Resources

Empirical commander popularity and co-occurrence

EDHREC is a public index of Commander decks that have been published to the major list sites. For each legendary creature it records two quantities: the number of decks in which that creature is the commander, and the rate at which every other card appears in that sample. Aetherwell’s commander table is ordered by those two quantities. The High synergy, Most played, land, and mana-rock suggestions presented while a list is being built are the same per-commander series, restricted to the color identity of the selected commander and to cards still legal in the chosen format.

The percentages therefore describe a public sample, not a hidden population. If one hundred published decks with a given commander include Sol Ring, EDHREC reports that rate. Unpublished binders, unlisted kitchen-table piles, and decks that were never uploaded do not enter the denominator. A high inclusion rate means that the card is common in the observed sample. It does not entail that the card is necessary, nor that it is strong in every construction. When EDHREC revises a commander’s page, Aetherwell follows the revision.

Oracle text, legality, and card images

Scryfall is the reference database for Magic: The Gathering cards. The names, type lines, mana costs, color identities, oracle texts, and format legalities displayed on Aetherwell are taken from that database. The full-card images, including those shown on hover, are served from Scryfall’s image hosts. Aetherwell does not retain a local archive of the artwork.

A search conducted inside a format is submitted to Scryfall with that format’s legality restriction, so that a card not legal in the format does not appear among the results. When the catalog is refreshed, the same interface is queried for current oracle text and current legalities, in order that a reprint, a rules change, or a ban not remain stale on this site. Where Scryfall and Aetherwell diverge, Scryfall is authoritative.

Documented combinatorial lines

Commander Spellbook is a catalog of documented infinite and near-infinite combinations. Each entry records the cards required, the result the combination produces, and a brief description of the line. Every combination shown on Aetherwell’s Combos page, and every combination attached to a particular list, is one of those published variants. Aetherwell does not originate loops.

A two-card combination is counted toward Commander Brackets only when Spellbook records it as a two-card recipe and both cards are present in the list. Combinations of three or more cards are offered as suggestions; they do not, of themselves, alter the official class. If a recipe as rendered here is in error, the error is one of transcription. Spellbook remains the authority.

The official taxonomy of Commander power

For most of the format’s history, players described their decks with an informal integer from one to ten. That integer had no shared definition. One table’s seven was another table’s four. In 2025 the Commander Format Panel at Wizards of the Coast published five named brackets. The purpose of the publication was a common vocabulary. When one player says Core and another says Optimized, both are referring to the same published constraints, not to a private impression of strength.

The classes are Exhibition, Core, Upgraded, Optimized, and cEDH. Assignment is not a matter of taste. It follows those constraints. Brackets 1 and 2 permit no cards from the Game Changers list. Bracket 3 permits at most three cards from the Game Changers list. Bracket 4 and Bracket 5 permit any number from the Game Changers list. Mass land denial, extra-turn effects, and two-card combinations are likewise constrained: they are absent from Exhibition, sparse in Core, late in Upgraded, and unrestricted in Optimized. Aetherwell applies the Game Changers published on 21 October 2025, together with subsequent additions such as Biorhythm.

Bracket 5 is the designation Wizards assigned to competitive EDH. Optimized and cEDH are permitted the same cards. Both may include any number of Game Changers. Both may include mass land denial, extra-turn effects, and two-card combinations capable of ending the game in its opening turns. The distinction is not a further restricted list. The distinction is the table. A Bracket 4 game remains a game of Commander among players who have brought fast, consistent decks. A Bracket 5 game is a known competitive metagame: the participants have studied the same commanders and the same lines, and they have sat down in order to win that metagame. Intention of that kind cannot be recovered from a list of names.

That is why Aetherwell will not assign cEDH on the basis of the list alone. A construction that contains Thassa’s Oracle, Demonic Tutor, and a two-card combination may be an Optimized deck brought to a living-room table, or it may be a tournament list. The cards do not disclose which. To print Bracket 5 from the list would be to assert what the evidence cannot support. The highest class that can be honestly reported from cards alone is therefore Optimized.

What Aetherwell will do is record when a list would not be out of place at a competitive table. If the construction is already Optimized, and it further contains five or more Game Changers, an early two-card combination, and dense fast mana, the site appends a note: the list could sit at Bracket 5 if that is the table intended. The note is an invitation to say so. It is not a fifth classification.

