Book of War: Variance

The Book of War mass-warfare game (see sidebar) is designed to quickly model standard D&D combat in as statistically high-fidelity a fashion as possible. Each d6 roll represents the attacks of 10 men or creatures (or more specifically: attacks by 5 men in the front row, over 3 rounds of combat, i.e., 15 individual attacks). The average casualties from the system are known to be very close to that of actual D&D combat played out en masse.

An outstanding question for me, however, has been: what of the variance? One fairly obvious observation would be that reducing the overall number of die rolls would make the game more "swingy" (i.e., reduced sample size causes higher variation). But on the other hand, by eliminating any damage rolls, that's actually one source of variation removed, so perhaps it pulls things back in the other direction. For a long time my assumption was that BOW combat would be about 3 times more swingy (in standard deviation) than normal D&D combat, but I didn't take the time to find out exactly.

Here are more exact comparisons, made possible by an edit to the BOWCore Rule program seen previously in the link above; revised version here). Numbers below are in terms of individual men killed per turn (3 rounds) by a single mass figure (5 across the front).

Variance in D&D & BOW

AC Mean Stdev Mean Stdev
10 5.9 3.3 6.7 4.7
7 4.4 3.1 5.0 5.0
4 3.0 2.7 3.3 4.7
1 1.5 2.0 1.7 3.7

Conclusion: The mean number killed is, as designed, as close as we could make it using a simple d6 roll. The BOW mechanic is indeed a bit more "swingy" (higher standard deviation/variance; more chance for an underdog to surprise a powerful opponent), but only on the order of about 1.5 times that of regular D&D (not triple the standard deviation as previously presumed).


Playing at the World

Many of you will already be aware of this, making me late to the party, but if you don't: Jon Peterson is a computer engineer and historian of D&D who self-published a 700-page book called “Playing at the World” which digs deep into the original development of D&D, its historical antecedents over several centuries, and the digital gaming revolution which came afterward. Apparently in some cases he traveled the world to dig up rare archival documents around the time of the first D&D game writings by Arneson and Gygax (and others). His book is available on Amazon, and he also started an associated blog a few months ago. (And I discovered his effort via an interview at Wired.) Work like this is enormously appreciated by people like myself, and I plan to eagerly read what other insights he posts to his blog in the future.


Jaquaying the Dungeon at the Alexandrian

A really fine series of articles I just found at Justin Alexander's blog site: "Jaquaying the Dungeon". This mostly concerns the kinds of complex and non-linear dungeon connection features seen in the work of Paul Jaquays, about whom I'm not alone in writing positively about in the past. Justin's series is highly recommended (and probably the verbing of Jaquays' name alone was enough to sell me on it).


Random Dungeon Stocking

Today I'd like to consider rules for semi-randomly stocking dungeons. I'll start with what I currently do, and then consider the history of such rules in the classic D&D book.

Delta's Rule: On a standard piece of graph paper (30×40 squares), I seem to get about 40 rooms or so, that will be dealt with in groups of sixths (i.e., about 6 rooms per category). First I make a list of about 6 "special" rooms I want to place (major sites, or monster clan lairs, usually with significant treasures); about 6 tricks or traps along the way (often corridor/navigation blockage or misdirection); and about 6 mundane "guard" areas in conjunction with the specials. Then I actually draw the map and place these indicated areas, filling about half the rooms of the map. Then I go through the remaining half of the areas (usually side-rooms, rarely main thoroughfare chambers) and roll dice: 2-in-6 indicates random monsters from the wandering charts (or semi-random; move or up down the list to something that generally fits). A second roll indicates treasure, 3-in-6 for monsters and 1-in-6 for otherwise empty rooms. If this random method comes out to greatly more or less than the expected 6 rooms or so (again, 1/6 of the total) then I go back and add some or take some away for the right distribution.

The end result of the preceding is 1/6 each "special" (major areas), 1/6 tricks/traps, 1/3 monsters (half selected guards, half random), and 1/3 empty -- which you'll see replicated in many of the classic rules below. One thing that's added for me are the pre-planned mundane guardrooms, which serve to flesh out the "theme" of who controls that part of the dungeon, and serve as buildup and clues to the major "special" lairs (trying to do that part randomly just doesn't build a proper theme). I think the overall effect is similar to Gygax's B2 or G1-3, with a primary monster type in each region, a general buildup to the encounters, and various and sundry surprise lurkers in the side areas (with perhaps somewhat more empty rooms).

Original D&D: OD&D is most similar to what I do above. It says, "It is a good idea to thoughtfully place several of the most important treasures, with or without monstrous guardians, and then switch to a random determination for the balance of the level" (Vol-3, p. 6). The random rolls are identical to what I use above (2-in-6 has a monster; 3-in-6 monsters with treasure, 1-in-6 otherwise empty rooms), with a helpful table provided for these treasures (balanced by dungeon level, and quite different from the otherwise well-known Treasure Types intended for large wilderness groups).

One thing that's missing here is the overall monster distribution-- how many of those "special" areas are expected? A second thing is that there's no stated allowance for tricks/traps, although a list of suggestions comes immediately prior to this section. One very interesting thing about that list of tricks/traps -- they're almost all navigation-misdirection elements that are most likely to appear in a corridor (e.g., false stairs, slanting passages, sinking rooms, teleportation, etc.). Using that as inspiration, it's one reason I feel compelled to decide on a list of traps prior to mapping the dungeon; the plan of the corridors and rooms is intimately tied to what navigational trickery is possible. But keep in mind that "specials" as used in Original D&D means the major treasure caches and monster lairs, not weird magical trickery, as the phrase might be used later.

