Hack & Slash posted On the Visual History of the Illithid the other day and pointed out that, in the original Monster Manual, the portrait of the mind flayer was surrounded by an irregular octagon that was completely unique within that tome:

“Although several creatures in the monster manual have borders, most are square. Only two other creatures, the Bugbear and Type V demons have octagonal borders and both of their borders are more regular. Each pane of the mind flayer border is of a different length, no two matching.” Which feels oddly appropriate, given the dimension-rending origins of the mind flayer in many versions of their mythos.

I was struck by the idea that you might be able to take that octagonal border and turn it into an iconic symbol or badge. An Icon of the Flayer. A couple dozen minutes of fiddling around in Photoshop gave me this:

After the review I wrote, several people have asked me whether I’ll be doing a remix of Hoard of the Dragon Queen to “fix” it the same way that I did for Keep on the Shadowfell.

Short answer: Nope.

Why? Because nothing about Hoard of the Dragon Queen made me want to run it. Which means I’m not going to run it. Which means that I’m not going to put the necessary effort into revamping it.

With that being said, while discussing the campaign over the past few days I have had a few thoughts about the approach I would take to remixing the campaign if I were going to do it.

NODE-BASED DESIGN

As I mentioned in my original review, the basic back story of Hoard of the Dragon Queen kinda screams out for a node-based design: Multiple factions of a cult simultaneously pursuing multiple artifacts should lend itself pretty much instantly to the PCs being able to choose which threats they want to prioritize.

One thing to consider, however,  is that D&D characters increase in power level over time, so you can’t just have a wide open playing field without risking either crippling difficulty at the beginning of the campaign and/or push-overs at the end of the campaign. With that being said, my understanding of 5E is that the “bounded accuracy” design is specifically meant to increase the range of tolerance for this sort of thing and pre-4E had at least a 3-4 level range of tolerance.

So what you need to do is design around funnels that refocus the investigation. (Or, alternatively, layer cakes while accepting that occasionally the PCs will backtrack and roll over some easier material.)

Having a bulkier initiating node that the PCs can gain a couple of levels during also makes sense (to take the edge of fragility off their characters). So it makes sense to keep the general idea of “siege on Greenest, followed by investigations at the cult camp.” In the camp you’d want to seed clues to three different nodes:

• Leads to wherever they’re transporting the stuff.
• An envoy from another faction of the cult.
• Reports from agents who are currently spying on a third faction of the cult (that somehow threatens this faction’s interests).

This 1st Funnel is primarily focused on figuring out what the cult’s true agenda is. The second prong of clues within the scenarios of this funnel, therefore, would show their fascination with Tiamat, hint that they’re looking for Tiamat-themed artifacts, and also reveal the “five-headed” structure of the cult. The first prong is the structural branch in which clues point them to the conclusion of the 1st Funnel: An “all-faction” meeting of the cult. At this meeting, the PCs would discover (or verify) that the cult is specifically interested in the five dragon masks. Furthermore, they would get clues pointing them towards 2-4 more dragon masks. These might include:

• Ongoing expeditions being run by the cult.
• Expeditions that are about to begin. (Do the PCs sabotage the mission? Race to meet them to the site?)
• Fortresses where the cult is keeping masks they have already obtained.

And so forth. This 2nd Funnel would lead them to Tiamat’s Prison where the final ritual is being performed (or whatever). If I was invested in the idea of the cult getting all five masks and raising Tiamat, I’d probably arrange things so that the masks have already been shipping to the final site before the PCs ever arrive at the various expedition locations. Given my predilections, however, what I’d probably do is design the ritual in a way that the PCs holding a single mask won’t completely disrupt it. Maybe something like:

• Each mask allows the cult to summon one Aspect of Tiamat (i.e., a deity-infused dragon of matching color). Or maybe the ritual just involves an appropriate dragon wearing the mask and, thus, channeling a shard of Tiamat’s soul. Either way, the point is that if the PCs manage to hold onto one or more of the masks… great! They have substantially reduced (but not eliminated) the effectiveness of the plans and Tiamat’s manifestation on this plane.
• Also designing 2-3 proactive nodes of assassins or thieves or the like that the cult would send after the PCs in order to retrieve the masks they’ve “stolen”.

USING THE RAW MATERIAL

If I’m looking to preserve as much of Hoard of the Dragon Queen as possible, I’d probably look at something like this:

Node 1: This consists of Episodes 1-3 from the book.

