Full rules

No. After many Bracket 3 tournaments, we can say with confidence that a well-built deck in this format is closer to a cEDH deck than to a Bracket 2 one. What changes is the pace: games run longer, key pieces are more tightly controlled and the paths to victory aren’t as explosive. Even so, the level of optimization, decision-making and deckbuilding remains highly competitive. Showing up to a Bracket 3 event with a casual deck can create serious mismatches and a bad game experience for both you and your opponents.

Bracket 3 was born as a power classification, but the community and stores need more than a label. The desire to play competitive decks that are more accessible than cEDH is real, and making that work in a tournament setting takes clear guidelines. These rules give the spirit of Bracket 3 a formal, fair shape, with verifiable rules that remove subjectivity and let organizers run events with confidence.

In cEDH, the Game Changers we limit to 3 here are precisely the most-played staples. A cEDH deck can run 20 or 30 of these cards without restriction. That is a significant barrier to entry, both in cost and in card availability, especially in regions like ours. Bracket 3 aims to keep the competitive spirit with a more democratic entry point: less reliance on expensive staples, more creativity in deckbuilding.

We did extensive research based on Wizards of the Coast’s official guidelines for the Brackets system, the philosophy behind each bracket and the experience gathered at real tournaments. We took that spirit and turned it into clear, measurable rules: lists of banned combos, mass land destruction cards, a Game Changers limit, a ban on cards that grant extra turns, and a web checker that automates validation to make deckbuilding easier.

It’s designed for every store and community that adopts the Bracket 3 format and wants to run tournament-style events with consistent rules. It’s also useful for individual players who want to validate their deck before an event, or for organizers who need an objective reference when reviewing decklists.

Most existing tools assign a Bracket generically, without breaking down exactly what keeps your deck from qualifying. Our checker analyzes several dimensions at once: it detects the 1061 banned two-card combos, identifies mass land destruction (MLD) cards, checks the Game Changers limit, reviews color identity, verifies the singleton rule, detects extra-turn cards and checks against the official Commander banlist. All in a single pass, taking the subjectivity out of the process.

There are definitely decks and strategies we have identified as disproportionately powerful even within Bracket 3’s restrictions. The more data we collect from real tournaments, the more precise our observations will be. Our goal is for this information to serve as input for suggesting, to those responsible for Wizards’ official guidelines, additions to the Game Changers list or other changes that strengthen the format.

We are not affiliated with Wizards of the Coast in this project, only through the WPN (Wizards Play Network) of our brick-and-mortar stores. The Bracket 3 Rules are an independent initiative created for Commander communities. That said, we are looking to open lines of communication with Wizards, because we believe tools that broaden the metagame and offer variety for every kind of player can contribute positively to the format’s ecosystem.

The previous version of the project tried to regulate cards by their mana value in combinations and "Cheat Into Play" effects (putting permanents onto the battlefield without paying their cost). Although the intent was good, in practice it proved extremely hard to maintain: every new Magic set adds dozens of cards that might qualify, judging what counts as a "cheat" is inherently subjective, and players found it hard to build decks without knowing exactly where the line was. It also invited inconsistent interpretations among judges and organizers about what was fair or unfair to the spirit of the format. That’s why we decided to focus on objective, verifiable restrictions: documented two-card combos, Game Changers on a closed list and specific MLD cards.

These are two questions, and they have different answers. First, why Ghost Quarter is never banned: what we regulate isn’t "destroying a land", it’s leaving your opponent without mana. Ghost Quarter and Field of Ruin destroy the land and its owner searches for a basic (they lose the utility land but keep the mana); Wasteland replaces nothing. It’s an objective line: either the card gives you a land back, or it doesn’t. And note that getting a creature doesn’t count as a replacement: a 3/3 doesn’t produce mana. Second, why the format split: at a four-player table, sacrificing a Wasteland to destroy a land is a 1-for-1 (both of you end up one land down) that only sets back one of your three opponents while the other two develop freely. The table punishes the technique on its own. Turning it into a prison takes several more pieces: recursion like Crucible of Worlds or Life from the Loam, plus extra land drops like Exploration or Azusa. That’s an assembly of three or more cards, the same category these rules explicitly allow, because Rule 1.1 bans combos of exactly two. In 1v1 the math changes completely: the same exchange is a 1-for-1 against your only opponent, with tempo in your favor, and repeated it’s a prison. That’s why Rule 1.4 only applies in 1v1. In both formats you can still answer a Gaea's Cradle or a Cabal Coffers with Ghost Quarter.

