If eladrin are the elf-kin with the strongest remaining connection to the Feywild, shadar-kai are those whose nature has been shaped by the grim Shadowfell. Three types are described in Mordenkainen’s Tome of Foes: the shadow dancer, gloom weaver and soul monger.
Shadow dancers are shock troops optimized for operations in darkness. They can function in dim light, but bright light hobbles them severely, so they’ll never willingly choose to fight in daylight or comparable illumination. Dexterity is their one outstanding ability, which they rely on for both offense and defense, and they’re proficient in Stealth, predisposing them to ambush.
Although they’re tough, their stat block is short, not especially complex and mostly passive. Their two standout features are Shadow Jump and Spiked Chain (which they can use three times as a Multiattack action).
Shadow Jump is a mobility feature and action economy enhancer that lets shadow dancers teleport from one dark or dimly lit point to another up to 30 feet away. Depending on the environment and the positioning of combatants, they can use this trait either to engage in melee or to disengage from it. Shadow Jumping to engage is a more desirable tactic when fighting in total darkness, as we’ll see in a moment.
Spiked Chain is a melee attack with a 10-foot reach and a few different riders that can be invoked if the target fails a DC 14 Dexterity saving throw. The first option grapples and restrains the target (a shadow dancer can entrap only one target this way at a time), the second knocks it prone, and the third deals a heap of necrotic damage.
The most effective way to use these options begins with grappling. Restraining the target confers advantage on subsequent attack rolls as well as disadvantage on Dex saves, so if the shadow dancer initially uses Spiked Chain to grapple and the target fails their save, its chances of knocking the target prone or dealing the necrotic damage become significantly better. A shadow dancer always chooses the grappling option if it lands a hit on the first attack of its Multiattack.
Once the target is already restrained, however, knocking them prone is a bit redundant. Combining the grappled and prone conditions can be very effective, as the fact that being grappled reduces the target’s speed to 0 makes it impossible for them to get up again. But the prone condition doesn’t offer anything that the restrained condition doesn’t—in fact, it offers a small measure of protection against ranged attacks. The only reason for the shadow dancer to choose the proning option over the necrotic damage option against a restrained target is if the target has a strong chance of escaping from the grapple, so that even if it does, it’s still on the ground and has to spend movement to get up. Unfortunately for the shadow dancer, its Intelligence and Wisdom aren’t high enough for it to have any criterion for judging a target’s likelihood of escape beyond “looks strong.”
The best possible outcome for the shadow dancer occurs when it lands hits on every one of its three attacks, its target fails all three Dex saves, and it chooses the grapple option the first time and the necrotic damage option the second and third. The value of the grapple option declines significantly relative to the value of the necrotic damage option if either the shadow dancer misses on its first attack or its target makes that save. Even if its second attack hits and the target fails their save, the fact that the shadow dancer didn’t get to apply a rider on the first attack makes it more likely that it won’t get to apply one on the third, even with the advantage of having a restrained target. Thus, it becomes preferable to go for the sure thing—the necrotic damage—on both the second attack and the third, unless the target looks like someone who only succeeds on Dexterity saves by dumb luck. (If the shadow dancer has many non–shadow dancer allies on the field, it may go for the grapple anyway so that they gain advantage on their attack rolls.)
If the shadow dancer has advantage on the first of its three attacks, its total expected damage increases by roughly 50 percent, so starting its turn hidden is important. At the outset of combat, it can use its Stealth proficiency to make its first attack while unseen, but once combat is under way, it needs to attack from darkness as much as possible—and even this doesn’t help it against targets with darkvision. To avoid being spotted by them, a shadow dancer has to spend actions (and thus turns) Hiding, which is simply too slow. Therefore, opponents with darkvision are to be taken out in the initial ambush if possible.