1st Funnel:

• For tracking the looted goods, I’d use a possibly abbreviated version of Episodes 4-5 but have it conclude at the Roadhouse. (With records there pointing to the conclusion.)
• The envoy would be from the Hunting Lodge (Episode 7, with the same relationship to the Castle in the Clouds for the conclusion).
• The agents would be spying upon Castle Naerytar. (Preparations are being made to send their delegation to the conclusion.)

The conclusion of the 1st Funnel would be a meeting at the Castle in the Clouds (Episode 8).

Remember to include cross-node clues: So, for example, the Hunting Lodge is also shipping goods to the Roadhouse (providing a place where you might be able to use additional material from Episodes 4-5) and has a spy from Castle Naerytar. And so forth.

The 2nd Funnel would be a hypothetical remix of the material from Rise of Tiamat (if it proves to have anything usable upon release).

THE FIRST EPISODE

 The other thing you have to do if you’re going to run Hoard of the Dragon Queen is fix the opening scenario.

First, the hook of “you have some vague reason to be in Greenest, but when you get there you find that the town is being attacked by a dragon” is far too fragile. One option would be to drastically ramp up the importance of whatever they need to accomplish in Greenest, while also making it massively time-sensitive and also being of a nature that a dragon besieging the town doesn’t render it moot.

Nothing comes to mind that fits the bill, though, so I’m going to recommend an easier fix: The PCs are already in Greenest when the cult attacks.

The far more problematic aspect of the first episode, however, is the lengthy section with the World of Warcraft quest-giver standing atop his castle walls and ordering the PCs to venture forth over and over and over again. I suspect you could probably salvage this section of the adventure by having the players receive a complete tactical overview of the situation in town and then venturing forth once to accomplish their missions all at once. That would look something like this:

• Dragon Attack (the dragon assaults the wall shortly after the PCs arrive in the castle)
• Old Tunnel (after the PCs prove themselves against the dragon, Escobert immediately tells them about it)
• Sallying Forth (Save the Mill, Sanctuary, and Prisoners)
• Sally Port (as they return to the keep, the sally port buckles and they have to help drive the raiding parties back out)

They can then question the prisoner with Escobert and you can cut that interrogation short whenever it becomes boring by having the half-dragon champion issue his challenge.

BRINGING THE AWESOME

Pearce Shea summarizes the big problem with Hoard of the Dragon Queen eloquently at games with others:

What we learn about kobolds [in Hoard]: They are small and they like dragons. They comport themselves menacingly.

When Paizo released Rise of the Runelords they reimagined goblins with Wayne Reynolds, and Goblins lit themselves on fire by mistake, drowned in half-full barrels, feared and hated horses (horses are kind of like their dragons) and one eats a man’s face off (is eating it, through a hole in the wall, when you find it). They roast limbs for fun, carry molten tongs and try to shove adventurers into a furnace. There are optional feral goblin babies to kill (or to try to raise). They have a song (it is lame), and a druid that moves through their bramble walls as if the brambles were no obstacle at all, and a chief that rides a giant gecko. There will be a man encased in glass, an aasimar becoming a demon, a barghest, seduction, romance, betrayal, two patricides, a boat hunt, family squabbles, rangers giving reports about goblin activity, flirting, grave robbing, lost mega weapon-type defense systems, ancient temples and fonts of evil power, demons, a mutant goblin and an imp that imagines itself queen. That’s in 60 pages or so (the first of six chapters). Half the length of Hoard of the Dragon Queen. fuck.

So the biggest thing you’d have to do during a revamp of Hoard of the Dragon Queen is liberally inject the scenario with awesome stuff: The cult needs weird rituals. Strange Tiamat lore needs to be scrawled in living inks. Feel free to take advantage of the bassabal culture I created for kobolds in my Shadowfell remix. Yank open your fantasy spice rack and start using it liberally.

On that note: Check out the remix Hack & Slash is embarking upon.

Something I’ve noticed for years is that serialized narratives with a lot of complexity and moving parts have a tendency to bloat up: You look at the first season of a TV series or the first two dozen issues of a comic book and there’s a ton of stuff happening and a lot of progress seems to be getting made… and then a couple years later it seems like the pace of everything has slowed to a crawl and it takes three or four or five times longer for anything to happen.

Sometimes, of course, this is just because the writing has gotten sloppy: You cut some corners to make a deadline and release something that’s not as tight as you would like. You realize that nobody seems to have really cared about the sloppiness of what you released and so you give yourself permission to be a little more sloppy in the future.

But I’ve just realized that there’s something else going on here and it can be explicitly quantified.