It’s true that the cost varies enormously: Strip Mine and Wasteland activate for free, Tectonic Edge asks for {1}, Dust Bowl {3}, Encroaching Wastes {4}, Memorial to War {4}{R} and Gnottvold Slumbermound {3}{R}{G}{G}. Even so, drawing the line by cost would be a mistake, for three reasons. The first is that the most dangerous card on the list isn’t Wasteland: it’s Dust Bowl. It’s the only one that doesn’t sacrifice itself (it sacrifices another land), so it stays on the battlefield and destroys a land every turn indefinitely, with no recursion needed. A line drawn by mana would leave legal the only self-contained engine in the group. The second is that a mana threshold isn’t stable: do you set it at zero, one, three? Any number is arbitrary, and every new set forces the debate all over again. The replacement criterion is derived from what the card does, not what it costs, so it updates itself. The third is that banning an expensive card costs nobody anything: nobody plays a six-mana Strip Mine. Over-including costs nothing; under-including leaves a gap.

Precisely because the criterion isn’t what gets played today. There are five cards (Terastodon, Numot the Devastator, Violent Ultimatum, Lavaball Trap and Rain of Salt), and they are every card in Commander that can destroy two or more lands in a single resolution. Almost nobody plays four of them, and that’s exactly the reason to include them: the list is derived from card text, not from the metagame, so it closes off the whole category instead of chasing individual cases. The cost of banning a card nobody plays is zero, because the checker reviews it for you; the cost of leaving a gap is that someone finds it at a tournament. And there are two more reasons. One: even if the mana cost seems prohibitive (Terastodon costs eight), there is always a way to put them onto the battlefield without paying it, with reanimation or similar effects, and in fact Terastodon is the only card in Commander that destroys two or more lands with an enters trigger, which makes it the only one of the five you can reanimate. Two: every new set can print another one like them, and the list rebuilds itself every Monday from card text, so it absorbs it without anyone having to remember. In multiplayer all five remain legal: spreading two or three lands across three opponents isn’t worth the card or the mana, and you fall behind the other two.

Because banning it would backfire. Stone Rain costs {2}{R} and destroys a land. Beast Within costs {2}{G} (exactly the same) and destroys any permanent, lands included, without replacing anything. Beast Within does strictly more for the same mana. If we banned Stone Rain and left Beast Within legal, we would be banning the weak version and legalizing the strong one: anyone who wanted to destroy lands would just play the other. And closing that gap would mean banning all of the format’s flexible removal (Beast Within, Generous Gift, Demolish, Creeping Mold, Aftershock, Vindicate), that is, a huge list that would punish decks with no intention of destroying lands at all. That’s why the rule only reaches lands: it’s the only case where destroying is free, because it costs you your land drop rather than a card or mana. A Stone Rain at a four-player table is a 1-for-1 that sets you back against the other three; a Wasteland costs you nothing, and any recursion gives it back.

Because recursion is a multiplier, not the cause, and on top of that it can’t be enumerated. On the first point: Crucible of Worlds with repeated Ghost Quarter leaves the opponent with the same number of lands every time; it does no harm. And a Wasteland with no recursion at all still takes away a land forever. If the base is zero, no multiplier raises it. On the second: we tried to ban the recursive half and it’s impossible. Life from the Loam returns three lands from the graveyard to your hand and recurs itself with dredge; Eternal Witness, Regrowth and Noxious Revival do the same on a one-off basis. That’s hundreds of cards, and all of them have legitimate uses that have nothing to do with lands. The destructive half, on the other hand, is a closed list of cards whose only job is destroying lands. The rule is always written about the half that can be enumerated.

