In Numenera cyphers are one-use items scavenged from the ruined technology of an elder age. Often the utility of these items as perceived by the PCs will be only tangentially related to the item’s original function; the ultra-tech equivalent of ripping a laser out of a CD player and using to signal your squad mates. I’ve talked about cyphers at greater length in Numenera: Identifying Items, but the core concept is that they let the PCs do something cool once and then they’re cycled out for another cypher. Examples of cyphers include stuff like:

• A metallic nodule that can be attached to an item, allowing one to telekinetically manipulate it with a paired rubber glove.
• A disk-like device that shoots out paralyzing beams.
• A set of goggles that can be used to perceive out-of-phase creatures and objects.
• A metal amulet that surrounds the user in a field of absolute blackness.
• A cannister dispensing foam that transforms metal into a substance as brittle as glass.

Now that you know what a cypher is, here’s the tip:

To quickly create a memorable bad guy, roll on the random cypher table in Numenera and have the bad guy do that as “their thing.” When the fight’s done, the PCs will be able to scavenge the matching cypher from their corpse.

For example:

• A synthetic octopodal creature that telekinetically hurls large items at the PCs.
• A slender ungulate with “antlers” that are metallic discs emitting paralyzing beams.
• An invisible assassin droid that launches attacks from the out-of-phase interdimensional space where it lurks.

This technique works particularly well in Numenera because, as I discuss in Numenera: Fractal NPCs, creating NPCs in the Cypher System can be literally as easy as saying, “He’s level 3.” But then you can expand that to whatever level of detail you want (hence the “fractal” in “fractal NPCs”), which in this case would be adding the cypher-based ability.

However, you can achieve a similar effect in almost any system by simply grabbing an existing NPC stat block and slapping on the ability. (The only drawback is that scavenging may not be an assumed part of play in your game of choice, so you may lose that depth of experience unless you make a special effort to incorporate it.)

For example, if you wanted to make a tribe of orcs a little special, you might have them worship Glaubrau, Demon of the Nethershade. Take a standard orc stat block and then have them cloaked in absolute blackness. (The PCs can harvest their black blood to make an oily potion that imbues a similar effect.)

I’ve actually found myself using the Numenera cypher lists to achieve this effect across multiple systems, but there are often local equivalents. For example, random potion lists in D&D are somewhat more limited in their range of effect, but can work in a pinch. (If you’re running 5^th Edition, however, you might just cut to the chase by grabbing a copy of Arcana of the Ancients, which adapts a whole slew of cool stuff from Numenera for your D&D game, including some random cypher tables.)

Go to Part 1

SUNBLIGHT

The biggest problem with Sunblight (and you’re going to be hearing me say this a lot) is that the lore just doesn’t hold together.

To start with, there are a couple weird conspiracies in the fortress.

The first is that Xardorok thinks he’s worshiping and following the commands of Deep Duerra (a duergar goddess), but in reality Asmodeus has punked him and is only pretending to be Deep Duerra!

This doesn’t really make a lot of sense and doesn’t really go anywhere. Further research suggests that 4^th Edition killed Deep Duerra, but Sword Coast Adventurers Guide for 5^th Edition just listed her as an active god again. So maybe this was an attempt to clean that continuity up (Duerra is dead, but Asmodeus has been impersonating her)? But mostly my take-away is that Wizards of the Coast really, really, really, really, really, really likes Asmodeus and needs him to be in every single adventure.

In any case, the conspiracy goes hilariously dumb when we get to this bit:

Infernal Tablets. The barbed devil [pretending to be a priest of Deep Duerra] spends its time chiseling on granite tablets to inscribe them with Infernal runes. Characters who can read Infernal script can learn about the devil’s manifesto, in which Klondorn reveals that Asmodeus, in the guise of Deep Duerra, is using the duergar to further his interests.

There are ninety-two granite tablets stacked about his room. Each tablet weighs 50 pounds and is 1 foot wide, ½ feet long, and 2 ½ inches thick.

“I have infiltrated Xardorok’s fortress. As an infernal spy, I will spend my days carving large stone tablets that say I AM A SPY and stacking them in large, clearly visible piles.”

The other conspiracy involves Grandolpha, a rival duergar despot that Xardorok wants to seduce and marry.

She’s not interested. But rather than marrying Xardorok, slitting his throat, and taking over the joint, Grandolpha has concocted a scheme where she suborns the guards one by one, then slits his throat, and takes over the joint.

Eh. Maybe she just enjoys suborning people.

But where I’m really left scratching my head is that Grandolpha orders her guards to unlock the doors of the fortress and let the PCs in.

The “plan” seems to be a vague hope that the PCs will politely kill Xardorok, but leave Grandolpha and her conspirators alone. But Grandolpha has no way of knowing that the PCs are coming. So is it just a standing order to the guard on the front door? “If anybody shows up, let them in on the off-chance they’re here to kill Xardorok! Boy, that sure would be great!”

This central silliness aside, Sunblight is a well-designed set piece. It has a good map with a lot of strategic interest and a good key with a lot of exploratory rewards. It would probably benefit from an adversary roster, but it would be relatively easy to throw one together.

AURIL’S ABODE

This chapter leaves me scratching my head.

It starts off by saying that the characters “might come to this island on their own, hoping to put an end to Icewind Dale’s everlasting winter.” But there’s no meaningful mechanism for that: Other than one random rumor, the only way for the PCs to discover where Auril lives is for Vellynne the Necromancer Ex Machina to tell them.

When I first heard that Auril was going to have three different forms (which was a major element in the pre-release publicity for the campaign), I girded my loins for an epic confrontation of mythological proportions: You can’t defeat a goddess just once! You’ll have to face her three separate times!

But… no. She’s just a video game boss shouting:

Even the setting of the fight isn’t inspiring. It’s Area G8 below, which is a frost giant’s ruined bedroom:

The way it’s scripted is that the PCs walk into the bedroom and Auril “staggers” out of the side chamber. FIGHT!

It’s a strategically boring space lacking in all gravitas.

(If the PCs don’t defeat Auril here, there’s a kind of half-assed bit later where she shows up at the end of the adventure and is like, “HOW DARE YOU SKIP MY FIGHT SCENE? RAWR!”)