If you use the optional Facing rule (Dungeon Master’s Guide, chapter 8), you can give the shadow dancer a little more leeway, allowing it to Shadow Jump beyond a target’s peripheral vision and gain advantage if and only if it can win a Stealth-vs.-Perception contest to avoid being seen and heard before immediately attacking (akin to how I treat phase spiders’ Ethereal Jaunt). If you play it this way, to be fair, the shadow dancer mustn’t do anything between Shadow Jumping and attacking—no movement, not even a free interaction. And even so, it’s gaining its unseen-attacker advantage from positioning, not from obscurity; as soon as it attacks, its darkvision-enabled target sees it.
Shadow dancers’ reach and ability to grapple gives them a strong incentive always to attack from 10 feet away, so that their grappled targets can’t even reach them without a long weapon. Simple visibility gives them an incentive to attack from darkness, and Shadow Jump gives them a way to place themselves in such a spot just before attacking. Therefore, if a shadow dancer needs to use Shadow Jump at the start of its turn to reach its desired target, it does. But if it doesn’t use it at the start of its turn, it may also use it at the end—say, after a particularly successful Multiattack that obviates its need to continue to grapple and restrain its target.
Most of this analysis assumes that shadow dancers are fighting alone, or alongside other shadow dancers. In mixed company, they’ll fight the same way if they’re taking the lead. If they’re fighting in more of a supporting role, though, they’ll deploy their Spiked Chains in service of their allies’ needs—which could result in such madcap sequences as using Spiked Chain to knock one melee-engaged enemy prone, then another, then to grapple a third, Shadow Jumping and moving normally in between to travel from foe to foe.
Gloom weavers, with their exceptional Dexterity and Charisma, are spellcasters—specifically, according to the mechanics of their spellcasting, sorcerer-warlock hybrids. A curious synergy between two of their features, Burden of Time and Misty Escape, makes them willing and able to get closer to their enemies than spellslingers normally prefer to be: Burden of Time gives them an incentive to allow foes to approach within 10 feet, while Misty Escape lets them turn invisible and teleport away as a reaction
whenever they take damage once after taking damage, which changes things—see below.
They have two sets of spells: their innate set, which are all cast at their base level, and their pact magic set, which are cast at 5th level, except for the cantrips.
- Arcane gate sets up a teleportation portal with a 500-foot range which is persistent and lasts for up to 10 minutes, with concentration. It’s not a subtle spell, and not as useful for a quick getaway as dimension door or plane shift, since enemies can jump through the portals as well. As a method of initiating combat, it does allow a large squad to storm in at once, but its conspicuousness forfeits any element of surprise. A more interesting application of this spell, I think, is to allow a squad of shadar-kai to infiltrate a position from an unexpected direction, such as by gating up to the ramparts of a castle.
- Bane is a dependable debuff, but one that requires concentration and that the gloom weaver can’t boost beyond 1st level. It seems like a poor use of the gloom weaver’s time and concentration; the opportunity cost is much too high, given how powerful its other spells are.
- Compulsion can be used in a variety of ways: to pull an enemy within range of Burden of Time, to push away an enemy who’s already that close, to drive an enemy within reach of a shadow dancer’s Spiked Chain, to cause an enemy engaged in melee to provoke one or more opportunity attacks, or—and this is kind of brilliant—to compel an enemy to march through another gloom weaver’s arcane gate, whereupon the spell can be immediately dropped, trapping the target 500 feet away. (A single gloom weaver can’t pull this stunt, because both arcane gate and compulsion require concentration, but two can. Also note that compulsion can’t make a target “move into an obviously deadly hazard,” but the portals of arcane gate are filled with opaque mist, so it’s never obvious what’s on the other side.)
- Confusion is useful against at two or more clustered enemies (cf. Targets in Area of Effect table, Dungeon Master’s Guide, chapter 8), preferably ones with Extra Attack or some other feature that makes their action economies more robust. Also requires concentration.
- True seeing doesn’t require concentration and is worth switching on when the opposing side is making mischief with such spells as invisibility/greater invisibility, darkness, blink and phantasmal force, or before combat if a gloom weaver has good reason to expect that its opponents are going to.
