Final Fantasy: The Draft Codex

The Record

what actually happened17Lands archive

Every other chapter here is expert judgment. This one is the account of what followed. Card tiles throughout carry a win-rate chip: the games-in-hand win rate, meaning how often the player won when that card was in hand or in play. Set median is 55.9 percent.

Games analyzed

769,659

Premier draft, whole format

On the play

58.6%

vs 52.9 on the draw

Play advantage

+5.6

points

Average game

9.2

turns

Pair	Party	Win rate	Games
UR	Big spells	57.3%	104,526
WR	Equipment aggro	56.9%	61,693
BG	Graveyard	56.7%	87,880
WB	Sacrifice	56.6%	63,167
WG	Go wide	56.2%	43,682
WU	Artifacts	55.9%	75,662
UG	Towns ramp	55.9%	49,347
UB	Control	55.5%	78,464
BR	Noncreature spells	55.2%	51,479
RG	Landfall	53.8%	30,682

The single most useful correction in this guide: the prerelease advice was to avoid the blue-red big-spells deck unless you were experienced. It finished first, and it was the most-drafted pair in the format by a wide margin. Meanwhile red-green landfall, pitched as a straightforward beatdown plan, finished last and was the least-drafted. If you take one thing from this page, let it be that the expensive-spells deck was real.

The Rankings Were Inverted

the honest scorecardColours are closer than anyone said

The primer ranked the colours black first, then green, blue, red, and white last. Measured across every pair each colour appears in, the order came out very nearly upside down.

Colour	Predicted	Actual	Mean win rate
White	5th, weakest	1st	56.40%
Blue	3rd	2nd	56.15%
Black	1st, deepest	3rd	56.00%
Red	4th	4th	55.80%
Green	2nd	5th	55.65%

But read the numbers before you overcorrect. The gap from the best colour to the worst is three quarters of one percentage point. These colours are close enough that the ranking is nearly noise, which means the primer's own headline advice was better than its rankings: draft the open colour, and use any ranking only as a tiebreaker on a close pick. The lesson is not that white is secretly the best colour. It is that colour rankings in a balanced set are worth very little, and reading signals is worth a great deal.

And the same pattern in the rares

**Ultima**55.6%Called the absolute strongest rare in white. It finished at the 45th percentile, below the set median.

**Summon: Bahamut**57.2%An absurd top of curve, in theory. The 63rd percentile: good, not absurd, and nine mana is a lot to ask.

**Bartz and Boko**56.3%An absurd payoff for bird decks, but the bird deck finished last, and it landed at the 55th percentile.

**Buster Sword**61.4%Called the best of the colorless rares, and correct: the 93rd percentile.

**Memories Returning**60.5%Called the best rare for control, and correct: the 92nd percentile.

**Dark Confidant**60.1%Called the single best high-rarity card in the set. The 90th percentile, so genuinely elite.

The Six Systems

mechanics

Two are new and both change how you count your deck. Learn them before you evaluate a single card.

Job Select newOnly on equipment: when it enters it makes a 1/1 Hero token and attaches itself to it for free, solving equipment's usual weakness. The Hero stays if you move the equipment and the equipment stays if the Hero dies, so count every job select equipment toward your creature count. The Hero tokens are artifacts too, which switches on the artifact deck.

Tiered newA base cost plus three modes, one free and two costing more. You always choose exactly one: unlike Spree you cannot buy several. It is excellent in limited because one card is an early answer and a late kill spell.

Saga Creatures a firstEnchantment creatures that are also Sagas, with all properties of both. They attack and block, but they still sacrifice themselves at the final chapter, so a three-chapter saga creature attacks only once. Treat the body as temporary, and trade it off in combat before the end rather than mourning it.

TownsThe common fixing: ten nonbasic lands that enter tapped and tap for multiple colours. A town appears in the basic-land slot of roughly 55 percent of packs, so you will see them constantly and can fix without spending spell slots.

Adventure, Flashback, KickerAll returning, all strong here for the same reason: flexibility and card advantage. Adventure turns up notably on the rare town lands.

Through the AgesAbout a third of packs carry a bonus-sheet card with alternate art. They are fully legal in your deck and simply look different from the rest.

Two habits worth building now. With the small three-chapter Summons, often the right play is to take a hit or two and let the Saga sacrifice itself, trading your life total for their card. With the big four-chapter one at common, do the opposite and kill it, because it attacks twice for eight damage and you cannot afford to wait it out.

Always Draft These

the reliables10 for 10, the best list in the series

The cleanest prediction any guide in this series has made: every card here landed between the 77th and 94th percentile, tightly clustered and with no misses at all. Flexible cards that any deck in their colour wants, rather than narrow archetype payoffs.

**Dragoon's Wyvern**58.9%A flier plus a ground body, so trading one off still leaves you ahead. Every blue deck wants it, and it wheels far too often.