The PCs don’t need to defeat Auril to end her winter curse on Icewind Dale, though: If they kill her roc (which is roosting on the roof), she can no longer fly into the sky to perform the ritual each night and the curse ends.

But… uh… Auril can fly.

And it’s D&D, so there’s like a zajillion work-arounds.

Meanwhile, killing her to end the curse doesn’t really make a lot of sense, either: The book is quite clear that if the PCs kill her, she’ll just re-manifest at the next Winter’s Solstice, restored to her full power. We’re later told that she’ll just decide to give up on the whole plan – so no worries! – but there’s no meaningful reason given for her making that decision.

Okay, but what about what brought the PCs to the island in the first place?

First: They’re looking for Nass Lantomir. They will, as far as I can tell, never find her.

The ship Nass took to the island sunk. She swam to the island and then died of frostbite (despite having multiple spells that would have prevented that from happening). The PCs:

• Have no way of knowing she came to the island.
• Have no reason to search the island.

They will, therefore, never find her corpse, which is in the middle of nowhere and mostly buried in snow.

… it’s a pity she’s carrying a magic item that’s fairly integral (albeit not strictly necessary) to the rest of the adventure.

The second reason the PCs have come to the island is to retrieve the Codicil of White, which is a religious primer for Auril’s worshipers. This mostly calls attention to the fact that the worship of Auril is a big ??? in the book and doesn’t make a lot of sense. But, regardless, the Codicil is locked in Auril’s basement and the security system is…

… well.

It’s triply dumb.

Dumb #1: The authors really want you to play through the security system, so we get some amateur hour railroading.

The door is otherwise unopenable, indestructible, and impassable. Any spell cast with the intent of bypassing the door fails and is wasted.

Cool story, bro.

Dumb #2: “This is a very important vault! Only authorized people should be let in! What’s our security system?”

“Only people who pass four tests can enter.”

Uh huh. Where are the tests located?

“Right outside the door of the vault.”

… uh huh.

Dumb #3: These fucking tests.

Here’s how they work: The PCs walk through a magic door and are teleported to a Raghed tribe which is very conveniently facing a crisis perfectly themed to the test (Cruelty, Endurance, Isolation, Preservation) at the very moment that the PCs arrive.

This is not an illusion or fake-out. The book is VERY INSISTENT that this is REALLY HAPPENING.

So, like, the Preservation test has them arrive in a Raghed camp where everyone has been killed except a nine-year-old kid. The PCs have to save him.

Wow.

Lucky thing that literally happened JUST as the PCs opened the magic door, right?

It gets dumber, though, because the PCs explicitly don’t have to all take the test at the same time. So after the first group goes through the test… what happens when the second group opens the same door?

In multiple places the adventure talks about the logistics of taking tests at different times on the VAULT side of the door, but never once explains what the second or third or fourth pass on a given test will look like for the characters actually taking it.

By the way, if you don’t get the Codicil out of the vault, the entire campaign breaks and you can’t play any more. So if the PCs fail the tests, some NPCs stop by and let them in.

…

All in all, Auril’s Abode is the only part of the adventure that I think is actually straight up crap. (Except for the art, which continues to be fabulous.)

This is a pity because this is THE adventure promised by the cover and title of the book.

When I got to this chapter I was literally rubbing my hands together with glee. HERE WE GO!

And then… Meh.

THE LOST CITY OF YTHRYN

Eighteen hundred years ago, Ythryn was a flying city and part of the Netherese Empire. A strange artifact caused the city’s mythallar to malfunction and it crashed into the Great Glacier. It has remained completely sealed off from the outside world ever since.

On that note, I’ve noticed a weird design tic in WotC adventures: Ancient ruins that have supposedly remained undiscovered and unknown for centuries, but which the PCs nevertheless reach by journeying through cosmopolitan caves filled to the brim with people.

Brief segue while we’re in these caves. This room is one of the funniest things I have ever read in a D&D adventure:

H21. FROZEN FROST GIANT

Entombed in the icy floor of this twenty-foot-high cave is the frozen, well-preserved corpse of a frost giant. Scratch marks in the ice suggest a half-hearted attempt to excavate the remains.

Really fantastic imagery here! Frozen ice cave. The creepy giant corpse. It’s super atmospheric. Memorable. Very cool. (Pun intended.) And enigmatic! How did he get here?

A city fell on him.

…

Not making this up. He was crushed to death when the flying city fell on top of him.

And, as we all know, when a city(!) falls on top of you, what happens is that your perfectly preserved corpse ends up flawlessly preserved in ice.

The best part is that if you look at the map, the city actually ended up directly BELOW him. So the city fell on him, crushed him to death (without mangling his body in any way), and then… phased through him to its final resting place?

In any case, the PCs eventually make their way through the caves and end up in the buried city of Ythryn. The book is a little vague on this, but the intention seems to be for the PCs to enter a huge cavern and be able to look down at the ruined city below them.

At which point you show them the Players’ Map of the whole thing:

Then they can explore the city by basically just pointing at cool stuff and saying, “Let’s go check that out!”

This isn’t the only way to run a ruined city scenario, but it’s a pretty great way to do it.

I do have a couple of quibbles here.

First, the scale of the city is such that you can walk across the whole thing in ten minutes, which seems to sap some of the epic scope. To be fair, this in itself is an attempt to compromise with the harsh realities of the 5^th edition core rules: Slow travel pace is 200 feet per minute in the core rules; here the writers invent a “cautious rate” of 200 feet every 5 minutes. On the other hand, the scale also doesn’t seem to sync up well with the individual building maps, so you could probably double the size of the city to no ill effect.

Second, one of my personal pet peeves is when building floorplans don’t match the picture of the building. But this becomes particularly egregious when the intended interaction of the game is literally “point at picture, then go to that place.” For example, the Spire of Iriolarthas:

The inset on this map actually hides the full extent of the problem here. If you look at the Players’ Map above, you can see that this is just the tippy-top of the central structure in the city. Where the heck is the rest of the building?!

But even just looking at the inset is rife with problems: Where are the missing floors that are clearly depicted between the three floors shown on the map? And did you notice the strut clearly connected to the High Court that leads to the other half of the spire?

