Why Hypnotic Pattern Stays S-Tier All 3 Acts (And Which Spells Fall Off After Act 1) — BG3 Spells Guide 2026

Verified on Patch 8. Values may change with future updates.

Most players pick spells based on how impressive they sound at level 1 and never re-evaluate. The result: by Act 3 you’re burning 3rd-level slots on Chromatic Orb against enemies with 140+ hit points, averaging 22 damage, while the player next to you just froze six of them with a single Hypnotic Pattern and let the party clean house with basic attacks.

This guide ranks BG3 spells by how useful they remain across all three acts — not just at level 1 when everything feels powerful. Some spells are S-tier from the Nautiloid to the final battle. Others earn their spot in Act 1 and quietly become dead weight before you leave the Underdark. Knowing the difference saves spell slots, saves concentration, and wins fights.

For a full list of every available spell by school and level, see our BG3 spell list. For picking the right class to cast them, check the best class guide.

Quick Start: 5 Spells to Lock In Before You Leave Act 1

If you’re short on time, these five spells pay dividends from the goblin camp through the endgame. Prioritise these before experimenting with anything else:

  1. Bless (Level 1, Cleric/Paladin) — +1d4 to every attack roll and saving throw for three allies. The best spell per slot in the game.
  2. Healing Word (Level 1, Cleric/Bard/Druid) — Revive a downed ally at range as a bonus action. Keeps your healer mobile and out of melee reach.
  3. Hypnotic Pattern (Level 3, Bard/Sorcerer/Warlock/Wizard) — Incapacitates entire groups with a Wisdom save. Unchanged in power from Act 1 boss fights to the endgame.
  4. Haste (Level 3, Sorcerer/Wizard) — Extra action, +2 AC, Advantage on Dex saves, +9m movement for 10 turns. Wins encounters by itself on the right character.
  5. Hold Person (Level 2, most spellcasters) — Paralyses a humanoid enemy, guaranteeing critical hits from melee allies. Upgrade to Hold Monster in Act 3.

How to Evaluate Any Spell: The 3-Question Slot Test

Before learning a new spell, run it through three questions:

  1. Does it scale? Chromatic Orb adds +1d8 per upcast level. Witch Bolt adds nothing. Enemy HP in Act 3 is 3-8× higher than Act 1 — flat-damage spells become irrelevant.
  2. Does it compete for concentration? Concentration spells block each other. Hex blocking Hypnotic Pattern is a losing trade by Act 2.
  3. Does it affect groups or just one target? Single-target spells need to hit something worth the slot. Crowd-control spells that freeze three enemies are worth two slot levels above their listed cost.

S-Tier: Every Slot Is Worth It in All 3 Acts

Hypnotic Pattern

The best crowd-control spell in the game — not just Act 1, every act. Cast at 18m range into a 9m radius and everything that fails a Wisdom save becomes Hypnotised: no movement, no actions, no bonus actions, no reactions. The condition breaks if the target takes damage or gets shoved, so your job after casting is to clean up quietly [1].

What keeps it S-tier in Act 3 when enemy HP has tripled? It doesn’t care about hit points. A 200 HP boss’s minions are still Wisdom-saving against DC 15-17 from your build, and most of them fail. The immunity list is real — Undead (except Zombies), Constructs, and Doppelgangers are immune [1] — but the spell still trivialises any encounter with mixed enemy types.

One limitation to manage: allies within range are also affected unless you’re using the Sorcerer’s Careful Spell metamagic. Position your party before casting, or accept a -10 approval penalty if a companion gets caught.

Haste

Haste does something no other spell does: it gives a character a full extra action per turn for 10 turns [2]. Not a bonus action — a real action. On a Fighter with Action Surge, that’s four attacks in a single turn. On a Sorcerer, it’s two spells. On a Paladin, it enables two Smites without the resource trade-off.

The mechanical benefits stack: +2 AC, Advantage on Dexterity saving throws, and +9m of movement that compounds with Dash actions. Sorcerers with Twinned Spell metamagic can Haste two characters simultaneously from one 3rd-level slot — one of the strongest single actions in the game [7].

The one rule: never cast Haste on yourself and then immediately switch to another concentration spell. Breaking concentration before the Haste expires gives the target the Lethargic condition, ending their turn immediately [2]. Cast Haste first on a different character, then maintain that concentration slot for the fight.

Bless

Every attack roll and saving throw gains +1d4 (average +2.5) for three targets, lasting 10 turns. Against enemies with 18 AC, that extra 2.5 average converts missed hits to landed attacks several times per round across your party. Against a Wisdom save DC 17 on your own saves, that same +2.5 flips roughly 25% of failures into successes [6].