Because "chaining" describes what happens at the table, and a decklist doesn’t contain what’s going to happen at the table. The previous version of the rule (2.1) tried it: it banned "strategies whose goal is to chain turns", and since that can’t be read from a list, the checker ended up applying, on its own, a threshold of three or four cards that wasn’t written anywhere in the rules. Two judges looking at the same deck could disagree, which is exactly what these rules exist to prevent. The category, on the other hand, can be enumerated: it’s 53 cards and the list is derived from card text, so Rule 1.5 bans all of them. There’s also a telling figure: of the 1,061 combos the community had already voted unsafe for Bracket 3, 36 use an extra-turn card, and all 36 have the same result: infinite turns and lockout. The EDHREC vote had already flagged this category 36 separate times; 1.5 simply stops chasing it case by case. And the cost of over-including is zero: an occasional extra turn can be swapped for another card without the deck losing its identity, while the gap is paid for by the tournament.

It’s the fairest question of all, and the answer is that the clause can be stripped away. Last Chance, Final Fortune, Warrior's Oath, Chance for Glory and Alchemist's Gambit (five of the 53) give you the turn and then you lose at the beginning of the end step. But "you lose the game" there is a trigger that happens in a specific step, and all it takes is never reaching that step. The case that shows it best is Obeka, Brute Chronologist: tap her and the turn ends immediately, skipping the end step where your loss was waiting. And Obeka isn’t a card you have to draw: she’s a legal commander, so she’s available from the command zone every game, every turn, without searching for her. Her identity is Grixis, which covers four of the five: Final Fortune, Last Chance, Warrior's Oath and Alchemist's Gambit. With her, those cards stop being suicide and become two-mana Time Walks, chainable as long as you keep recurring them. Sundial of the Infinite, Angel's Grace or Platinum Angel achieve the same by other routes. And here’s the point: none of those pairs appears among the 1,061 voted combos (Obeka isn’t in any of them), so until Rule 1.5, an Obeka deck with the four suicide cards was perfectly legal at a Bracket 3 tournament. To ban only the "true suicide" ones we would have to maintain a list of everything that cancels a loss or skips a step, and that list can’t be closed: every set can print another one. That’s why the rules do what they always do and legislate the half that can be enumerated, which is the cards that grant the turn. Obeka, by the way, is still legal: she grants no extra turn, and without those cards she has nothing to chain.

Remove the card and verify again. Rule 1.5 took effect with the August 20, 2026 data cutoff, and certificates issued against an earlier cutoff still show the list they were issued against: that’s exactly what a dated cutoff is for. But to register you need a current certificate, and a current one is issued against the current cutoff. It helps to know the real size of the change: of the clean verifications we had on record, 13.9% included at least one extra-turn card, so roughly one in seven decks needs an adjustment. The most affected cards are Temporal Trespass, Temporal Mastery and Emrakul, the Promised End. In almost every case it’s solved by swapping a single card, because these effects usually show up one or two per deck and not as the main engine. If your deck really did revolve around taking turns, that deck belongs in Bracket 4 and always did; the difference is that now the rules say so before the tournament instead of finding out at the table.

No, and the distinction is deliberate: the rule reaches cards that GRANT an extra turn, not cards that talk about them. Stranglehold stops opponents from taking extra turns and Trouble in Pairs triggers when someone takes one; neither grants any, so neither violates Rule 1.5 and both remain legal. The difference matters because the list is no longer written by hand: it’s derived from card text, and if the criterion were "mentions extra turns" we would have banned exactly the cards that stop them. In fact, that mistake was already in the previous list (Gerrard's Hourglass Pendant had counted as an extra-turn card for months without granting any), and it was fixed when the list was derived from the text.

You paste your card list in any format (Moxfield, Archidekt, Topdeck.gg, MTGO, Arena or plain text) and the checker instantly analyzes it against every rule of the format. It automatically detects your commander, validates color identity, checks the singleton rule, searches among 1061 banned combos, identifies MLD cards, counts your Game Changers and detects extra-turn cards. Results are shown with card images and clear messages about what to adjust.

The list of 1,061 combos started from EDHREC’s "Early Game 2-Card Combos" category, which comes from the Commander Bracket Combo Vote, and was frozen with the WolfCon 2026 data cutoff, where the format was officially born. For each two-card combo, three questions are answered: can it reasonably be executed within roughly the first six turns? Are the prerequisites (ignoring mana) simple enough to still call it a two-card combo? And is it easy to turn into a win? The 1,061 inherited combos are voted 4-5 on EDHREC: unsafe in Bracket 3. That’s the starting point, not the authority. Since September 2, 2026 the list is no longer rebuilt from EDHREC, and from November 2, 2026 it is decided by a measured formula, not by a vote or by anyone’s judgment: each combo earns points for how fast it works, what it produces, the setup it needs and how much it wins in tournaments, and at 4 points or more it is banned. Changes take effect on the first Monday of each month and every combo’s score is public. Every change is published in the public changelog with its date and its data cutoff. The "six turns" threshold comes from the definition of early game by Gavin Verhey, of Wizards of the Coast.