**Sazh's Chocobo**60.3%A one-drop that grows on every land entering. Cast turn one it is often effectively a 4/5 for one, in ANY green deck.

**Dragoon's Lance**61.6%Job select makes it a two-mana flier, then it re-equips to grant flying to anything. The best-performing card on this list.

**Thunder Magic**58.6%Tiered removal covering all three sizes at instant speed. Every red deck wants it.

**Cloud of Darkness**60.4%A flier that shrinks a creature by your graveyard's permanent count, effectively killing anything in a self-mill shell.

**Choco-Comet**61.2%Early it kills something small and leaves a bird; late it throws seven or more at the face.

**Summon: Fenrir**61.4%Ramps, then adds a counter, then draws. And being a creature you can trade it off for a free two-for-one.

**Samurai's Katana**60.8%A three-mana hasty trampler that later hands the same to anything, and it counts as a noncreature spell for pinger triggers.

**Overkill**59.5%Premium, efficient, unconditional removal. Every black deck takes it highly.

**Shantotto, Tactician Magician**59.6%Grows on every noncreature spell and draws on the big ones. Kill it immediately or lose to it.

The Ten Parties

archetypesStandings from real games

Every pair has a signpost pair of gold legendary uncommons. You do not need them to build the deck, but they are the tiebreaker between two otherwise similar directions.

Big Spells57.3% · best · most drafted

Ramp into expensive noncreature spells

Cast noncreature spells of four or more mana with payoffs that reward each big cast, buying time until they take over. Widely doubted before release as too slow for limited. It was the best deck in the format, and prioritising big spells that also leave a body behind is what makes it work.

The Emperor of Palamecia // The Lord Master of Hell — **The Emperor of Palamecia**59.7%

Equipment Aggro56.9% · second

Job select, then push damage

Cheap aggressive bodies plus job select equipment, which is at its absolute best here because each one brings its own creature. Cost reducers matter: the equip costs are high, so anything that makes attaching cheaper multiplies the whole deck.

Graveyard56.7% · third

Fill the yard, then cash it in

Deliberately stock the graveyard with permanents and convert them into removal, token armies, and huge threats. Self-mill is not optional: none of the archetype's best cards function on an empty yard, so the enablers come before the payoffs.

Sacrifice56.6%

Fodder into value

Use creatures and artifacts as fodder for sacrifice payoffs, grinding value while still hitting hard. It lives or dies on opening a repeatable sacrifice outlet, and there are only a couple, so do not commit to the plan without one.

Go Wide56.2%

Bodies, then a mass pump

Flood the board with creatures and tokens, then convert with counters and team pumps for a lethal swing. Doubted before release as neither fast enough to win early nor grindy enough to win late; it finished comfortably mid-pack, which is a better result than its reputation.

Artifacts55.9%

Critical mass

Build a high density of artifacts and lean on the count payoffs and cost reducers. Aim for roughly half your nonland cards, and remember the job select Hero tokens are artifacts, so the equipment counts twice for you.

Towns Ramp55.9%

The only real splash deck

Accelerate and fetch lands, including towns, then refuel and deploy the payoffs. This is the one archetype whose mana genuinely supports a light third colour high on the curve, because the town duals do the work for you.

Control55.5%

Answer, filter, then win late

Spot removal, card filtering, and a resilient finisher. Its weakness is the early game, so deliberately include cheap low-cost cards: the late game takes care of itself, but a fast start will simply run you over before you get there.

Ultimecia, Time Sorceress // Ultimecia, Omnipotent — **Ultimecia, Time Sorceress**55.5%

Noncreature Spells55.2%

Wizards and pings

Put 0/1 Wizard tokens into play that ping whenever you cast a spell, then chain cheap spells for the kill. Nearly every card should be a noncreature spell or a creature that cares about them.

Garland, Knight of Cornelia // Chaos, the Endless — **Garland, Knight of Cornelia**56.7%

Landfall53.8% · last

Birds and land drops

Creatures that grow or pump on every land entering, chaining extra land drops for multiple triggers a turn. It reads like a clean beatdown plan and it was pitched as one, but it finished last by a full point and a half, and it was the least-drafted pair. Hold extra lands late if you are here.

Required Reading

top commons by colour

White

**Coeurl**54.3%Efficient stats early, and late the extra mana taps down their unenchanted creatures.

Blue

**Stuck in Summoner's Sanctum**55.9%Flash aura that taps, keeps it tapped, AND shuts off activated abilities.

**Ice Magic**58.4%Instant tiered interaction: bounce, tuck, or shuffle away.

Black

**Vayne's Treachery**57.9%Efficient early, and kicks up to kill big things late. Good at both ends of the game.

**Sephiroth's Intervention**58.9%Instant unconditional kill that also gains two life, so an early trade still stabilises.

