Commander Deck Rebuild Workflow: Keep, Cut, Proxy, Then Commit

TLDR: Treat a Commander deck rebuild with proxies as a controlled revision, not an invitation to throw 40 exciting cards into the maybeboard and call it progress. Sort the current deck into four buckets: Keep, Repair, Replace, and Test Separately. Give every incoming card a named outgoing card. Freeze the active list only when it contains exactly 100 legal cards, its functional packages have been reviewed, and all unresolved alternatives live outside the deck in a clearly labeled test package.

A Commander deck rebuild with proxies works best when it preserves the deck’s identity while fixing the parts that make it stumble. The process starts with the deck you actually own or play, not a blank page. Keep the commander, themes, pet cards, and engines that still serve the goal; repair weak foundations; replace cards with specific reasons; and quarantine uncertain upgrades until testing earns them a slot.

This is physical-deck version control. Your active 100 is the stable release. The maybeboard is not secretly part of it, tokens are not deck cards, and six possible finishers do not become a coherent win condition because they share a folder.

Set the rebuild target before cutting cards

Write one sentence describing the intended deck before touching the list. Include the commander, main plan, expected game experience, and any theme you refuse to sacrifice. For example: “A mid-powered Muldrotha permanent-value deck that wins through accumulated board advantage, without fast combo lines.” That sentence becomes the standard against which every keep and cut is judged.

Decide whether the commander itself is locked. Changing commanders can alter color identity, card roles, mana requirements, and the deck’s entire reason for existing. If the commander is unsettled, you are comparing separate builds rather than rebuilding one deck. Put the alternate commander and its exclusive package into the test bucket.

Discuss the intended experience with the people likely to play against the finished deck. The official Commander format overview presents Brackets and Game Changers as tools for describing expected games rather than replacements for a pregame conversation. It also summarizes the format’s 100-card construction, singleton structure, and color-identity restrictions. A proxy does not change those construction rules, and access to stronger cards does not automatically make them appropriate for every pod.

If you need more structure for that conversation, the guide to building a Commander Brackets proxy deck that fits your playgroup can help separate raw card strength from the kind of game the table wants.

Use four buckets for your Commander deck rebuild with proxies

Review every card in the current deck once and assign it to one of four buckets. Do not start by browsing replacements. First establish what each existing card is doing—or failing to do.

Bucket	What belongs there	Required action
Keep	Commander-defining cards, proven engines, valued theme pieces, reliable interaction, and payoffs that fit the plan	Leave it in the active list unless a later package audit exposes a conflict
Repair	Cards in an important but underperforming package, such as mana, ramp, draw, removal, protection, or finishers	Review the whole package before choosing individual upgrades
Replace	Cards that miss the plan, strain the mana, duplicate an unnecessary role, or fit a different game environment	Name the reason for the cut and pair it with a specific incoming card
Test Separately	Disputed flex slots, alternate packages, uncertain power upgrades, alternate commanders, and helper pieces	Keep these outside the active 100 and label their intended swaps

Keep what makes the deck worth rebuilding

A keep is not merely a strong card. It is a card whose role and fit are already clear. That can include a commander-specific engine, a dependable removal spell, a payoff the deck can support, or a flavorful card that expresses the theme without materially sabotaging the deck.

This distinction matters because “weak” and “flavorful” are not synonyms. A slightly inefficient tribal payoff may deserve its slot if the deck is intentionally theme-first. A generically powerful staple may deserve a cut if it competes with the commander’s plan or pushes the deck beyond the intended table. Judge a flavor card by whether it contributes enough to the stated experience, not whether an optimized list would select it.

Repair packages instead of swapping cards at random

The Repair bucket is for functional areas that need diagnosis. If the deck routinely misses early plays, replacing one flashy six-mana card with another does not repair anything. Audit the foundation before adding premium staples or finishers.

Mana: Can the deck produce its colors when its important spells need them?
Ramp: Does acceleration match the curve, colors, and commander rather than merely looking familiar?
Card advantage: Can the deck recover after spending its opening hand or losing a board?
Interaction: Can it answer the permanent types and stack-based threats common in its intended games?
Board control: Does it have a realistic reset plan when opponents build faster?
Protection and recursion: Which essential engines need protection, and what is worth recovering?
Win conditions: Can the deck convert its normal game plan into an ending without relying on unrelated haymakers?

Repairing a package often produces several coordinated changes. A color-hungry deck may need different lands, cheaper fixing, and fewer spells with difficult pip requirements. A graveyard deck may need more setup and fewer expensive recursion effects. The goal is not to maximize every category; it is to make the deck’s pieces support one another.

Replace cards with a written reason

Move a card to Replace when you can finish the sentence, “This leaves because…” Useful answers include: it does not advance the primary plan, it is too expensive for the effect, it creates avoidable color problems, it duplicates a role already covered, or it belongs at a different power target.

Avoid cutting only because a card failed in one memorable game. Expensive cards can be mismatched, and modest cards can be structural glue. Focus on role, opportunity cost, and repeated deck behavior. If you cannot explain the cut, move the card to Test Separately instead of pretending the decision is final.

Quarantine uncertainty in Test Separately

The test bucket protects the active deck from maybeboard leakage. It is the correct home for two competing draw engines, an optional combo package, a more powerful tutor suite, alternate win conditions, an alternate commander, or a card whose usefulness depends heavily on the local pod.