- Arcane eye is a surveillance tool, largely inappropriate for combat but very useful for pre-combat reconnaissance. The one stunt a gloom weaver may be able to use it to pull off during combat is to cast blight at a creature it couldn’t otherwise get at, but it has other spells (e.g., chill touch, hypnotic pattern) that it can cast at any point within range without having to see its target, so this trick doesn’t seem to offer much net benefit.
- Mage armor is a no-brainer; unless it’s surprised, a gloom weaver doesn’t enter combat without this spell cast upon itself already. If it’s caught unprepared, a gloom weaver casts it after its first use of Misty Escape. (To be honest, since it can cast this spell at will without spending a spell slot, and since it lasts 8 hours per casting, I’m not sure why gloom weavers wouldn’t just keep it up all the time, making its casting into a daily ritual routine like the Muslim prayer schedule.)
- Speak with dead is a cool thing the gloom weaver can do in a social interaction encounter, but the only purpose I can see it serving in combat is to try to interrogate a corpse that the player characters, for whatever reason, want not to be interrogated—providing the reason for combat to take place.
- Armor of Agathys, like mage armor, is common-sense pre-combat prophylaxis, but the catch is that unlike mage armor, it requires a spell slot—and gloom weavers don’t have many of those. We’ll have to compare it to the gloom weaver’s other spells to decide whether it’s worth the opportunity cost.
- Blight is a big gun with one bullet, one that deals an average 40 necrotic damage on a failed saving throw and 20 on a success. If you figure a 50/50 chance of making the save (which would be true for an opponent with a +5 Constitution save modifier), that’s an expected 30 points of damage, which the gloom weaver can crank up to 35 by casting it at an opponent affected by Burden of Time. Since blight has only a 30-foot range anyway, why not reserve it for use against opponents 10 feet away or closer?
- Darkness has the two pesky drawbacks that even creatures with darkvision can’t see through it and that a creature within its target area can simply walk out of it. At least having true seeing active makes the first of those drawbacks go away—for you, not for the rest of your side.
- Dream has a 1-minute casting time. Pass.
- Invisibility cast only on the gloom weaver itself would be a trap choice, since invisibility comes free with Misty Escape. No, the real benefit of this spell comes in its scaling: Since the gloom weaver casts it at 5th level, it can affect itself plus four other allies, giving them a way to initiate combat with a volley of unseen attacks.
- Fear doesn’t scale and causes opponents to run away rather than get hurt and die, so aside from the possibility of provoking opportunity attacks, where’s the benefit? In causing targets to drop their magic weapons so that the gloom weaver and its allies can thief them up, that’s where. But to make this spell worthwhile, the gloom weaver really needs to invoke that Burden of Time effect to impose disadvantage on their saves. Requires concentration, but once the loot’s been scooped up, there’s no particular need to keep sustaining it, although it does impose disadvantage on affected targets’ attack rolls.
- Hypnotic pattern has a long range, is good against three or more opponents (again, cf. Targets in Area of Effect) and is useful against the same kinds of enemies as confusion It doesn’t scale, but it shuts affected targets down completely, which is especially good if they have Extra Attack or some other action economy–enhancing feature.
- Major image requires concentration, costs a spell slot and provides no combat benefit, only a bit of deception. A gloom weaver should only ever cast this spell to prevent an encounter from taking place.
- Contact other plane falls into the same category as dream and speak with dead: cool stuff the gloom weaver can do that has nothing to do with combat.
- Vampiric touch is double-edged, because it’s a melee spell attack with only a 5-foot reach. If the gloom weaver is that close to its target, it would rather be casting spells that require saving throws, upon which Burden of Time will impose disadvantage. Also, even scaled, the damage is lackluster in comparison with blight, and the hit points restored are equal to only half the damage dealt. If they were equal to the full damage dealt, it might be competitive with blight, but they’re not.