Imagine that a given multi-episode plot thread in a TV series takes 40 minutes to resolve. Assume each episode of the series devotes 75% of its time (30 minutes) to the primary plot of the episode and 25% (10 minutes) to developing these multi-thread plots. Assume a new multi-episode plot thread gets introduced every 3 episodes and equal time is given to each one. What happens?

• EPISODE 1: Plot A +10 minutes.
• EPISODE 2: Plot A +10 minutes (20 minutes total)
• EPISODE 3: Plot A +5 minutes (25 minutes total). Plot B +5 minutes (5 minutes total).
• EPISODE 4: Plot A +5 minutes (30 minutes total). Plot B +5 minutes (10 minutes total).
• EPISODE 5: Plot A +5 minutes (35 minutes total). Plot B +5 minutes (15 minutes total).
• EPISODE 6: Plot A +5 minutes (concluded). Plot B +3 minutes (18 minutes total). Plot C +2 minutes (2 minutes total).
• EPISODE 7: Plot B +5 minutes (23 minutes total). Plot C +5 minutes (7 minutes total).
• EPISODE 8: Plot B +5 minutes (28 minutes total). Plot C +5 minutes (12 minutes total).
• EPISODE 9: Plot B +4 minutes (32 minutes total). Plot C +3 minutes (15 minutes total). Plot D +3 minutes (3 minutes total)
• EPISODE 10: Plot B +4 minutes (36 minutes total). Plot C +3 minutes (18 minutes total). Plot D +3 minutes (6 minutes total).
• EPISODE 11: Plot B +4 minutes (concluded). Plot C +3 minutes (21 minutes total). Plot D +3 minutes (9 minutes total).

In actual practice, obviously, some of these plots would be ignored for several episodes and then given more screen time when they come back. And the actual amount of time devoted to each plot would vary. But you can see what happens: Early on there’s a limited number of focal points and they rapidly accumulate the time required to resolve them. But because you’re introducing these arc-plots at a rate slightly faster than they can be resolved, the number you’re juggling gradually increases over time (which further reduces the rate at which they’re resolved).

The key thing to understand here is that the bloat isn’t the result of the writing on a series changing: The bloat was there the entire time. It just took awhile before the symptoms of the disease became noticeable.

One TV series that seemed to largely avoid this problem while also enjoying the benefits of arc-plotting was Buffy the Vampire Slayer. The technique they used seems relatively straight-forward: They limited the number of arc-plots they introduced in each season and made sure that most or all of them were resolved by the end of the season. As a result, when they launched season 6, for example, they didn’t have any more balls in the air than they did when they were launching season 3.

A few days ago my Ptolus campaign had a 100 Session Review to reorient the group as we get ready to launch into the next hundred sessions. In reviewing our campaign journals for the first hundred sessions, however, I noticed once again how much the PCs seemed to accomplish in the first dozen or so sessions of the campaign and how everything seems to take much longer in the more recent sessions. But when I drill down and actually look at any given session, it’s not that there’s less stuff happening: It’s just that there’s more balls in the air.

I don’t know if I would necessarily change anything in the current campaign. But the example set by Buffy in controlling the pace at which new arc-plots are introduced in the campaign is something I’ll be keeping in mind for the future.

An idea that I’ve toyed around with for years is creating a hex map for the Underdark. I still haven’t done it. But recently I’ve been running a huge technological complex for Numenera with a hex map that shares a lot of similarities with the Underdark. If the idea of running a hexcrawl through the Underdark is something you’d like to try,  I think there are a few key points to consider:

(1) What makes a hex map work is that it abstracts the actual terrain of the game world. If you’re doing a wilderness hexcrawl, you shouldn’t try to map every tree… or even every single country lane. If you do that, you’re defeating the entire point of the hex map. Similarly, if you’re designing your Underdark with a hex map you should not try to map every individual tunnel. (You might map major thoroughfares, the same way that major highways or rivers would be indicated on your wilderness hex map.)

(2) One key distinction between a wilderness hex map and an Underdark hex map is that, generally speaking, travel is always assumed to be possible through the side of a wilderness hex. This is not necessarily the case in the Underdark and one thing you’ll want to develop is a key indicating a minimum of three states for each side of the hex:

• Open (there are lots of tunnels leading from this hex to that hex)
• Closed (there are no tunnels leading from this hex to that hex)
• Chokepoint (you can get from this hex to that hex, but only by passing through a specific keyed location)

Note that the existence of a given chokepoint could also be a secret that needs to be discovered (by either obtaining the information elsewhere or perhaps by performing a detailed survey of the area).