Label every test card with its intended swap. “Try X instead of Y” is actionable; “maybe add X” is how a 100-card deck becomes a 113-card philosophical discussion. For a more formal testing cycle, use a change log and the workflow in how to playtest a Commander deck with proxies.

Helper pieces also belong outside the shuffled deck. Track tokens, emblems, counters, monarch or initiative reminders, and double-faced-card solutions in their own inventory. They can be essential to playing the deck cleanly without being part of the active 100. A separate token and emblem checklist is useful when the rebuild creates new board objects.

Build a one-in, one-out replacement map

Once the four-bucket audit is complete, create a replacement map. Every incoming card intended for the active deck must point to one outgoing card. This keeps the count stable and forces each upgrade to solve a defined problem.

Incoming card or package	Outgoing card or package	Reason	Status
Cheaper color fixing	Slow or unreliable mana source	Improve access to early colors	Approved
Commander-specific draw engine	Generic high-cost draw spell	Advance the main plan at a better point on the curve	Approved
Optional high-power interaction	Current flex spell	Depends on pod expectations	Test separately
Alternate-art basic lands	Current basic lands	Cosmetic change only	Approved after final count

Packages still need exact accounting. If three new ramp cards enter, name three outgoing cards. Do not hide extra deck cards under labels such as “options” or “sideboard”; Commander does not turn those labels into additional active slots. Keep alternatives in a separate revision batch.

When should you freeze the 100-card list?

Freeze the list when decisions that affect gameplay are settled enough that a full physical build will remain useful even if you later tune a few flex slots. Perfection is not the threshold. Controlled uncertainty is.

A list is ready to freeze when all of the following are true:

One decklist is the source of truth; duplicate apps, spreadsheets, and handwritten notes agree with it.
The deck contains exactly 100 active cards, including the commander, and follows current color-identity, singleton, legality, and banned-card requirements.
The commander and intended game experience are settled.
Mana, ramp, card advantage, interaction, board control, protection, recursion, and win conditions have been reviewed as packages.
Every approved incoming card has a named outgoing card.
Unresolved choices are limited to a distinct test package with labeled swaps.
Tokens, emblems, counters, and other helper pieces are inventoried separately.
Card versions and treatments are readable enough for normal table play.

Do not freeze if you are still making daily structural edits, arguing over multiple unresolved cuts, changing the land count without testing, or exporting a list that mixes the main deck with a maybeboard. An undecided power target is another stop sign: a rebuild cannot consistently serve two incompatible tables without explicit swap packages.

Before a large order, save a dated snapshot such as “Muldrotha Rebuild v2.1 — 2026-09-10.” Preserve the export, replacement map, revision label, and separate helper/test inventory together. Existing finished-deck guidance likewise recommends treating the deck list and physical build as controlled, checkable versions rather than mixing unresolved alternatives into the active deck.

Prepare the proxy order without reopening the rebuild

After freezing, preparation should translate the approved list into physical cards—not restart card selection. Confirm the commander, card count, versions, readable rules information, double-faced-card handling, and helper pieces. The Commander proxy order checklist covers the final decklist, tokens, DFCs, extras, and arrival inspection.

Choose between a full rebuild and a smaller revision batch based on how much of the deck changed. A full build makes sense when the deck has undergone major structural changes, when presentation consistency matters, or when you want a self-contained physical deck. A revision batch is more sensible when the current shell is stable and only a handful of mapped swaps have been approved. If you use a print-on-demand option such as PrintMTG for the approved proxy list, submit the frozen active deck separately from experimental extras so the two cannot be confused later.

On arrival, compare the physical cards against the dated snapshot before sleeving everything into regular play. Check the commander, total count, replacement map, double-faced cards, and helper inventory. This is the moment to catch an old version, accidental duplicate, or missing land—not halfway through a game while everyone watches you search the deck box like it contains legal counsel.

Common questions about rebuilding with proxies

Does using proxies change the 100-card requirement?

No. A standard Commander deck still consists of one commander and 99 other cards, subject to color identity, singleton rules, the basic-land exception, and current legality. Tokens, emblems, and similar helper pieces are kept outside the deck.

Should mana and ramp be rebuilt before adding finishers?

Usually, yes. Finishers only matter if the deck can cast its spells and reach the stage where those cards function. Audit mana, fixing, curve, and acceleration first, then choose payoffs that the repaired foundation can support.

How should a proxy rebuild be described to the playgroup?

State the deck’s intended game experience and any cards or packages likely to affect expectations. The important disclosure is not the printing method alone; it is what the deck is designed to do, how quickly it can do it, and whether it includes unusually efficient tutors, interaction, resource acceleration, or compact wins.

Can a proxy deck be used at a store or sanctioned event?

Ask the store or event organizer before attending. Casual playgroup acceptance does not establish event eligibility. Wizards’ January 2016 policy explainer said sanctioned events required authentic cards except for narrow judge-issued replacement circumstances during an event, but that document is dated and should not substitute for current organizer guidance.

Commit to a stable version, not a perfect one

The best rebuild keeps the cards that give the deck identity, repairs the packages that make it function, replaces mismatched pieces for explicit reasons, and isolates uncertainty where it cannot contaminate the active list. Once the 100 is exact and the remaining experiments have named swaps, save the snapshot and commit.

You can revise the deck later. That is the point of version control: today’s finished deck remains playable while tomorrow’s ideas wait their turn outside the sleeves.