They’re conditional combos: they’re only infinite if your color identity reaches a certain number of colors. This happens with cards like Bloom Tender and Faeburrow Elder, which produce one mana for each color among your permanents. Bloom Tender with Vigean Graftmage produces 2 mana in a Simic deck and untapping costs 2: it breaks even, it isn’t infinite. In a three-color deck it produces 3 and pays 2, and then it is. The same pair of cards is impossible in one deck and illegal in another, so each of these combos lists how many colors it needs and the checker only flags it if your commander reaches them. There are 13 such combos on the list and the threshold is shown on each one. An honest caveat: the rule looks at your commander’s identity, which is what can be verified in a list. If during the game you end up controlling permanents of more colors than your identity allows (by stealing a permanent from an opponent, for example), the combo would work, and that’s your responsibility at the table, not something the checker can see.

It’s a small, steady group, and we’d rather say so precisely. The 1,061 combos inherited from EDHREC were voted on by a median of 17 people per combo; the middle half of the list ranges from 15 to 19, the minimum is 10 and the maximum 31. Demonic Consultation + Thassa's Oracle, the most-played combo in the format, has 14 votes. The first thing to clarify is how much that weighs in the rules: of the eight restrictions the checker applies, seven don’t depend on any vote. Color identity, singleton, deck size, the Commander banlist, the lands of Rule 1.4 and the extra-turn cards of 1.5 are derived from card text on Scryfall, and the three-Game-Changer limit was set by Wizards. The vote fed Rule 1.1 and nothing else. That small electorate (and an outside one: none of those people play our tournaments) is the reason the list was frozen on September 2, 2026 and stopped following EDHREC. What it did have going for it was consistency: it went through the entire list, so an obscure combo got the same scrutiny as a famous one, and that’s why the inherited list works as a starting point. Since November 2, 2026 no vote decides anymore: Rule 1.1 is applied by a measured formula, the same for every combo, and each combo’s score is public. Every deck is verified against a dated data cutoff: if a monthly announcement moves a combo, your certificate still shows which list it was issued against.

The card database (color identity, legal commanders, images, the land-destruction lands that give nothing back and the extra-turn cards) is regenerated automatically every Monday from Scryfall’s bulk data, so it absorbs every new set without intervention. The banned combo list works differently: no refresh touches it. It has been frozen since September 2, 2026 and, from November 2, 2026, only changes on the first Monday of each month, by a measured formula. Banning a combo changes the legality of decks that are already certified, so every change is announced and published with a dated cutoff. Each tournament states which cutoff it’s played under. If a recent card isn’t recognized, the checker flags it as a warning and can’t certify the deck until it’s fixed.

The checker supports the most popular formats: Moxfield, Archidekt, Topdeck.gg, Mana Box, MTGO and Arena. It also accepts plain text in the "1 Card Name" format. It automatically detects commander sections, handles double-faced cards (DFC), recognizes commanders with Partner and correctly processes foil markers, set codes and categories.

Yes, it’s stored, and we’d rather say so precisely. The checker runs first in your browser to give you an immediate answer, but from each verification we keep the list you pasted along with the commander, the card count and the violations found: without that sample we couldn’t reproduce a verdict or detect that the checker got it wrong. If you also request a certificate, we keep your full decklist and your email, because a certificate that lived only in your inbox wouldn’t be auditable: the judge needs to check it against a source that doesn’t depend on whoever is showing it. None of this is shared with or sold to third parties, and your certificate is only accessible to whoever has its link. If you want us to delete a specific issuance, write to us.

Absolutely. This is a community project that keeps evolving. If you spot a combo that should be on the list, a misclassified card, or you have suggestions to improve the rules, you can contact MTG Wolf México. Feedback from players, judges and organizers is essential to keep improving the format.