THE END OF THE CAMPAIGN

In any case, here in Ythryn we have reached the end of the campaign… and it all just falls apart.

There are three options given in the book for how Rime of the Frostmaiden can end.

ENDING OPTION #1: Activate the Reset Obelisk (a powerful artifact in the city) and save Ythryn.

Here’s how that works:

1. The PCs take Iriolarthus’ staff of power.
2. The PCs use it to activate the Obelisk.

Why is this a problem?

Well, Iriolarthas is actually a Netherse wizard who has spent 1,800+ years trying to solve this problem. Despite his “best efforts” he’s been unable to do so, despite having trivial access to everything he needs. (The staff is literally sitting in his office.)

Now, Iriolarthas is currently a demi-lich and that would probably prevent him from doing this himself. But there are other people in the city who would gladly help (and do). Also, he didn’t become a demi-lich for a long, long time. (Which is why there was time for his “best efforts.”)

Also: He’s only a demi-lich because his phyalactery was buried in the ice when the city crashed, so he wasn’t able to physically access it to feed souls into it. (This is not how phylacteries work. As the Monster Manual says, “The phylactery must be on the same plane as the lich for the [soul-feeding] spell to work.”)

But the BIG problem is that:

• The city is stuck because X needs to be done.
• X could have been done at any time.

And it doesn’t make any sense. Nor is it made to make sense for the players.

Maybe the players don’t realize it and it’s fine. But I still think it’s a problem.

And the solution would be easy: The fix to the Obelisk just needs to be an out-of-context element. This would force the PCs to leave the city, get the thing they need (and which the people trapped in the city lacked), and then come back.

This would also help explain some of the timeline issues the finale struggles with, which brings us to…

ENDING OPTION #2: Use the Ythryn mythallar to cast control weather and end Auril’s winter!

The problem here is that this doesn’t actually work:

You can use the Ythryn mythallar to cast the control weather spell without requiring any components and without the need for you to be outdoors. This casting of the spell has a 50-mile radius. For the duration of the spell’s casting time, you must be within 30 feet of the Ythryn mythallar or the spell fails.

This is actually the second time the book has asserted that a magic item that casts control weather will solve the problem. But it doesn’t really hold up, because the spell only lasts for 8 hours and requires the user’s concentration. So unless the PCs want to spend the rest of their lives in Ythryn perpetually casting the spell three times a day, it won’t help.

Also, the 50 mile radius of the spell isn’t far enough to reach all of Ten-Towns.

On the other hand, it’s nice to have an alternative resolution to the Auril’s Winter plot than just killing Auril…

…except not really, because the adventure says that using the mythallar immediately causes Auril to show up and fight them to the death.

So it’s all pointless.

ENDING OPTION #3: Twenty-four hours after the PCs enter Ythryn, Auril shows up with a loudspeaker and says, “YOU HAVE TAKEN TOO LONG! LET’S FIGHT!”

And then she fights to the death. Or whatever.

Mostly I don’t understand the design decision behind the timeline here:

• The PCs reach Ythryn.
• 12 hours later, competing archaeological teams arrive in the city.
• 12 hours after that, Auril shows up, summarily executes the competing archaeological teams, and triggers the endgame.

This seems like an interesting, dynamic situation for the PCs to interact with. (Think about the cat-and-mouse game between Indiana Jones and Belloq in Raiders of the Lost Ark.) But it seems as if the timeline has been set up just to punish PCs who take long rests in the city by giving the “bad ending.”

The other problem here is that Auril’s appearance feels arbitrary. You could trigger this, “THE TIME HAS COME FOR THE END OF THIS CAMPAIGN!” encounter at literally any point in the campaign and it would make as much sense as it does here. It doesn’t really proceed from everything the PCs have been doing, so it doesn’t feel like a culmination or a conclusion. It doesn’t resolve the story, because the PCs’ story hasn’t been about Auril.

You might also note that both Ending Option #2 and Ending Option #3 become null and void if the PCs killed Auril back at Auril’s Abode. This is a problem that often blights adventures in which a pseudo-sandbox gets welded onto the side of a railroad: The author (or GM) basically says, “Look at how much freedom you have! You even have the freedom to screw up the plot I’ve prepared!”

But then, if the players do that, the story doesn’t actually move in a new direction, it just continues shambling through the now empty and meaningless scaffolding of the prepared plot.

THE MISCELLANEOUS MISSES

Now that we’ve reached the end, let’s talk about a handful of other problems I had with the book — the smaller details that haven’t really fit into the discussion so far.

For example, three hundred page adventures need indexes. And if you don’t have an index, you’d better make really, really, really sure that the table of contents doesn’t have errors in it.

On a similar note, there’s a lot of basic continuity errors in the book. In one of the most egregious examples, a key piece of information the players need to finish the campaign is given in two contradictory forms on the same page.

REPUTATION SYSTEM: Given the nature of the Icewind Dale sandbox, the designers clearly realized that it would be really cool to have a reputation system. For example, they write:

Once the characters reach 4^th level, they no long [sic] gain levels by completing the quests in this chapter. Even so, completing more than the required number of quests can improve their standing in Ten-Towns (see “Reputation in Ten-Towns” below)…

Unfortunately, two paragraphs later they say, “Eh. Fuck it.”

The adventuring party’s reputation in Ten-Towns improves as the characters gain levels, with the following results:

When the characters reach 3rd level, they [list of stuff].

When the characters reach 4th level, they [list of stuff].

So you keep gaining reputation even when you’re not gaining levels, but the only way reputation is tracked is by what level you are?

C’mon, man.

Furthermore, the only function of “reputation” is to unlock the Chapter 2 rumor table.

So why not just… do that? As written, it would be best to just drop the reputation “system” entirely.

Which is unfortunate, because a solid reputation system could really heighten the story of Icewind Dale and make it viscerally meaningful: The players would really be able to feel themselves becoming heroes of the Dale, and those mechanics would feed back into the narrative of rising up, overthrowing the Frostmaiden, and leading Icewind Dale out of her horrible winter.

IGNORE THE RULES! There are a lot of places in Icewind Dale where the book says, “The rules say this doesn’t work, so the rules don’t apply.”