Bless is a 1st-level concentration spell. In Act 3 you have 9th-level slots available. Running Bless from a 1st-level slot while your other casters handle Haste and Hypnotic Pattern is pure slot efficiency — and it’s still doing meaningful work in the endgame that a 5th-level spell from Act 1 can’t match.

Spirit Guardians

Cleric-exclusive, but if your party has one, Spirit Guardians is one of the most reliable damage sources in the game. Every enemy that enters or starts its turn in a 3m radius around the Cleric takes 3d8 radiant or necrotic damage and has its movement speed halved. No turn limit — as long as concentration holds, the aura persists [7].

The reason it scales into Act 3: the 3d8 (average 13.5) ticks every turn for every enemy in range, and enemies in Act 3 group up. Two enemies in range for five turns is 135 average damage from a single 3rd-level slot without using any other actions. Upcast to 4th-level for +1d8, and it keeps pace with endgame HP totals.

A-Tier: Reliable Through All 3 Acts

Counterspell

The situational pick that becomes less situational as the game progresses. Act 1 has few dangerous casters. Acts 2 and 3 are full of them — Mind Flayers throwing Dominate Person, Githyanki Wizards casting Fireball into your formation, final-act enemies with high-damage spells. Counterspell as a reaction stops any spell of 3rd level or lower automatically, and contests higher-level spells with an Arcana check [6].

Keep one caster with this available in any Act 2-3 fight where enemies have spellcasting tooltips. The action it denies is worth more than whatever else you would have cast that reaction on.

Hold Person / Hold Monster

Hold Person paralysis is good in Act 1 and Act 2 against humanoids. What makes it A-tier across the game is the critical hit guarantee: paralysed enemies have every attack roll against them made with Advantage, and attacks within 3m score automatic critical hits. A Fighter with three attacks inside 3m of a paralysed enemy deals roughly double expected damage on that turn [4].

Upgrade to Hold Monster in Act 3 — same mechanic, works on non-humanoids. The upgrade is mandatory for the final act where many key enemies fall outside the “humanoid” type restriction.

Fireball

Classic area damage. 8d6 fire damage in a 6m radius, no concentration, no save-for-half against fire immunity in BG3 (save halves). Fire isn’t the most resisted damage type in the game — that distinction goes to poison — so Fireball remains effective in most Act 3 encounters. Its main advantage is no concentration cost: cast Haste, maintain it, and throw Fireballs with the rest of your actions.

Call Lightning

Druid-specific but worth flagging: Call Lightning deals 3d10 lightning damage per action without consuming additional spell slots after the initial cast. Against wet enemies, it triggers a doubling effect. Once cast from a 3rd-level slot, it keeps paying dividends every turn without burning resources [7]. In long encounters, the damage-per-slot efficiency surpasses most offensive spells.

The Falloff Problem: Why Act 1 Stars Become Act 3 Slot Waste

Enemy hit points in BG3 scale sharply. A goblin in Act 1 has 15-30 HP. A Steel Watcher in Act 3 has around 150. A flat-damage spell that wasn’t impressive against 30 HP is genuinely useless against 150.

The math is simple: if a spell’s average damage is less than 15% of the target’s HP and it requires your action and a spell slot, it’s almost always better to cast something else. Use this as the minimum threshold when evaluating whether to keep a spell prepared in Acts 2 and 3.

Spells to Drop Before Act 2

SpellAct 1 ValueWhy It Falls OffReplace With
Chromatic OrbSolid — flexible damage type, 3d8 Thunder baselineEven upcast to 3rd level (avg 22.5 dmg), barely dents 140+ HP enemies. Surface creation can hurt allies in tight corridors [3]Shatter or Misty Step
Witch BoltDecent — sustain damage without extra slotsDamage doesn’t scale with character level. 1d12 avg (6.5) on a 1st-level slot is worse than a cantrip by Act 2 [5]Any concentration damage spell
HexGood — extra 1d6 per hit for WarlocksCompetes for concentration that should be Hypnotic Pattern or Hunger of Hadar by Act 2. The 1d6 bonus doesn’t scale [5]Hunger of Hadar or Hypnotic Pattern
Color SprayStrong — blinds enemies below HP thresholdHP threshold triggers only on low-HP enemies. By Act 2, most enemies exceed it before the effect appliesHold Person

Spells to Avoid in Every Act

These spells underperform their slot cost from Act 1 through Act 3. Swap them out as soon as alternatives are available:

