Know Your Spells: A Tier List, Synergies, and Matchups for Sigil

Draft for the publisher's blog. Ratings are from our own play experience; the statistics come from 2,411 completed games on Sigil Online.

Every game of Sigil deals nine spells onto the board: three rituals (five nodes), three sorceries (three nodes), and three charms (one node). Both players can see them from the first move, and both can fill and cast any of them. The spells themselves are the shared terrain of the game, so knowing which ones are worth fighting for, which ones work together, and which ones beat which is most of what separates a strong opener from a weak one.

This article covers the 39 spells in the core set and the eight expansion packs (Springtime, Celestial, Inferno, Tempest, Flood, Autumn, Gloom, and Covenant). It has four parts: a tier list, a guide to which spells prefer open or crowded boards, a table of synergies and anti-synergies, and a table of head-to-head matchups. A final section looks at what the recorded games say, and where they disagree with us.

A reminder on two rules that come up repeatedly below. Casting a three- or five-node spell locks it for you until your lock moves (one-node charms never lock), and you get one cast per turn unless you hold Seal of Summer. Red moves first, and to compensate, blue starts one point ahead on the score track (a phantom stone that is not on the board). A player wins when the track shows them three ahead, so blue wins with two more stones on the board than red, while red needs four more than blue.

The tier list

We ranked all 39 spells by direct comparison, answering "which of these two is better?" for 241 pairs and fitting a Bradley-Terry model to the answers. The order below is the result. The tiers break that order at its largest gaps in strength, keeping every tier to at least three spells.

Full order, best to worst:

A few notes on the placements that tend to surprise people.

Gather and Harvest are at the top because they work with everything. On their own they do nothing except refill themselves. But cast any other non-static spell first and then Gather or Harvest, and the second cast refills the first spell, unlocks it, and nets you a stone in the process. Paired with a soft-move spell they become an engine: the two spells refill each other, you cast one every turn, and once both are filled you gain two or three stones over four to six turns. The fastest win we know of uses exactly this (see the end of the article).

Fireblast rewards patience. It destroys every enemy stone touching yours, then costs you one stone. On a board where the two sides are in contact in many places it can swing the game by four or five stones in one cast. Its weakness is that building that contact takes time, which is why Carnage beats it head to head (see Matchups).

Scatter and Blossom rank third and fourth for how they combine. Both drop stones into several spells at once, which both develops the board and sets up Erupt (two moves into every spell you already have a stone in). Their weakness is a specific counter: Hail Storm destroys one enemy stone in every three- and five-node spell, and Decay destroys every enemy stone touching two or more empty nodes. Against either, a freshly cast Blossom can be wiped in one turn.

Charge and Slash are the best charms. Their value shows at the front line. When you and your opponent are fighting over a spell and you hold one of these charms behind your side of it, you can spend the charm's stone to push forward. At minimum your opponent can never push you back past the charm, so they lose ground over time; at best the push crushes an enemy stone outright. Slash is the strongest charm in the core set; Charge (a move into any three- or five-node spell) is the strongest overall.

Grow is near the bottom, despite being a fine partner. Two soft moves for a three-node investment does not net stones. Grow is good in one situation only: next to Harvest, where the pair refill each other. Torrent, one soft move then one hard move, is the weakest spell in the game.

The static seals are mostly weak. Seal of Lightning and Seal of Wind are the exceptions and both sit in the strong tier: Lightning makes a dash cost one sacrifice instead of two, and a blink for your first move every turn is real mobility. Seal of Summer looks powerful on paper, but two casts a turn only matters on a board crowded enough to have several spells filled at once, or when one spell fills another. With one or no spells filled it does nothing. Seal of Destruction is a special case: it destroys everything touching you at the end of your turn, but if it is still filled at the start of your next turn you lose, and only your opponent moves in between. It is a finisher, not a spell you develop toward, and it is much stronger for blue, who wins on a two-stone lead: from level material, bordering two enemy stones and filling the seal is enough, where red would need to destroy four. The way to land it is with a spell that fills it in one cast, such as Tsunami, Grow, Flourish, or Eclipse, at a moment when you already border many enemy stones elsewhere: the seal fills and fires in the same turn, before your opponent can react. The way to lose with it is to build it slowly while Gust is on the board. Gust picks up every enemy stone touching the caster and drops them anywhere, including the five nodes of your half-built seal, and a seal that is full at the start of your turn ends the game against you.

Open board or crowded board?

Some spells want empty space and some want contact. Where a spell falls on this axis tells you when in the game to fill it.

The clearest cases are the two destruction sorceries. Fireblast wants as many points of contact as possible, so it is at its best late, when the board is full. Decay destroys enemy stones touching two or more empty nodes, so it wants the opposite: an early board with thin, spread-out formations.