For example, the chardalyn dragon is allowed to ignore environmental effects so that it doesn’t fall out of the sky. In another place, the authors want to have a very specific haunted house story, so they just arbitrarily declare that the ghost can’t be turned.

This is often just gratuitous. For example, the dragon would fall out of the sky because it doesn’t have a hover speed. But it’s a completely artificial dragon… why not just give it a hover speed? (Given that it’s kind of a steampunk creation made from evil crystals, I argue that giving it VTOL steam jets would just make it more badass.)

You probably won’t be surprised to discover that I’m not a big fan of this. It’s just not that hard to design stuff that works right with the rules.

BREAKING THE FOURTH WALL: I’ve mentioned this briefly above, but there are actually multiple points in the book where an NPC basically says, “Thou art not yet experienced enough! Return when thou hast attained a higher level!”

It’s kind of gauche. Frankly, I find this sort of, “You need to gain 500 XP before you can access this content!” stuff pretty off-putting even in a video game. In a tabletop RPG? It’s completely unacceptable.

BOXED TEXT: Quite a bit of boxed text in Rime of the Frostmaiden is very bad. Take this example from the caves around Yrkyth:

Frost-covered blocks of stone jut from the floor of this ten-foot-high cave of ice. Perched atop the largest stone is an emaciated kobold with glowing red eyes. It bares elongated fangs as it hisses at you, then scampers away.

First, there are too many places where the action (or inaction) of the PCs are assumed. That’s a big no-no.

But the really odd thing I found here was how often the boxed text wasn’t actually describing the thing it was purportedly describing.

It was here in the caves that I finally figured out why, because the boxed text in the caves is constantly telling you the height of the ceiling but absolutely nothing else about the shape or dimensions of the area.

…

Did you figure it out?

The adventure is written exclusively for GMs using virtual tabletops. The assumption is that the GM won’t bother describing the room because the players can just see the map displayed on their computer screens.

Here’s a more egregious example from earlier in the book:

The room is clearly designed for the PCs to investigate the tool-covered table and the bookcase, but the description of the room doesn’t mention that they exist. And, yes, the answer is that you can see the table and the bookcase on the map.

THE STANDOUT MOMENTS

Remember at the beginning of all this that I said I liked the book quite a bit? But I’ve been analyzing its shortcomings for a few thousand words now, and that can leave a somewhat lop-sided impression. So I’d like to close things out by calling out some of the really nice stuff that the book does.

For example, the book introduces domesticated axe beaks which are used as mounts and pack beasts in Icewind Dale. And they’re awesome:

CUSTOMIZED CHARACTER CREATION: Icewind Dale also does a fantastic job of customizing character creation:

• It personalizes the scenario hooks by tying them to specific Backgrounds.
• It includes customized starting equipment appropriate for the setting.
• It includes a new option for a PC race in the Goliaths. (Even if no one picks the option, its availability still sets a tone.)
• It has a cool set of Character Secrets, most of which are tangibly tied to various elements of the sandbox.

There are all fairly minor things, but that’s kind of what makes them great. It really shows how just a few light touches can make character creation feel unique and special; investing the players in the feel and color of the campaign before play ever begins.

Playtest Tip: I’ve found that the Character Secrets tend to be the definitional aspect of character concept/background. So if you’re using them, you’ll probably want to deal them out up front so that people can build their characters around their secret. You can ameliorate this to some extent by dealing two or three secrets to each player and letting them choose.

WEATHER CONDITIONS: The designers use extreme weather conditions (particularly blizzards) to attempt to reintroduce some of the difficulties (or, more accurately, the interesting complications) of overland travel that 5^th Edition’s core design generally stripped out of the game.

Icewind Dale isn’t really exploration-based, so getting lost in a blizzard, for example, doesn’t contribute to the navigational puzzle which would be found in such games. But it DOES create a dramatic tone: it brings Auril’s curse into the narrative at a fundamental level.

NON-COMBAT SUPPORT: Icewind Dale also does a great job of supporting non-combat resolutions and goals.

In Caer-Dineval, for example, the Knights of the Black Sword push hard for an alliance with the PCs, in a way that feels like it could reasonably happen (at least for awhile).

The Caer-Konig quest that takes you to a duergar outpost is primarily framed around “get our stolen stuff back,” which is a sharp contrast to the usual “clear the dungeon” goal found in a lot of published modules. (This pairs well with the xandered entrance to the outpost, which allows the PCs to either assault the main door or attempt to breach the complex through a bunker.)

The book also generally provides this support while avoiding contingency planning (which is the other typical pit trap scenarios can fall into with this).

For similar reasons, I like the Good Mead quest in which the PCs clear the dungeon… but then, after they’ve done so, a different bad guy shows up while they’re still there, likely inverting their relationship with the dungeon (so that they need to defend what they just invaded).

GREAT GRAPHICAL DESIGN: It’s unsurprising to discover that Icewind Dale: Rime of the Frostmaiden continues Wizards of the Coast’s fantastic graphical design (samples of which you’ve seen throughout this review.)

I’ve already mentioned the fantastic art illustrating even the weaker parts of the adventure, like Auril’s Abode. But this also extends to even the smallest details. For example, I simply adore this magic ring:

I want to own one in real life.

The advantages of Wizard’s graphical design extend beyond the book itself. For example, consider the stunningly beautiful (and incredibly massive!) chardalyn dragon miniature from Wizkids:

These rich graphical resources and other enhancements are a huge value add to your campaign that can be difficult or impossible to replicate otherwise.

These strong graphics are frequently paired to some fantastically flavorful material.

For example, I love the random encounter with Arveiaturace the White Wyrm: “…the dragon is buried under heavy snow. Her [dead rider who she once served and is still strapped to her back] is visible above the surface, looking like a frozen corpse in the snow. If the characters are close enough to touch the corpse, they’re already standing on the dragon’s back.”

That’s so unique, cool, and totally terrifying.

Gotta love it. Can’t wait to run it.