  • Barkskin — Sets AC to 16 as a concentration spell. Anyone already wearing medium armour or above gets zero benefit [5]. The concentration cost alone disqualifies it next to Bless or Haste.
  • Sacred Flame (Cantrip) — Targets Dexterity saving throws. Most enemies have good Dex saves, and on a success they take no damage, not half [5]. Fire Bolt (attack roll, 1d10) is reliably better.
  • Prayer of Healing — Cannot be cast during combat [5]. Short rests are available after every two encounters and cost nothing. This spell is never worth preparing over a combat-relevant option.
  • Blade Ward (Cantrip) — Halves physical damage for two turns. Using your action to avoid damage instead of dealing it is a losing trade in almost every encounter [5].
  • Charm Person — Prevents the target from attacking you and grants Advantage on Charisma checks against them. Effect ends instantly if anyone deals damage to them [5]. In a combat encounter with four party members, this lasts approximately one action.

Spell Priority by Player Type

Player TypeCore SpellsSkip These
New PlayerBless, Healing Word, Hypnotic Pattern — these three win most fights without complex setupsWitch Bolt, Color Spray, Barkskin
CasualHypnotic Pattern, Haste, Fireball — freeze enemies, buff one attacker, area clear the restAny spell requiring Concentration + combo setup in same turn
OptimizerTwinned Haste (two characters hastened), Spirit Guardians (upcast), Counterspell reactive — squeeze every action economy advantageSacred Flame, Charm Person, Phantasmal Force
CompletionistExperiment with every school — some niche picks (Speak With Animals, Speak With Dead) unlock unique story contentNothing permanently — test everything once

Spell Slot Decision Tree

Use this in the first round of any encounter:

  • Multiple enemies grouped? → Hypnotic Pattern (if not Undead/Construct-heavy) or Fireball
  • Single high-priority target? → Hold Person/Monster + melee pile-on for crit damage
  • Party needs sustained damage? → Spirit Guardians (Cleric) or Haste (best martial character)
  • Enemy casters in the group? → Save a reaction for Counterspell; open with Hypnotic Pattern on their melee
  • First round, limited info? → Always Bless. The +1d4 bonus starts paying off immediately and doesn’t need a target assessment.

For more on how saving throws interact with these spells, the BG3 saving throw chart breaks down which enemy types are most likely to fail which saves — useful for targeting Hypnotic Pattern at Wisdom-weak groups and Hold Person at Dexterity-weak targets.

FAQ

Does Hypnotic Pattern work on bosses?

It works on most humanoid and beast-type bosses. Immune enemies — Undead (except Zombies), Constructs, and Doppelgangers — will be listed in the tooltip when you hover over a target [1]. Always check before spending your concentration slot. The upside: most of the dangerous Act 3 boss encounters mix immune and non-immune enemies, so you can still freeze the surrounding group even when the main boss resists.

What’s the single best spell at Level 1?

Bless. No contest. +1d4 to every attack roll and saving throw for three party members turns marginal hit chances into reliable ones and turns risky saving throws into manageable odds. It scales in value as enemy AC and save DCs increase in Acts 2 and 3.

Should I swap Chromatic Orb out before Act 2?

Yes. By Act 2, 3d8 Thunder damage (average 13.5 from a 1st-level slot) no longer dents anything relevant. Replace it with Misty Step for repositioning or Shatter for a no-concentration AoE that stays useful through Act 2. If you’re a Sorcerer, slot in Hypnotic Pattern the moment you hit level 5.

What’s the best use of a 3rd-level spell slot?

For arcane casters: Hypnotic Pattern for crowd control or Haste for action economy. For Clerics: Spirit Guardians — the sustained AoE damage out-values almost every other 3rd-level option across a full encounter. For Sorcerers with Twinned Spell: Haste on two characters from one slot is the best action in the game at that slot level.

Find the right build to pair with these spells in our BG3 beginner’s guide, which covers party composition and class synergies across all three acts.

Sources

  1. bg3.wiki — Hypnotic Pattern
  2. bg3.wiki — Haste
  3. bg3.wiki — Chromatic Orb
  4. Game Rant — Baldur’s Gate 3: 23 Best Spells
  5. Game Rant — The Worst Baldur’s Gate 3 Spells
  6. sevenswords.uk — Baldur’s Gate 3: All Spells Ranked by Usefulness
  7. Pro Game Guides — Best Concentration Spells in BG3
Michael R.
Michael R.

I've been playing video games for over 20 years, spanning everything from early PC titles to modern open-world games. I started Switchblade Gaming to publish the kind of accurate, well-researched guides I always wanted to find — built on primary sources, tested in-game, and kept up to date after patches. I currently focus on Minecraft and Pokémon GO.