Dungeon Geomorphs: Gygax's Dungeon Geomorphs product (featuring the older "use-every-space", thin-blue-line wall design for the basic dungeons) has stocking suggestions on the inside cover. It says: "Roughly one third of the rooms should remain empty. One-third should contain monsters with or without treasure (possibly selected randomly using the Dungeons & Dragons Monster & Treasure Assortment), one-sixth traps and/or tricks, and the remaining one-sixth should be specially designed areas with monsters and treasures selected by the DM (rather than randomly  determined). Slides, teleport areas, and sloping passages should be added sparingly."

Note that these are the same overall distributions as I use today. One mistake I made in the past, I think, was to take this a simple random d6 roll in every area of the map (1-2 empty, 3-4 monsters, 5 trap, 6 special). The problem with that is that it handcuffs you into doing work to freshly invent traps and specials at times you don't want and places that don't fit (there being no random follow-up method for those). Moreover, the all-random monster method produces no intelligible theme or pattern, so nowadays I fill at least half of those intentionally beforehand (pretty simple, mostly just duplicating the "special" types).

Monster & Treasure Assortment: This collection of treasures to be used in different dungeon levels has a different distribution suggestion on the cover: "However, it is recommended that the DM selectively place as many treasures as possible, doubling up in some cases, and augmenting the whole by noting where and how the treasures are protected and/or hidden. It should also be noted that just as a dungeon level should have monsters in only 20% or so of the available rooms and chambers, about 20% of the monsters should have no treasure whatsoever."

It seems to me that having monsters in only 20% of the rooms is way too low and wasteful. I don't think any other product ever suggested such a low number, and I think it can be pretty much ignored. Note that my current rule places monsters (specials, guards, and random) in 3 rooms in 6, i.e. 50%, or more than twice as many as suggested by the MT&A rule. With my rule 1 room in 3 is empty, which some people might already consider excessive. Furthermore, while it explicitly talks about the DM selectively placing only treasures and not monsters, it would clearly be a significant error to just go through a dungeon level rolling 2-in-10 for all the monster placements.

AD&D Dungeon Masters Guide: Again, no global dungeon distribution rules are given, although we can see a table in Appendix A: Random Dungeon Generation that gives some clues (p. 171). This table uses a d20 roll, with results of 1-12: Empty; 13-14: Monster only; 15-17: Monster and treasure; 18: Special (or else stairs), 19: Trick/Trap; 20: Treasure.

A few comments: This seems to be similar to the MT&A rule, with fully 60% of the rooms empty, only 25% with any kind of random monsters, and just a tiny number of specials or tricks/traps. In practice this seems incredibly sparse, and when using those rules I personally increase the numbers of non-empty results. One obvious sticking point are the "specials" which have no random method for completing them, and would be an outright conflict of interest for those playing a game solo by those rules. However, this does lead to the fascinating rule, "For special areas you can have a friend or correspondent send you sealed information." (p. 173)

Holmes Basic D&D: The "blue book" Basic D&D by Eric Holmes somewhat surprisingly doesn't have any distribution suggestions. It has a very short couple of paragraphs on Dungeon Mastering as a Fine Art, where it states that "the Dungeon Master must sit down, pencil in hand, and map out the dungeons on graph paper... the introductory module 'In Search of the Unknown'... will be usable for initial adventuring as well as provide ideas for dungeon construction" (p. 39). My understanding is that earlier version of the D&D Basic set instead included the Dungeon Geomorphs product (above), which would thus pick up the slack both for mapping and distribution suggestions.

Moldvay Basic D&D: Moldvay's "red book" rules give a detailed step-by-step process for creating a dungeon, starting with: A. Choose a Scenario (listing 10 themes), B. Decide on a Setting (6 types), C. Decide on Special Monsters to be Used, D. Draw the Map of the Dungeon, E. Stock the Dungeon, and F. Filling in Final Details. In the Stock the Dungeon stage, he says, "Special monsters should be first placed in the appropriate rooms with special treasures. The remaining rooms can be stocked as the DM wishes. If there is no preference as to how certain rooms are stocked, then the following system may be used" (p. 52). This system is (as usual) determined by a d6: 1-2 Monster, 3 Trap, 4 Special, 5-6 Empty (treasure is indicated 3-in-6 for monsters, 2-in-6 for traps, and 1-in-6 for empty rooms). Traps and specials must be developed by the DM, but lists of ideas follow (e.g., navigational hazards a la OD&D, moaning room, talking statue, flying weapons, etc.)

Most of the time I'm highly impressed by the clear and detailed understanding that Moldvay provides; but here I think he falls down a little bit. First of all I'm not fond of the random method creating design work for the DM in 1/3 of the cases (traps and specials), since in theory the random generation should be relieving and completing the DM's labor, not making more of it. Secondly he's mangled the notion of "specials" a bit, using the phrase both for specially-selected major monsters and treasure, as well as randomly-indicated weird trickery, which is not terribly helpful. Using this in the past caused me quite a bit of frustration -- I'm not sure that drawing the map first, and then having traps randomly placed and determined (really unique to these rules except for the very small number in the DMG solo section), is the way to go. Furthermore, at the point where the DM turns to random stocking, the random system really needs to be a complete turnkey solution (and not demanding more work of the DM).

So, what system works best for you? (By the way, this marks the blog's 400th post. Thanks for reading!)