Alexandria — the Bahamut Attack

Story mode. Written for a first playthrough and kept free of spoilers — it will not tell you what happens here, only what the game is asking of you.

Alexandria is burning and you are not exploring it. This is a corridor with two decisions in it.

Give the soldiers the right orders

Beatrix asks you to assign four pairs of the Knights of Pluto. All four right gets you the Angel Earrings.

• Blutzen and Kohel — gather information
• Weimar and Haagen — protect the townspeople
• Breireicht and Laudo — contact Lindblum
• Dojebon and Mullenkedheim — the cannons

The reward degrades rather than disappearing. None right is a Hi-Potion, one is an Ether, two is an Elixir. You cannot get exactly three — assign three correctly and the fourth is right by elimination.

The scene happens once.

Do not skip the library

There is a boss in it called Tantarian, and this is the last chance in the game to fight it. There were only ever two.

Examine a book on the bottom-left standing bookshelf to start it.

21,997 HP, weak to holy. It fights with its covers shut and is hard to hurt that way; keep hitting it and it opens to the middle pages, and then it is vulnerable. Then it closes again.

Steal first — an Ether, an Elixir, a Silver Fork, and, very rarely, a Demon’s Mail.

It drops Running Shoes guaranteed, plus a Ramuh Card. Running Shoes teach Auto-Haste and there are only two other sources in the game.

The Mistodons

Seven fights, fixed, as Steiner and Beatrix cross the town. They are undead, so Life kills them instantly if you would rather not grind through.

The last one puts Steiner in Trance on its own.

Everything else

There is a moogle in the guest room, which is being used as a hospital, and it is the only save point in the chapter.

There is nothing to pick up. No chests, no gil, no ATEs — eight guides cover this chapter and none of them has a room-by-room list, because there is no room to go back to.

Game Notes

The Knights of Pluto have been a joke for the entire game.

They are the palace guard of a country that has been conquering its neighbours, and every time you have met them they have been losing a fight, falling over, or arguing about lunch. Steiner captains them and is faintly embarrassed by it. The game names all eight of them and gives them personalities so that the joke has somewhere to live.

And then the city burns, and the game asks you — not Steiner, you — to decide what four pairs of them should be doing while it does.

There is a correct answer. Two of them are good at gathering information. Two are strong enough to move people. Two can be trusted to carry a message. Two can work a cannon. You know all of this, if you have been paying attention to a running gag.

That is the whole trick: the game spent twenty hours making you fond of them, and then made the reward for being fond of them an accessory.

Completionist mode. Full spoilers, in play order, with every permanent loss called out before you reach it. 3 entries below are marked where our sources disagree with each other — we have not picked a winner for you.

The ladder is published, and our data confirmed the part that is hard to see

We found four award sites at even spacing in this chapter’s script — Hi-Potion, Ether, Elixir, Angel Earrings — against a research pass that had found two publishers calling the outcome binary. We wrote, on the previous Alexandria page, that this was a ladder nobody had ever observed.

That was wrong. Jegged publishes the exact table, and Final Fantasy Wiki publishes the same one independently:

The correction is written up on that page. What is worth saying here is what survived it, because it is better than what we claimed.

Three is missing from that table, and it is not an omission. Jegged states the rule: you cannot get exactly three correct, because assigning three pairs correctly leaves only one pair and one task, so the fourth is right by elimination.

So the ladder needs four rungs for five conceivable outcomes. Our script has exactly four award sites. We read the shape out of the data without knowing the rule, and the rule explains the shape precisely.

That is the most satisfying kind of agreement this project produces — not our data telling the corpus something, and not the corpus correcting us, but a structural fact and a gameplay rule arriving from opposite directions and fitting.

What the corpus genuinely lacks is player-side corroboration. A search aimed specifically at first-hand accounts — forum posts, trophy guides, talk pages — found nobody reporting having landed on the Ether or Elixir rung themselves. The guides document it; the player-report record does not confirm it. That is a narrower claim than the one we made and it is the one the evidence supports.

Most publishers are silent rather than wrong. They simply do not mention the wrong-answer cases. Only one states an outright binary — Angel Earrings if right, Ether if not — and that is the shape of a guide describing the two outcomes its author has seen.

The assignments

Unanimous across seven publishers:

Nobody documents whether the assignment can be backed out of before it is confirmed. Not one of nine publishers addresses it. The scene happens once and there is no new game plus, so a wrong set forecloses the Angel Earrings for that save — but no source labels it missable, and one publisher’s dedicated missables FAQ flags only Tantarian for this chapter.