Style: 5
Substance: 3

Story Creator & Lead Writer: Christopher Perkins
Writing Team: Stacey Allan, Bill Benham, H.H. Carlan, Celese Conowitch, Dan Dillon, Will Doyle, Mikayla Ebel, Anne Gregersen, Chad Quandt, Morrigan Robbins, Ashley Warren
Rules Development: Jeremy Crawford, Dan Dillon, Ben Petrisor, Taymoor Rehman
World Building: John Francis Daley, Crystal Frasier, Jonathan Goldstein, Ed Greenwood, Amanda Hamon, Adam Lee, Ari Levitch, Mike Mearls, Christopher Perkins, Jessica Price, R.A. Salvatore, Kate Welch, Shawn Wood

Publisher: Wizards of the Coast
Cost: $49.95
Page Count: 320

Go to Icewind Dale Index

Auril the Frostmaiden has claimed Icewind Dale, laying her enchantment upon it: a terrible curse of perpetual winter. The denizens of Ten-Towns – ten settlements clustered around the lakes at the center of the Dale, nestled between the Spine of the World and the Great Glacier – grow increasingly desperate for a spring which never comes. When the PCs arrive in this gloom-riven land, they will discover that the cold of the wintry north has leeched into the hearts of men. Surrounded by darkness, can they be the flame that rekindles the light of hope?

As the campaign begins, the players are presented with an open world sandbox: They’re largely free to wander through Ten-Towns as they please, helping people and engaging with crises that present themselves in each of the settlements. A bonafide sandbox is unusual (if not unique) among official D&D campaigns, and the town-based structure used by Icewind Dale is intriguing and ripe with possibility.

Unfortunately, like several of the other official D&D campaigns I’ve seen, Rime of the Frostmaiden only flirts with being something innovative and unique before abruptly ripping off its mask and shouting, “Aha! Just kidding! I was a railroad the whole time!”

In this case, at least, the sandbox is at least fairly legitimate for as long as it lasts. But after just four or five levels, the whole thing abruptly collapses down into the linear plot. (Which is, itself, beset with problems.)

With that being said, Icewind Dale: Rime of the Frostmaiden is bursting with a ton of cool stuff. The covers are metaphorically strained with a bevy of good sandbox material, a handful of epic set-pieces, stunning artwork, and fifty pages of chilling new monsters (plus a thematic miscellanea of other useful elements). Before we go any further, I’m just going to say that I like it. The book is not without its shortcomings (we’ll get to those), but I liked it enough that I felt comfortable launching a campaign in the Icewind Dale sandbox without first making major alterations to the material. Which is, for me, fairly high praise indeed.

THE SANDBOX

One of the primary problems with Rime of the Frostmaiden is how poorly it explains its structure. This is particularly true of the sandbox, the explanation of which is both inadequate and filled with vague contradictions that further complicate comprehension. But this is, more or less, how it works:

There are two starting quests. An easy-to-miss feature here is that these quests are tonally distinct from each other — one is tracking a serial killer that ties into the darkness that has seeped into Icewind Dale; the other is a hunt for fanciful fairy creatures called chwingas. Both, however, thematically tie into the over-arching campaign: The Frostmaiden is a dark and feyish goddess, and the two starting quests reflect that from different angles. (The DM can also use this to dial in the tone they want for the campaign, by choosing or emphasizing one quest over the other.)

In any case, these starting quests are designed as light framing devices that will motivate the PCs to move from one town to another. Each town has an additional quest keyed to it and opportunities to pick up rumors about quests in other towns. So as the PCs journey around, they will collect additional quests and most likely begin doing those while continuing to work towards accomplishing their original quests.

The starting quests thus effectively provide a default action for the beginning of the campaign: If the players have any doubt about what they should be doing next, they can simply go to a new town and look for their starting quest item (the serial killer or the chwinga).

Once the PCs have accomplished a certain number of quests, they level up and effectively “unlock” an additional rumor table (which the book confusingly refers to as “tall tales,” despite the fact that they are completely reliable sources of information) that will begin pointing the PCs towards what I’m going to call the Chapter 2 quests (because that’s where they’re described in the book). These quests are, obviously, more difficult; they also tend to take the PCs further out into the wilderness around Ten-Towns.

Now, there are some caveats here.

First, Rime of the Frostmaiden instructs the DM to only use one of the starting quests. This can be an option for aesthetic reasons, but objectively speaking it’s almost certainly wrong: The structural function of the starting quests, as noted, is to provide a default action that will move the PCs through several towns. Both of the quests, however, can end essentially at random (i.e., the players choose the correct town, find the thing they’re looking for, and the quest ends). You want both quests on the table to provide redundancy. Having both quests in play will also deepen the default interaction with each community (because they need to look for multiple things).

Most importantly, however, giving the PCs two different quests to simultaneously pursue at the start of the campaign will immediately break the players of the expectation that they’re going to be doing a linear set of assigned tasks.

The second caveat is that the way Icewind Dale handles rumors is mostly wrong. The advice in the book for using the rumor tables is radically inconsistent, and there are simply too many places where the DM is told to spoonfeed rumors to the PCs one at a time. This is more or less the exact opposite of what you should actually be doing.

See, in a sandbox campaign you want LOTS of rumors to be in play at any time. The essence of a sandbox campaign is that the players have the ability to choose or define what their next scenario is going to be. In order for that to work, the PCs need to be in an information-rich environment and rumors are, broadly speaking, how you accomplish that. If you spoonfeed them the rumors (i.e., scenario hooks) one at a time, you’re choking the life out of your sandbox. (See Juggling Scenario Hooks in a Sandbox for a longer discussion of this in detail.)

To be fair, there are other sections of the book that correctly tell you to be profligate with the rumors. But even some of this advice can go awry. For example, there’s one passage where it suggests that the GM should just randomly give players rumors out of the blue. (“Lo! From the heavens I give unto you… a rumor!”) The reality is that the acquisition of rumors should flow from the actions taken by the PCs.

(Which is also why the investigative action taken in each town related to the starting quests should, generally speaking, explicitly trigger the local rumor table.)

These are regrettable shortcomings in the book (particularly if one imagines it being used by a neophyte GM), but not seriously debilitating as you can, for the most part, simply ignore the bad instructions on how to use the material.

The third caveat, however, deals with the actual adventure material itself. While many of the sandbox adventures presented in Icewind Dale are well done, there are too many that miss the mark. (And some badly so.)

For example, consider the starting quest in which the PCs need to track down a serial killer named Sephek. The scenario hook for this quest is one of the worst I’ve ever read. It’s comically bad.