Two entries deserve a comment. Bewitch converts two touching enemy stones, so you might expect it to want a crowded board. It does not need one: every turn starts with a regular move that places a stone next to one of your own, so you almost always end your turn with a touching pair. The only way to deny Bewitch its targets is to dash and sacrifice every turn, which costs more than being Bewitched. Bewitch's real value is on an open board, where converting two stones swings territory.

Hurricane destroys the smallest contiguous enemy group. Early in the game your opponent may have a single group, and a Hurricane then removes all of it. On a crowded board the groups are already split and Hurricane only picks off stragglers. We list it as no preference because its real dependency is not crowding but group structure, which the opponent controls (see Matchups).

Synergies

Two spells have synergy when holding both lets you do something neither does alone. Three general principles explain most of the table.

Refill loops. Casting a three- or five-node spell locks it. Gather and Harvest place stones into your locked spell, so cast a spell and follow with Gather or Harvest and you have both refilled the first spell and gained a stone. Soft-move spells are the best partners, because they can refill Gather or Harvest in turn. This is the strongest synergy in the game. Gather and Harvest have no synergy with each other, because refilling each other nets nothing.

Destroying and converting are redundant. Every enemy stone Hail Storm or Meteor removes is one Bewitch or Corrupt can no longer convert. Both effects are good, but they eat each other's targets. Two destruction spells cast on consecutive turns thin each other out the same way.

You only need to reach a mana once. Meteor, Blossom, Seal of Wind, Comet, and Flourish all help you cover distance. Holding two of them is wasteful: once you have reached a distant mana, the second reach spell has nothing to do, unless the opponent retakes the area.

There is also a small, universal anti-synergy: with one cast per turn, any two spells compete for the same casting opportunity. We ignore it in the table, except where it is the main effect.

Positive synergies

Some of the entries worth explaining:

• Blossom or Scatter, then Erupt. The first seeds a stone into several spells; Erupt then makes two moves into every spell you have a stone in.
• Seal of Lightning and Surge. Lightning makes dashing cost one sacrifice; Surge gives a free move whenever you dashed. Together, a cheap dash every turn pays for itself.
• Gust and Decay. Gust picks up every enemy stone touching you and drops them wherever you like. Drop them in isolation and Decay destroys them all next turn.
• Comet and Flourish. Comet's blink can grab a mana; Flourish's four soft moves then convert two or more mana into a stone advantage. Comet pairs well with many spells for two reasons: the blink can fetch a mana that makes the next cast stronger, or it can drop a stone into or beside an enemy group to raise the number of enemy stones you border before a destruction or conversion spell.
• Seal of Wind with Fireblast or Corrupt. This is the same idea at full strength. Wind's blink is your regular move, so it does not use up your cast: drop a stone anywhere, then destroy or convert everything around it in the same turn.
• Bewitch and Fireblast. Convert part of an enemy group with Bewitch, and the rest of the group is now touching you for the Fireblast that follows.
• Seal of Destruction and a fast filler. Tsunami, Grow, Flourish, or Eclipse can fill the seal in one cast while you already border many enemy stones elsewhere, turning a slow finisher into a sudden one. Gust does the same from the other direction: it drops up to five enemy stones into the seal for you.

Negative synergies

Carnage, Slash, and Surge all trap and destroy cornered enemy stones, and there are only so many enemy stones you can corner. Carnage and Hail Storm both want a crowded board, then thin it out for each other.

Matchups

The matchup question is different from the synergy question: if you are building toward one spell and your opponent toward another, who is favoured? Some of these are about spells directly countering each other, and some are about tempo.

The stories behind the strongest entries:

Hurricane loses to any charm that makes a move. Hurricane destroys the smallest contiguous enemy group, which is devastating against a single group. The defence is cheap: place stones across a charm's node, then cast the charm. The vacated node splits your stones into two groups, and Hurricane now removes only the smaller one. Any charm that lets you place a stone (Sprout, Slash, Surge, Splash, Charge, Azimuth, Lurk) does this. Hurricane is far stronger for red, whose tempo lets it threaten the one-shot before blue can arrange a split; blue usually has to dash and sacrifice to stay safe.

Hail Storm and Decay punish spread. Blossom places one stone in each other three- and five-node spell, which is exactly one target per spell for Hail Storm. Those same lonely stones each touch two or more empty nodes, which is Decay's condition. Scatter has the same problem against Decay.

Carnage beats Fireblast. Fireblast needs many bordering stones to be worth its cost, and building that contact takes turns. Carnage's four hard moves can seize the contested area, or the Fireblast spell itself, before the buildup pays off. Comet beats Carnage and Bewitch for a different reason. Comet costs a sacrifice, and you choose which stone to give up: pick the one your opponent was about to push or convert, and their spell has less to work with.