Beside the bracket, Aetherwell prints a number from one to ten. That figure is not Wizards’ instrument, and it does not amend Wizards’ classes. It is an internal estimate of how rapidly the list can close a game. The named bracket remains the classification. The number is a commentary upon it.

The estimate begins at 3.4, the neighborhood of a Core list that carries none of the published flags. Each subsequent term measures a different way in which a game may be shortened. The terms are not prices attached to individual cards, as if Demonic Tutor were worth more than Cyclonic Rift. They are coefficients on orthogonal axes, and each axis is given an upper bound. The bound exists so that repetition along a single axis cannot exhaust the scale. Six instances of the same kind of advantage are not six times the first.

The Game Changers list is Wizards’ principal axis, and it is therefore counted card by card. Each such card contributes 0.38. One Game Changer is a modest increment. Three — the official maximum in Bracket 3 — contribute 1.14. The saturating bound on this axis is 2.4, reached at slightly more than six Game Changers. Further cards from that list do not increase the estimate. The 2.4 is not a heavier card, and it is not a second tariff levied on a special subset. It is the maximum this one dimension is permitted to add, so that a list composed of a dozen Game Changers and no winning line is not scored as if the game were already over.

Extra-turn spells are treated analogously: 0.35 apiece, saturating at 1.2. A single extra turn is a published flag. Three extra-turn effects are the threshold at which Aetherwell already treats the list as chaining turns for the official bracket, and they are also the point at which this axis is closed. Mass land denial is not analogously incremental. Wizards constrains the presence of the effect, not its multiplicity. The estimate therefore adds 1.3 if any such card is present, and does not add it again. One Armageddon already reconstitutes the game. A second copy does not reconstitute it a second time.

Two-card combinations receive the largest single coefficient because they are the least ambiguous statement that the game can end on a line rather than on attrition. When both pieces are in the list and their mana values sum to seven or less, the line is treated as early and receives 1.6 — more than four Game Changers. That ordering is deliberate. An early two-card win compresses the expected length of the game more than a cluster of powerful singles. A two-card combination whose pieces sum to more than seven mana is a late line and receives 0.55, comparable to one additional Game Changer, because the game has already been played by the time it resolves. Fast mana — Sol Ring, Mana Vault, the Moxen, Grim Monolith, Jeweled Lotus, and Mana Crypt — contributes 0.28 per piece and saturates at 1.4. One Sol Ring is commonplace. Five pieces constitute a dense acceleration engine, and that is where the axis closes, so that a collection of Moxen without a terminal line cannot be scored as if it were a combination.

Finally, the average mana value of nonland cards is compared with 3.3. A cheaper curve raises the estimate; a dearer curve lowers it. The residual is confined to a narrow interval, from 0.8 below the running total to 1.1 above it, because an inexpensive tribal list is not, by that fact, Optimized. The terms are then summed, rounded to one decimal place, and confined to the closed interval from 1.0 to 10.0.

A community one-to-ten demand scale

Before the official brackets were published, several community instruments attempted to answer, with a single number, how severe a deck was. EDH Power Level is the most widely cited of those instruments. It inquires after the efficiency of the mana curve, the density of high-impact cards, and the speed with which the deck can end a game.

The one-to-ten figure printed beside an Aetherwell bracket stands in that tradition. A cheaper nonland curve, a greater number of Game Changers, denser fast mana, and the presence of two-card combinations all raise the estimate. The coefficients are Aetherwell’s. They are not a reproduction of EDH Power Level’s unpublished model. The five named brackets remain the description that Wizards asked players to employ.

Tournament co-occurrence

MTGTOP8 publishes finishing lists from sanctioned and reported tournaments: the cards that shared a deck which placed in the Top 8 of an event. Aetherwell’s companion suggestions put a single question to that archive. Given the commander and the cards already chosen, which other cards appear beside them in those results, and with what frequency?

A Commander list is sought among cEDH and Commander events. Standard, Modern, Pioneer, Legacy, Vintage, Pauper, and the remaining constructed formats are sought under the format names that MTGTOP8 already publishes. Aetherwell neither conducts those tournaments nor hosts the lists. The record remains MTGTOP8’s.

Magic: The Gathering

Magic: The Gathering, including every card name, rules text, and format, is the property of Wizards of the Coast, a subsidiary of Hasbro. Aetherwell is an independent fan work. It is neither affiliated with, nor endorsed by, nor a substitute for Wizards of the Coast or Hasbro. Nothing on this site constitutes an official ruling.