Hlin has taken it upon herself to investigate the recent murders because no one else – not even the Council of Speakers – can be bothered. Hlin is studying the characters closely, trying to decide if they’re worth her time. Ultimately, she takes the chance and draws them into conversation, asking them to help her take down her only suspect: a man named Sephek Kaltro. Here’s what she knows about Sephek and the victims:

“Sephek Kaltro works for a small traveling merchant company called Torg’s, owned and operated by a shady dwarf named Torrga Icevein. In other words, Sephek gets around. He’s charming. Makes friends easily. He’s also Torrga’s bodyguard, so I’m guessing he’s good with a blade.

“His victims come from the only three towns that sacrifice people to the Frostmaiden on nights of the new moon. This is what passes for civilized behavior in Icewind Dale. Maybe the victims found a way to keep their names out of the drawings and Sephek found out they were cheating, so he killed them. Maybe, just maybe, Sephek is doing the Frostmaiden’s work.

“I followed Torg’s for a tenday as it moved from town to town. Quite the devious little enterprise, but that’s not my concern. What struck me is how comfortable Sephek Kaltro looked in this weather. No coat, no scarf, no gloves. It was like the cold couldn’t touch him. Kiss of the Frostmaiden, indeed.

“I will pay you a hundred gold pieces to apprehend Sephek Kaltro, ascertain his guilt, and deal with him, preferably without involving the authorities. When the job is done, return to me to collect your money.”

“Hello. Yes. I would like to pay you 100 gold pieces to kill this random person because he’s a serial killer… Well. Maybe. Who knows, really? I have no actual evidence he’s a killer. Hell, I don’t even know if he was actually in the same towns where the killings happened. But maybe.. just maybe, the killer is working for the Frostmaiden. Or maybe not. But if the killer IS working for the Frostmaiden, then MAYBE he would be immune to the cold. And this guy doesn’t wear a coat. So… yeah. Definitely the killer.”

That’s pretty dumb. But then it gets worse:

Quest-Giver: I followed Torg’s caravan for ten days.

PCs: So where is it now?

Quest-Giver: No idea.

The quest-giver followed the caravan for ten days, became convinced someone in the caravan was the killer, and then… left and went to a completely different town? So they could sit in a tavern and stare at strangers until randomly deciding which ones they would ask to go kill someone on their laughable say-so?

The crazy thing is that the quest-giver is completely right: The Frostmaiden is sending Sephek to kill people who bribe officials to take their names off the lottery list.

But… only three of the Ten-Towns even do the human sacrifice thing. If the Frostmaiden wants to kill people for not being available as sacrifices, why isn’t Sephek targeting the OTHER seven towns?

There are a number of similar head-scratchers strewn about Rime of the Frostmaiden, but there’s also a fair number of scenarios that are just badly designed.

For example, in the town of Bremen there’s some sort of murderous creature in the lake that’s attacking the local fishing boats. The PCs are supposed to grab a boat, head out onto the lake, and deal with the creature. (It’s an awakened plesiosaurus.) Here’s how the scenario works:

1. Get in the boat.
2. Roll on a random table of events until the plesiosaurus shows up.

The structure itself is obviously lackluster, but it’s made worse because (a) none of the results on the table are actually interesting and (b) they simply repeat until you roll the magic numbers on the d20 that end the misery.

Here are the actual rolls I made while simulating how this scenario would play out:

• Knucklehead trout hits you in the head.
• Knucklehead trout hits you in the head.
• Knucklehead trout hits you in the head.
• Nothing happens.
• Targos fishing boat shows up, then leaves.
• Nothing happens.
• Nothing happens.
• Nothing happens.
• Targos fishing boat shows up, then leaves.
• Plesiosaurus shows up.

It’s here! Finally!

… but then there’s a 1 in 3 chance that it just leaves before the PCs can interact with it! Which is, in fact, what I rolled. So then:

• Nothing happens.
• Knucklehead trout hits you in the head.
• Nothing happens.
• Nothing happens.
• Plesiosaurus comes back!

(It should be noted that each of these checks takes an hour, so the party also presumably went back to town and slept somewhere in there.)

Imagine running this at the table!

And yes, obviously, any DM worth their salt isn’t going to actually do this. But that just raises the question of why it was written this way in the first place, doesn’t it?

And then there’s Tali, the quest-giver. They’re a raging asshole.

Before the quest they say: “Can you go out on the lake and take notes on a dangerous beast that’s killing people?”

Then after the quest they say: “Oh! You’re back? In payment, here’s a potion that makes dangerous beasts friendly so that they won’t kill you.”

(I acknowledge that the potion wouldn’t work because the plesiosaurus has been awakened, but THEY don’t know that.)

To be clear, there are many quests that don’t have problems like this. (The book has more than twenty of these sandbox quests.) But there are, frankly, too many that do — like the dwarves who offer to pay the PCs more to retrieve a shipment of iron than the iron is worth; or the bad guys who captured a castle so they could live there and immediately dumped corpses into the water supply; or the one where all the NPCs are baffled about how they can find the bad guys, so they take the PCs to the tracks that the bad guys left in the snow and scratch their heads until one of the PCs says, “Maybe we could… follow the tracks?”

Speaking of tracks, it turns out that A LOT of the quest hooks in Icewind Dale are based around following tracks. Fail the Wisdom (Survival) check to follow the tracks? Guess you fail the quest! Sucks to be you! (And it’s frequently worse than this because the designers are actively sabotaging the already fragile structure. For example, there’s one quest where the PCs can fail to successfully follow the required tracks because they did so at the wrong time of day. And another where they follow the tracks of some thieves half way to their goal, but then the tracks automatically blow away in the wind, forcing the PCs to make blind Wisdom (Survival) checks to search the hills for… well, they don’t actually know WHAT they’re looking for, but if they find anything it will DEFINITELY be where the thieves are, right?) This is sort of okay in a sandbox like Rime of the Frostmaiden because you don’t need to succeed at every quest, but it’s still not great scenario design… particularly if you’re doing it over and over and over again.

Most of the material that’s compromised like this is, ultimately, salvageable. But you will need to put in the work to salvage it.

TRANSITION TO LINEAR

Let’s talk about the dragon in the room.