(3) The RPG industry has developed a fairly standard “vocabulary” of wilderness terrain types. (These actually predate D&D and were inherited from Avalon Hill’s Outdoor Survival when Arneson used it as a template.) These terrain types also have the benefit of being familiar to us in our every day lives: We know what forests are. We know what mountains are. And so forth. IMO, you’re going to want to develop a similarly interesting vocabulary of at least 4-5 different Underdark terrain types. And you’re going to have to figure out how to clearly communicate those differences to a group that probably doesn’t contain spelunkers (and certainly no fantasy spelunkers). The point of this, obviously, is to make the map more interesting: This both rewards exploration (a key component of any hexcrawl), but also to make the actual description of the PCs’ journey more engaging.

(4) The Underdark is fundamentally three dimensional in a way that the surface of the world is not. Keep that in mind, but don’t worry about it too much: The surface of our planet varies from 1,400 feet below sea level to 29,000 feet above sea level but we still successfully visualize it as a flat plane. Consider the minor elevation shifts I discussed in Xandering the Dungeon and apply the same logic at a macro-scale here: You can probably make your Underdark more interesting by saying “you have to go down and then over and then up to get to there”, but vast slopes and slants and descents and climbs can be abstracted onto a two-dimensional map. So go back to Point #1 above and remember to embrace the abstraction of the hex!

A quick review for those unfamiliar with the Eclipse Phase system: It’s a percentile system where you need to roll equal to or under your skill level in order to succeed. If you roll a success, your margin of success is equal to the number you rolled on the dice. If you roll a failure, your margin of failure is equal to the number you rolled minus your skill level.

(So if your Fray skill is 45 and you roll 27, your margin of success is 27. If you roll 89, your margin of failure is 89 – 45 = 44.)

Playing Eclipse Phase at Gencon this year, I noticed once again the difficulty some people have grokking this method of “calculating” margin of success. Part of the problem is that it’s discordant with how venerable percentile systems like Call of Cthulhu calculate margin of success (by subtracting the number you rolled from your skill rating). And part of the problem is that Eclipse Phase actually swapped methods between subtracting numbers and reading the die roll between printings. (The post-Catalyst Labs versions of the game should really have been clearly labeled a Revised Edition, frankly.)

But laying all of that aside, the huge advantage of the “read the die” method of calculating margin of success is that it completely eliminates calculation at the table when calculating margin of success: All you have to do is look at the dice. When you can get everyone to grok that (and to report their rolls as “XX out of YY” instead of just “succcess”) it makes the game run with incredible smoothness. (Margins of failure still require calculation, but the system doesn’t use them nearly as often.)

Having concluded that there’s a huge upside to calculating margin of success like this, without further ado I present several different conceptual frameworks that can help you (or someone else) grok the concept:

• Success starts at 00 and grows from there, so the higher you roll the better your success (assuming that you succeed).
• It’s like blackjack: You want to get as close to your target number as possible without going over.
• It’s like The Price is Right: The dice are naming a price and you want that price to be as close to the actual price (i.e., your skill rating) as possible.
  
• Your skill rating is like a gravity well: Successes start far away at 00, but the closer they get to your gravity well the faster they go and the bigger the explosion when you punch that guy in the face.

(For some reason face punching always features heavily whenever I’m teaching a new system to people.)

UNTESTED: MARGIN OF FAILURE

Okay, now that you’ve grokked how Eclipse Phase does margins of succcess, let me strain your credibility by proposing a similar method for handling margin of failure in the system. (This is really just a random thought that occurred to me as I was writing out the above.)

The key point here is that the system (a) rarely cares about margin of failure and (b) when it does, it only cares if you missed by either 30 points or 60 points. (The former are referred to as “severe failures” and in my system cheat sheet I refer to the latter as “horrific failures”, although I don’t believe the rulebook ever gives a formal term for them.)

So the method here is really simple:

• A roll of 70 or less is a severe failure
• A roll of 40 or less is a horrific failure.

The system also has a handful of effects which are determined “per 10 margin of failure”. (For example, shock damage can knock you unconscious for 1 round per 10 MoF.) To calculate that, simply subtract the tens digit of your result from 9. (So if you roll 77, you would be shocked for 9 – 7 = 2 rounds.)

If you’re looking for a conceptual framework, think of failure as emanating from 99 and growing in magnitude. Note, too, that higher is always better with this system: A higher success is a better success; a higher failure is a better failure. What my mind initially tries to interpret as a discontinuity actually makes sense if you just imagine success and failure emanating from opposite ends of the spectrum while the outcome is a linear comparison to your skill rating.