- Witch bolt looks decent at first: a ranged spell attack, good against targets between 15 and 30 feet away, sustainable with concentration, scaled up to deal 5d12 lightning damage per turn—legitimately competitive with blight, which is one-and-done. But there are a couple of drawbacks. If the attack misses, no damage is dealt, and the spell slot is lost—a steep cost for a warlock. And the gloom weaver may not get to sustain it past one round: if the target wheels on the gloom weaver, closes to melee distance and lands a weapon hit,
the gloom weaver’s smart play is to react with Misty Escape. Now, technically, rules as written, the witch bolt can remain in effect if the gloom weaver’s concentration isn’t broken and it teleports 30 feet or less away, and the gloom weaver can even remain invisible, because witch bolt, when sustained, “deal[s] damage to the target automatically,” so the gloom weaver isn’t attacking or casting a spell. But this is an affront to common sense, and besides, if the gloom weaver remains that close, its target is just going to run it down again. Like the dog in the joke that finally catches the car it’s been chasing and doesn’t know what to do with it, witch bolt can leave the gloom weaver in the situation of regretting that it got what it was trying to get. Tl, dr: Blight is better.[Gah. Forget all this. The real risk is losing concentration upon a melee hit, and the real weakness is that witch bolt deals boosted damage only when it’s first cast—to the extent that it may not even be worth maintaining, and certainly isn’t worth the cost of a warlock’s spell slot.]
- Chill touch is a ranged spell attack with very long range which doesn’t cost a spell slot. But why deal 3d8 necrotic damage when you can cast . . .
- Eldritch blast and deal
3d10 + 43d10 + 12 force damage with an implicit Agonizing Blast invocation, at the same range, also without spending a spell slot? Plus, since each beam requires a separate attack roll, if you don’t hit with all three, you can still at least hit with one or two. Chill touch is all-or-nothing.
- Minor illusion and prestidigitation are just playing around. When combat starts, playtime’s over.
Taking all the pros and cons into consideration, the gloom weaver’s plan for its spell slots looks like this: one slot for invisibility if the gloom weaver and its allies plan to initiate combat while invisible; one slot for armor of Agathys, pre-cast unless the gloom weaver wasn’t expecting a fight; remaining slot(s) for blight, fear and/or hypnotic pattern as needed. Blight is for a foe who gets too close, fear for a group wielding magic weapons totaling at least 3 pluses, hypnotic pattern for at least three opponents whose action economy is too good. In the first two cases, either the target(s) are close enough or the gloom weaver gets close enough for Burden of Time to affect them. Unfortunately, this doesn’t work for hypnotic pattern, because of the shape and size of the area of effect—well, the gloom weaver may get lucky and have all its targets, or all its most important targets, right at the front of the cube, but probably not, and it certainly can’t count on it. So blight and fear are preferred, and if the gloom weaver is considering hypnotic pattern, it should probably try confusion first. But then again, it can also take advantage of the fact that its mere presence on the battlefield has the potential to throw its opponents’ positioning all out of whack.
See, with all this discussion of the gloom weaver’s spell repertoire, I haven’t even had a chance to talk about its Shadow Spear attack or its Multiattack. That Multiattack comprises two Shadow Spear attacks and a spell. The gloom weaver is always going to wield that spear two-handed, because it has no shield or off-hand weapon to worry about. Also, the additional necrotic damage on a Shadow Spear hit is insane. When laid over the spear’s base piercing damage, it totals an average 34 damage on a hit—every bit as good as blight. And the gloom weaver gets to attack twice with it. And it can even cast blight in the same turn if it wants to. Or eldritch blast, or confusion, or fear . . . .
This is no ordinary spellslinger. This is a shock attacker who lays into one opponent while also lobbing spells at other opponents elsewhere on the battlefield. Out of both sound strategy and pure spite, it’s going to make its melee attacks against squishy backliners and its ranged attacks against the hardy front-liners who just happen to be in the wrong place altogether, because it can.