There comes a point in the campaign where the PCs have tracked down Sunblight, the duergar fortress in the mountains at the southern end of the Dale. As they approach the fortress, the chardalyn dragon that the duergar have been constructing out of magical, evil crystals flies out of the top of the fortress and begins winging towards Ten-Towns.

This is an incredibly epic, awesome moment.

It is also the point where the Icewind Dale sandbox begins collapsing into a linear scenario. This is, in my opinion, a bad decision. It seems to kick in at exactly the moment where, in my experience, the most interesting emergent gameplay in a sandbox will start appearing. In other words, stuff is just going to start getting awesome when the campaign abruptly says, “Eh. Fuck it. Let’s do something else.”

But beyond that, it’s just poorly done.

The intention is that the PCs have a choice: They can finish climbing up to Sunblight and assault the fortress OR they can climb down the mountain and race back to Ten-Towns to save it from the dragon.

Except the choice doesn’t actually work because the PCs have no way of understanding the stakes: It’s extremely unlikely that they’ll know what the dragon is going to do. “Dragon flying away from fortress” doesn’t auto-translate to “it’s going to destroy Ten-Towns.” I’d argue it’s far easier to read that as a signal to “hit the fortress fast before it gets back.” So there’s an opportunity for a cool choice here, but it’s a missed one.

But let’s assume, for the sake of argument, that the players do understand the choice: There’s a dragon flying to Ten-Towns and they have to catch it!

… except they can’t. It turns out the choice to chase it or not is irrelevant. If you use the travel times listed for the PCs and the dragon in the book, the decision to raid Sunblight is incredibly unlikely to have any impact on the outcome in Ten-Towns. And, in fact, the overwhelmingly likely outcome in Ten-Towns is that all ten towns will be destroyed except for Bryn Shander and maybe Targos. (Despite this, the book gives no guidance at all for what the post-apocalyptic Ten-Towns is going to look like.)

So the choice doesn’t work. But whether the PCs assault Sunblight first or not (and we’ll come back to Sunblight in a moment), when they come back down the mountain they find a woman waiting for them with enough dog sleds for all of them. She wants them to help her do a job and she offers them a lift back to Ten-Towns.

First: This entire setup is just inherently awkward in its execution. The players decide to head back to town and an NPC pops up in the middle of the wilderness to say, “Hey! Need a lift?”

It’s just silly.

Second: The entire back half of the book absolutely requires that the PCs hitch a lift and work with this NPC. (The book tells you this explicitly multiple times. As far as I can tell, it’s not wrong.) This is an incredibly weak structure to hang an entire campaign on!

But it’s actually worse than that, because this NPC is clearly a necromancer: She’s got undead minions hanging out with her and everything. That’s a problem, because I think it’s overwhelmingly likely that the PCs are going to kill her without a second thought.

I was interested to see what other people thought of this, so I ran a poll on Twitter. 190 people voted on the most likely action the PCs would take, and two-thirds said she was toast… along with the rest of the campaign.

It doesn’t help that even if the PCs do talk to her, she makes it clear in her pitch that she’s a member of the Arcane Brotherhood… who have been repeatedly established in the first half of the campaign as unrepentant bad guys who betray everyone who works for them.

And, again, the PCs have to agree to work with the necromancer ex machina or the rest of the campaign can’t happen. Here’s how that works:

• The necromancer gives the PCs a ride back to Ten-Towns.
• After they deal with the chardalyn dragon, she tells them that she has a cool job for them, but they’re going to have to level up first.
• Once the PCs have leveled up, she leads them to Auril’s Abode, where they have to steal a lorebook.
• The lorebook has a spell which will let them access the lost city of Ythryn.
• The necromancer leads them to the location of Ythryn and casts the spell.

In short, once the PCs choose to go to Sunblight they trigger a sharp transition from the sandbox to a linear sequence of set pieces:

• Destruction’s Light (chasing the chardalyn dragon, which we’ve already discussed)
• Sunblight Fortress
• Auril’s Abode
• The Lost City of Ythryn

For the rest of this review, we’ll be looking at each of these set-pieces in detail.

Go to Part 2

Over the past couple decades, a design concept that has become fairly entrenched in D&D culture is that the PCs need to face X encounters of Y difficulty per day. The general idea being that the game is balanced (either intentionally or unintentionally) around their resources being chewed up across multiple encounters: If they face fewer encounters, there’s no challenge because (a) they will still have lots of resources left at the end of the day (thus suffering no risk) and/or (b) they burn up LOTS of resources per encounter (making those encounters too easy).

There is obviously a kernel of truth here, but there are also, frankly speaking, A LOT of problems with this design ideology. But that’s somewhat beyond the scope of this article. What I’m mostly interested in focusing on today is one specific element of game play that becomes a really problematic dilemma when combined with this design ideology:

Wilderness encounters.

See, the basic assumption is that the design of Dungeons & Dragons should be finetuned around X encounters of Y difficulty per day in which the value of X reflects the number of encounters in a typical dungeon. There are some problems with this idea that dungeons should be designed as a one day excursion, but laying that aside, this makes a lot of sense: The dungeon is the assumed primary mode of D&D play; therefore the game should be balanced around dungeon adventures.

But, of course, the density of encounters in a dungeon is inherently much higher than the density of encounters in a vast wilderness. Furthermore, even if you increased the density of wilderness encounters, the result would be to turn every wilderness journey into an interminable slog. If you assume that

X = a dungeon’s worth of encounters = one day’s worth of encounters

then a wilderness journey that took ten days would “logically” have ten times the number of encounters in a typical dungeon.

You don’t want to do that, so you make one wandering monster check per day in the wilderness: The players know that they’ll have at most one “dangerous” encounter per day, so they just nova all of their most powerful abilities and render the encounter pointlessly easy.

So what do you do?

Do you just strip wilderness encounters out of the game entirely? They’re pointless, right? But to delete an entire aspect of game play (and something that routinely crops up in the fantasy fiction that inspires D&D) feels unsatisfactory.

Perhaps you could artificially pack one particular day of wilderness travel with X encounters, turning that one day into a challenging gauntlet for the PCs. But that’s really hard to justify on a regular basis AND it still has a tendency to turn wilderness travel into a slog.