Seal of Lightning beats Seal of Stone. Seal of Stone forces your opponent's regular move to be soft, but dashes and spell casts still push and destroy. Lightning makes the dash cheap, so the restriction hardly bites. For the same reason Seal of Stone does not really counter Carnage, Fury, or Slash. Seal of Autumn is the real answer to Lightning: a Lightning holder dashes every turn because it is nearly free, and Autumn forbids paying for dashes with stones that sit in spells, which chokes the supply.

Gust hard-counters Seal of Destruction. If your opponent is building toward the seal, Gust can pick up your bordering stones and drop them into the empty nodes of their seal. It is now full at the start of their turn, and they lose on the spot. Fireblast is the other natural enemy of the seal, since the same contact that powers the seal powers a Fireblast first.

Storm Front and Meteor do not counter seals. In theory knocking one stone out of a static seal switches it off. In practice the stone is easy to replace and the tempo is rarely worth it, so we rate those matchups even. Fissure, in the Tectonic pack (not covered here), is the real seal counter because its void disables the spell permanently.

What the recorded games say

We also looked at 2,411 completed games on Sigil Online. A word of caution first: only 21 of them are between two human players. Nearly all involve one of the site's AI opponents on at least one side, so the statistics reflect AI habits as much as human ones.

The obvious statistic is "how often does the player who cast spell X win?" It is also the most misleading one, because you only cast a spell when it is useful. A player who invests four moves into Hurricane and is then neutralised by a split group never casts it, and probably loses, and that game never shows up in Hurricane's numbers. So we looked at three views:

• (a) win rate of the side that cast the spell;
• (b) win rate of the side that started closer to the spell, once both players had a stone down;
• (c) win rate of the side with more stones in the spell after blue's fifth turn.

Views (b) and (c) measure commitment rather than payoff.

Four things stand out.

Selection bias is real, and largest for Hurricane and Surge. Hurricane wins 94% of games when cast and 51% for the side that invests in it. Surge wins 96% when cast, because casting it means you already dashed profitably, but the side that commits to it early wins only 39%.

Fireblast and Seal of Wind reward early commitment. They are the two spells where simply having more stones in them at turn 10 predicts winning, at 70% and 72% over hundreds of games each. If you are going to claim a spell early, claim one of these.

Leading in a ritual early predicts losing. The side with more stones in Flourish, Starfall, Bewitch, or Seal of Lightning at turn 10 wins only 41% to 45% of the time, across 650 or more games each. Five nodes is a lot of tempo to spend before the payoff.

The spell layout is fair. In standard games the nine spells are placed at random around the fixed starting positions. Over 900 games per spell, the side that starts closer to any given spell wins between 47% and 52%. No spell placement favours either player.

Where the data and our ratings disagree, we side with experience. The data rates Decay, Tsunami, and Gust higher than we do, on samples of 18, 43, and 4 casters respectively. It rates the charms as roughly interchangeable, because each side tends to hold whichever charm sits in its territory, so cast counts reflect geography rather than choice.

The fastest win

To close, here is how the refill loop plays out at full speed. Suppose the layout puts Harvest, Sprout, and Grow in a line from blue's starting mana, and red does not interfere.

1. Turns 1 to 3: spread through Harvest.
2. Turn 4: place in Sprout.
3. Turn 5: place in Grow, and cast Sprout to place Grow's second node.
4. Turn 6: fill Grow's third node and cast it, completing Harvest.
5. Turn 7: place in Grow and cast Harvest, netting one stone.
6. Turn 8: cast Grow.
7. Turn 9: cast Harvest again. Blue is two stones ahead and wins.

Nine turns, no opponent blunder required beyond leaving the engine alone. As far as we know this is the fastest win in the game that does not rely on the other player sacrificing stones or walking into a Hurricane or Seal of Destruction one-shot. It is also the clearest argument for why Gather and Harvest sit at the top of the tier list, and why, if you see this layout on your opponent's side of the board, you should go and break it up.

Method notes. Spell ratings and pairwise comparisons were entered by the author; the ranking is a Bradley-Terry fit over 241 pairwise comparisons, with all 39 spells connected and no preference cycles. Tiers are cut at the largest gaps between neighbouring strengths, skipping any cut that would leave a tier with fewer than three spells. Game statistics cover every completed game recorded on Sigil Online as of 5 September 2026, replayed through the site's own engine to recover board positions. Win rates are unadjusted; a version controlling for player strength and Elo gives the same qualitative picture.

New to the game? Start with the tutorial. Then try the puzzles, or read the Spells Matrix and the rulebook.

← Back to menu