And whenever it takes a hit, it uses Misty Escape to reposition itself alongside whichever of its opponents is the most isolated and vulnerable. [See below.]
Both shadow dancers and gloom weavers have normal self-preservation instincts and will try to retreat when seriously wounded (reduced to 28 hp or fewer in the case of the shadow dancer, 41 hp or fewer in the case of the gloom weaver). Shadow dancers use Shadow Jump, if possible, to avoid opportunity attacks, then Dash away; if they can’t Shadow Jump, they Dash away from a single melee opponent and Disengage from more than one. Gloom weavers use the head start granted by Misty Escape to get a safe distance away, then Dash.
ETA: As commenter idanbhk points out, I misread Misty Escape—the gloom weaver can use it only once per combat encounter. As the name suggests, this is an escape feature: if it can be used only once, the gloom weaver will use it when it needs to make a getaway. This means it’s not usable for bamfing from opponent to opponent. By extension, rather than attack one target with its Shadow Spear and a target somewhere else with spells, the gloom weaver may need to make more of an effort to finish one foe off before moving on to another. In fact, since Burden of Time imposes saving throw disadvantage on creatures within 10 feet, it’s implied that the gloom weaver ought to go ahead and hit the melee opponent with a spell as well. The incentive certainly exists. But make sure it’s a spell that calls for a saving throw, not a spell attack. Not only does eldritch blast gain no advantage from Burden of Time, casting it at point-blank range will incur disadvantage on the attack roll.
Soul mongers are also spellcasters with a nasty melee attack, but their spells are wizardy, completing the caster trifecta. They also have a trio of painful features—Soul Thirst, Weight of Ages and Wave of Weariness—that synergize to make them more dangerous as combat grinds on.
Wave of Weariness is a recharge ability that cools down on a roll of 4–6, meaning the soul monger gets to use it every other turn, on average. And it’s mean: In a cubic area of effect large enough to encompass an entire enemy side, targets have to make a DC 16 Con save or suffer a massive amount of psychic damage plus a level of exhaustion. One level of exhaustion is no big deal, two are bearable, but three or more are deadly.
This feature comes with a cost, though: Every creature in the area of effect gets hit, including other shadar-kai. (In contrast, Weight of Ages explicitly names other shadar-kai as being immune to its effect.) This makes it hard for a soul monger with allies around to get as much out of Wave of Weariness as it could on its own. On the other hand, if there are invisible gloom weavers lurking just outside the range of the weariness zone, Burden of Time imposes disadvantage on the saving throws of those within 10 feet of its perimeter.
The upshot is that a soul monger that’s not working alone needs to try to bang out as many uses of Wave of Weariness as it can manage before its allies jump in and join the fight. Once they do, even when it recharges, it can no longer take this action without risking harm to its own side.
Weight of Ages slows the movement speed of any creature that starts its turn adjacent to the soul monger by 20 feet. This effect makes it harder, if not impossible, for an opponent to run away from the soul monger, which in turn gives it an incentive to remain engaged with them and chase them down. Soul Thirst gives it a huge number of temporary hit points and advantage on attack rolls, which gives it an incentive to finish its opponents off. Other shadar-kai know this and will let a soul monger swoop in and take over against seriously wounded foes.
When it’s not taking the Wave of Weariness action, the soul monger’s primary choice is between Multiattacking and casting a spell. A hit with its Phantasmal Dagger, in addition to dealing hefty damage, imposes disadvantage on saving throws until the start of the soul monger’s next turn. The soul monger itself can’t benefit from this disadvantage—before it can cast a spell or take any other action that would require a saving throw, the effect wears off—so it chooses targets that it can soften up for its allies’ abilities, such as the shadow dancer’s Spiked Chain. If it can’t reach any of them, gloom weavers know to take opportunistic shots at opponents whom the soul monger has struck with its dagger.