For awhile there was a vogue for trying to solve the problem mechanically, primarily by treating rest in the dungeon differently from rest in the wilderness. In other words, you just sort of mechanically treat one day in the dungeon as being mechanically equivalent to ten days (or whatever) in the wilderness. The massive dissociation of such mechanics, however, could obviously never be resolved.

Long story short, the dynamic which has generally emerged is for wilderness encounters to be REALLY TOUGH: Since the PCs can nova their most powerful abilities when facing them and there is no long-term depletion of party resources (including hit points), it follows that you need to really ramp up the difficulty of the encounter to provide a meaningful challenge.

In other words, dungeons are built on attrition while wilderness encounters are deadly one-offs.

There are several problems with this, however.

First, it creates a really weird dynamic: Instead of being seen as dangerous pits in the earth, dungeons are where you go to take a breather from all the terrible things wandering the world above. That doesn’t seem right, does it?

Second, the risk posed by these deadly one-off wilderness encounters is unsatisfying. In a properly designed dungeon, players can strategically mitigate risk (by scouting, retreating, gathering intel, etc.). This is usually not true of wilderness encounters due to them being placed either randomly (i.e., roll on a table) and/or arbitrarily (“on the way to Cairwoth, the PCs will encounter Y”).

(A robustly designed hexcrawl can mitigate this, but only to some extent.)

Finally, this methodology – like any “one true way” – results in a very flat design: Every wilderness encounter needs to push the PCs to their limits and thus every wilderness encounter ends up feeling the same.

So what’s the solution?

THE FALSE DILEMMA

The unexamined premise in these attrition vs. big deadly encounters vs. skip overland travel debates is often the idea that challenging combat is the only way to create interesting gameplay.

Partly this is the assumption that all random encounters have to be fights (and that’s a big assumption all by itself). But it’s also the assumption that if the players easily dispatch a group of foes that means nothing interesting has happened because the fight wasn’t challenging.

Neither of these assumptions is true.

Let’s back up for a second.

You’ll often hear people say that they don’t like running random encounters because they’re just distractions from the campaign.

Those people are doing random encounters wrong.

The “random” in “random encounter” refers to the fact that they’re procedurally generated. It doesn’t mean that they’re capricious or disconnected from their environment. I talk about this in greater detail in Breathing Life Into the Wandering Monster, but when you roll a random encounter you need to contextualize that content; you need to connect it to the environment.

And because the encounter is connected to the environment, it’s meaningful to the PCs: Creatures can be tracked. They can be questioned. They can be recruited. They can be deceived. The mere existence of the encounter may actually provide crucial information about what’s happening in the region.

Random encounters — particularly in exploration scenarios — are often more important as CLUES than they are as fisticuffs.

They can also be roleplaying opportunities, exposition, dramatic cruxes, and basically any other type of interesting scene you can have in a roleplaying game. As such, when you roll a random encounter, you should frame it the same you would any other scene: Understand the agenda. Create a strong bang. Fill the frame.

Once you recognize the false dilemma here, the problem basically just disappears. If the PCs have camped for the night and a group of orcs approaches and asks if they can share their campfire… do I even care what their challenge rating is?

If I roll up the same encounter and:

• The PCs subdue and enslave the orcs.
• The PCs rescue the slaves the orcs were transporting.
• The PCs discover the orcs were carrying ancient dwarven coins from the Greatfall Armories, raising the question of how they got them.
• The PCs follow the orcs’ tracks back to the Caverns of Thraka Doom.

Does it matter that the PCs steamrolled the orcs in combat?

I’m not saying that combat encounters should never be challenging. I’m just saying that mechanically challenging combat isn’t the be-all and end-all of what happens at a D&D table. And once you let go of the false need to make every encounter a mechanically challenging combat, I think you will find the result incredibly liberating in what it makes possible in your game.

Go to Part 1

Once I came to see scenes as an emergent property of game play (rather than something that’s prepared), I began to realize that scenario prep is almost entirely about designing tools. Or, perhaps more accurately, toys. Running an RPG is about picking up these toys and actively playing with them, just like the players are actively playing their PCs.

In creating these tools you’re still thinking about how they’ll be used at the table. But it’s more like designing a hammer (“I know this tool will be useful for hammering nails”) than it is writing an IKEA instruction manual (“here is how the scene with Diego will work”).

At the table, players are going to say things like, “Let’s build a chair!” And the GM will be thinking, “Great! I’ve got a hammer, a bunch of wood, and some… screws? Well, let’s make it work.”

It may only reflect my predilections as a GM, but I’ve found over time that node-based scenario design seems to inherently lend itself to a naturalistic form of prep. I’m not sure if that’s the ethos of active play pushing node-based design in a particular direction or if it’s that node-based design has naturally fostered active play. Or both.

But, in short, the more comfortable I become with node-based scenario design, the more I find that the design arises out of the diegetic organization of the game world itself.

For example, when I’m designing something like the Lytekkas hypercorp, the nodes I design (and how those nodes are organized) overwhelmingly tends to mirror the organization of Lytekkas in the game world. (You can see this in the earlier examples, which were based on things like the geography of the game world or the corporate divisions of Lytekkas.)

When I’m struggling with node-based design, a trick I’ve learned is to pick an NPC and ask, “How would they see this stuff? How would they use it? How would they interact with it?” Whether that’s a hypercorp CEO in an orbital penthouse or a Triad 49er on the rough streets of Hong Kong, the in-game perspective is almost always illuminating.

Of course, this makes sense, doesn’t it?

If node-based scenario design is based on the flow of information – if the structure of a node-based scenario is made up of the paths formed by the information – then it follows logically that node-based scenarios will mirror how that information naturally flows through the game world. And, of course, that flow will be determined by the actual relationships between people, places, organizations, and events in the game world. In how the nodes are related to each other.

Thus, as you design your scenarios, you naturally break the game world apart into discrete chunks. Each chunk (or node) is either a toy or a collection of toys that you can pick up and play with, while the connections between nodes show how those toys are related to other toys, making it easy for them to be endlessly combined and recombined in actual play.

And that’s node-based scenario design.

FURTHER READING
Using Revelation Lists
Game Structures
Hexcrawls
5 Node Mystery
Gamemastery 101