Tantarian, and a card that came back

21,997 HP at level 41, weak to holy, unanimous among every source that prints a number.

Trigger it by examining a book on the bottom-left standing bookshelf. It fights shut and opens to the middle pages when hurt enough, and that is the window.

Steals, in order: Ether, Elixir, Silver Fork, Demon’s Mail — four publishers, same ladder, with the Demon’s Mail at about four tenths of one per cent.

Two chances in the whole game, and this is the second. Three publishers say so explicitly, one of them warning that this is the last one and the drop is strictly limited.

A correction to a dispute we were about to publish

A previous pass had one publisher listing a Ramuh Card alongside the Running Shoes and another apparently saying nothing else dropped, and we were going to render that as a dispute.

On a direct re-check, the source that appeared to withhold the card lists it too, in its own card field. Four sources give the card and none withholds it. The “nothing else” reading was an artifact of how the page was first read, not a publisher position.

So there is no dispute. Running Shoes and a Ramuh Card, and the Running Shoes are worth chasing because Auto-Haste has only three sources in the game — this one, the Amdusias fight at Treno’s Knight’s House, and the clock at Quan’s Dwelling. No single guide lists all three.

Two real disagreements remain and neither changes anything: whether Tantarian absorbs shadow or merely halves it, and whether Bio or Fira is the spell to use once it opens.

Seven Mistodons, and they are the small ones

Seven scripted fights, in a fixed order, as Steiner and Beatrix cross the town — four sources count them and agree on the shape, ending with Steiner entering Trance automatically.

They are the level 19 record at 1,473 HP, not the level 56 one that turns up in the endgame. No publisher states that in so many words; it follows from one enemy database tagging the low record’s locations as the Iifa Tree and Alexandria, and the wiki placing the high record only in post-Terra random encounters.

They are undead, so Life ends each fight in one action. No walkthrough recommends stealing, which is a reasonable omission given how fast the fights end — there is an Ore and a Hi-Potion on the table if you want them.

There is nothing else here, and that is the finding

No publisher gives a room-by-room sweep for this chapter. Eight of them cover it in detail and none lists a chest, a gil find or an ATE.

That is not a coverage gap. It is a forced march through a burning city with no backtracking, and the near-total silence across eight independent guides is itself the confirmation. Our field data agrees: the entire chapter’s award content is the Knights of Pluto ladder and Tantarian’s drop.

Zero ATEs, confirmed the same way — the scenes players associate with this part of the game belong to the Lindblum chapter immediately after, and one publisher’s chapter split makes that unambiguous.

Worth stating plainly because a reader trained by twenty-seven areas of sweeping will otherwise spend ten minutes looking for chests in a cutscene.

The moogle’s name

Two publishers call him Mosh; two call him Mogki. Split two-two, nobody reconciles it, and it does not matter — there is one moogle, it is in the guest room being used as a hospital, and it is the chapter’s only save point.

What comes back later

Alexandria reopens after you have the Hilda Garde 3, for a proper sweep — a Sapphire, an Ether, a Remedy, an Amethyst, 365 gil, a Tent, 4,832 gil, a Topaz, a Peridot, another Sapphire, and an Alexandria Card from a girl in the square.

One publisher’s chapter order places that stop immediately after the airship and before Daguerreo, and titles its page for a destroyed city. A second independent confirmation of that placement was not reachable, so we are carrying it as single-sourced and it belongs to a later page.

The jump-rope minigame is not available at that stop — a forum account puts its return much later, and a dedicated FAQ only says “Disc 3 or later” without pinning it down. Weak sourcing, flagged as such.

Game Notes

The four-rung ladder is the closest this project has come to a clean experiment.

We had a structural observation — four awards, evenly spaced, in one script — and no idea what it meant beyond “more outcomes than two.” The published record had a rule we had not read: three correct is impossible, because the fourth assignment has nowhere else to go.

Those two facts were discovered independently, by completely different methods, and they fit exactly. The script needs four rungs because the game can only ever produce four results, and the game can only produce four results because of a constraint on the puzzle that has nothing to do with scripts at all.

Neither side proves the other. But when a structural reading and a gameplay rule converge on the same number for different reasons, that is about as much confidence as you can get without playing it.

Which we will also do.