**Hecteyes**56.6%A body plus discard: value AND fodder to sacrifice, trade, or chump.

Red

**Suplex**57.1%Damage that EXILES, or instead exiles an artifact, which matters with this much equipment.

**Sabotender**52.8%Reach that pings on every land entering, quietly adding three or four damage a game.

Green

**Chocobo Kick**57.9%Fight-style removal, the best green gets, and the kicker doubles the damage so small creatures kill big ones.

**Balamb T-Rexaur**56.4%A big trampling lifelinker, with forest cycling as a bailout when you need land drops.

Party Combinations

two-card synergies

Every piece here is a common or uncommon. The first two are the ones worth actively drafting toward.

Attack with Garnet to remove a lore counter so the Saga never sacrifices itself. You re-trigger it for free lands every turn while Garnet grows, which simply runs away with the game without quick removal.

Get both onto the battlefield and pay five to meld them into Ragnarok: a giant with vigilance, menace, trample, reach, and haste that destroys a permanent and reanimates another. An enormous payoff for two uncommons.

Steal their creature, attack with it, then sacrifice it before the give-back. A temporary theft becomes permanent removal plus a free attack, and any sacrifice outlet works.

The Rod makes the equipped creature a Wizard, so suiting up a big body first turns a card most players treat as pure draw into a huge pump with lifelink.

The Moogle tutors the Down on entry; later you trade the Moogle off and the Down returns it. Nearly removal-proof, and it hands you a second search trigger.

Reanimation needs a target, so forestcycle the T-Rexaur to fill the yard on your terms and fetch a land at the same time. Turn five you rebuy a huge trampler.

Garuda only hits a tapped creature, which is dead weight when you are the aggressor. Coeurl taps one the turn before so it always has a target.

The Wall cannot attack without a counter and its own ability costs six. The Tutorial supplies one cheaply, turning it into an attacking menace as early as turn three.

Guard Your Flank

tricks and traps

White

You're Not Alone: one mana for a large pump, bigger when they have three creatures. The trick to keep in mind most.
Battle Menu: far more flexible than most two-mana spells

Blue

Magic Damper: hexproof, a pump, AND an untap, so a tapped creature ambushes you as a blocker
Syncopate and Swallowed by Leviathan: counterspells see real play, and the second scales as the game goes long

Black

Cornered by Black Mages: a one-mana edict played far more than usual because of the extra body. Keep a spare creature around.
Hecteyes and Malboro: hold a spare land against black so discard cannot strip a real spell

Red

Haste Magic: a pump that also grants HASTE, so a creature attacks from nowhere when you thought you had stabilised
Choco-Comet: ends games from five or six life once they have lands

Green

Blitzball Shot: a large pump with trample
Prishe's Wanderings: leaves a PERMANENT counter and still blows people out

Traps and oddities

Qutrub Forayer: not an instant, but it kills anything damaged this turn. Retroactive deathtouch, and it explains suspicious attacks.
Stolen Uniform: a trap. Limited rarely has equipment worth cheating on, and by the time you set it up you could have paid the equip cost.

The World Map

mana and fixing

This is firmly a two-colour format. Towns do the fixing, and outside of green and the ramp deck you should not plan on a third colour at all.

**Prishe's Wanderings**54.7%Fetches a basic OR a town onto the battlefield and adds a counter, which also triggers landfall. Better than a normal ramp spell.

**World Map**52.3%One of the few reliable NON-green fixers, and worth maindecking at a prerelease.

**Blitzball**51.2%The other common colorless fixer. Only play it with some legendary creatures.

**Torgal, A Fine Hound**58.7%An excellent multicolour mana dork. Green's fixing is exactly why it is the splash-friendly colour.

The Adventurer's Handbook

build and play

Sealed build, in order

Sort as you open, by colour, with separate piles for artifacts, gold cards, and nonbasic lands, then split commons from uncommons so the signposts jump out.
Count job select equipment as creatures. Each one brings a Hero, so your creature count is higher than it looks.
Identify your strongest rares first. Play boosters often give more than one per pack, and that is what separates a good sealed deck from an average one.
Two colours, firmly. The towns ramp deck is the one real exception, and landfall can occasionally splash.
Pick a lens from a strong signpost, then re-read the entire pool through it for removal, filtering, and a top end.
Do not commit to sacrifice without a repeatable outlet, and do not commit to graveyard without self-mill.

Lessons from a 7-1 draft

Take the broken rare and let it define you, but note what the pick would have been without it, so you know how big the gap really was.
Evaluate against your plan, not in a vacuum. Cheap answers become premium the moment your deck's job is buying time.
Double-colour top end cannot be splashed. Commit to it or leave it.
A splash is a hypothesis. Build it, test it in game one, and cut it between rounds if the mana punishes you.
Card advantage has diminishing returns. After two smoothers, take the body instead.
Read what they did not do. An opponent who let your engine resolve probably has no removal, so take the win rather than the extra turn.