With the exception of the cantrips chill touch (long-range, up to 120 feet) and poison spray (short-range, within 10 feet), all of the soul monger’s spells are one-shots. Bestow curse, gaseous form and phantasmal killer require concentration and are therefore mutually exclusive; chain lightning, finger of death and seeming can be cast anytime.
- Bestow curse is cast at its base level, which isn’t that hot—it requires concentration, lasts for a minute tops, and doesn’t accomplish anything significantly better than the soul monger’s other spells and traits. Why cast a spell to impose disadvantage on saves made with one ability score when you can just stab someone and impose disadvantage on all of them? What’s the great benefit of choosing only one opponent to have disadvantage on attacks against you? Are an extra 4 points of damage per attack, on average, worth your concentration when your simplest melee attack deals 32? The one option that might be worth the opportunity cost is forcing an opponent to make a Wisdom save every turn or lose its action, in order to shut down an opponent with lots of attacks and/or a devastating bonus action. The ideal target for this kind of curse is a monk, but only if the soul monger can team up with a gloom weaver to hit the monk with both Burden of Time and Weight of Ages at once, because monks are too good at Wisdom saves if you can’t impose disadvantage on them.
- Chain lightning is extra-magnificent when it can be made to target enemies restrained by shadow dancers’ Spiked Chains, afflicted by Burden of Time or struck by the Phantasmal Dagger. As soon as the number of foes subjected to one of these misfortunes equals or exceeds four, the soul monger seizes the moment.
- Finger of death, for some, is a finishing move, to be used against the squishiest opponents when they’re hanging by a thread—or already unconscious—to maximize the chance of raising them as zombies. But soul mongers are neutral, not evil; for them, the best application of this spell is to start the Soul Thirst snowball. Casting it at a target with 55 hp or fewer offers a two-thirds chance that the damage will reduce them to 0 hp, thereby activating Soul Thirst.
- Gaseous form is another spell that seems best suited to sneaky infiltration, but it’s only useful for this purpose is the soul monger is by itself.
- Phantasmal killer, even post-errata, sadly fails to live up to its name—but a team of shadar-kai may be able to salvage it, owing to their ability to impose disadvantage on saving throws in a variety of ways. The first failed Wisdom save, at the time of casting, merely imposes the frightened condition, but there’s nothing wrong with giving an opponent who makes multiple weapon attacks per round disadvantage on them. The target begins taking damage when it fails a second save on its own turn, and keeps taking damage as long as it keeps failing saves. So a soul monger bearing down on a foe with Weight of Ages and stabbing them with its Phantasmal Dagger, or working in concert with a gloom weaver and its Burden of Time, can potentially turn this into a reasonably assured average 22 damage on each of the target’s But all the pieces have to be in place.
- Seeming serves only to disguise the soul monger and its allies until they’re ready to start a fight—or to allow them to vanish into a crowd as they retreat.
Of these spells, chain lightning and finger of death are unambiguously valuable, and the soul monger will probably use them at some point in every fight; gaseous form and seeming are highly situational and particular in their applications; phantasmal killer is strong enough to consider if the soul monger and its allies can ensure that the target keeps failing its saving throws, but otherwise probably not worth bothering with; and bestow curse is worth considering only in the one very narrow circumstance described. As for poison spray, frankly, it’s a stupid choice when the soul monger can simply walk 5 feet closer and attack twice with its Phantasmal Dagger, and chill touch makes sense only when the soul monger has nothing better to do, which is almost never.
Despite its high Wisdom, the soul monger’s attitude toward self-preservation is weird. Although it’s desperate to maintain its existence, it illogically won’t retreat to save itself; instead, it insists on plundering the life force of others with Soul Thirst, and would rather let itself be destroyed than forgo a chance to do so. If its allies are retreating, it will grudgingly join them—but it will be the last to leave, doing its level best to consume its foes’ élan vital till the bitter end.
Next: conjured critters and how to manage them.