Mechanics
Fates keeps its arithmetic in five places, and the quickest way to tell them apart is to ask how many units each one is allowed to touch. A boon changes exactly one person. A growth table changes any unit you put in a class. A support rank needs two of them standing next to each other, marriage needs the same two plus a paralogue to collect the result, and recruitment settles the largest question of the lot, which is who you are allowed to field at all. Read in that order, the five pages below stop overlapping and start handing off to each other.
| System | Units in reach | What it fixes | When you set it |
|---|---|---|---|
| Boon and Bane | One, the avatar | One high base stat, one low one, plus growth and ceiling modifiers on a small group of stats each way | Character creation, before the first battle |
| Growth rates and stats | Any single unit | Base stats, stat ceilings and the percentage chance each level adds a point | Whenever that unit changes class |
| Max support | Two, as a pair | The rank between them and the boost they give each other in combat | Battles and castle visits, over many chapters |
| Marriage and children | Two, plus the unit they add | One class and one skill from each parent, and a stat modifier of both parents plus one | The paralogue that opens on an S rank |
| Recruitment | The route's whole cast | Which names exist for you to build in the first place | Chapter by chapter, decided by the path you picked |
One unit: the two picks you make for the avatar
The fifty-two named units on the growth rates page arrive with their percentages already printed. The avatar does not, and you fill two of them in by hand before the first map. Our boon and bane page calls the pair a boon and a bane; our growth rates page calls the same pair an asset and a flaw and tabulates what each one moves. The boon raises one base stat and gives it the best odds of gaining a point on level up, the bane does the reverse, and neither is limited to the stat it names.
The two pages index the same eight traits, and the boon and bane table's numbers line up with the growth column of the growth rates table trait for trait. An asset adds 15 percent growth for HP or Strength or Speed, 20 for Magic, 25 for Skill or Luck and 10 for Defence or Resistance, then hands 5 percent each to two neighbouring stats; a flaw removes 10, 15 or 20 from the headline stat and 5 from the same two neighbours. Which two neighbours move is fixed per trait, and both pages pair them identically: Strength moves Skill and Defence, Magic moves Speed and Resistance, Skill moves Strength and Defence, Speed moves Skill and Luck, Luck moves Strength and Magic, Defence moves Luck and Resistance, Resistance moves Magic and Speed, and HP moves Defence and Resistance.
Ceilings are tabulated on the growth rates page rather than the boon and bane one, and there the pattern has one exception worth knowing before you pick. Seven of the eight assets raise their own stat's ceiling by four and two neighbouring ceilings by two, with the matching flaw taking three off its own and one off each neighbour. The HP asset raises no HP ceiling at all: it adds one each to Strength and Magic and two each to Luck, Defence and Resistance, which is why that page's own example reads as five small gains rather than one large one. Growth is lopsided too. In six of the eight traits the asset gives five more percentage points than the flaw takes away, while Defence and Resistance sit level at ten in both directions.
As for what to take, the boon and bane page gives three loadouts and one piece of advice for a first run. Skill is its pick for a Master Ninja, for the sake of Lethality; Magic is its pick for a Dark Knight; Resistance is what the writer ran on a Malig Knight. The banes it names are Strength on that same Master Ninja, on the grounds that Lethality covers the loss, and Defence on the Dark Knight, since mounted units come with enough of it. For anyone new to the series it recommends the pair it labels Calm and Dull, chosen to keep magic damage survivable, and it warns off the traits labelled Sickly and Excitable unless a glass cannon is the point.
A third creation choice is easy to miss because it is filed under traits rather than stats. It fixes the avatar's secondary class and also an alternate that you can never enter yourself, only pass on through a seal or down to a child. The list runs to seventeen entries and it is almost perfectly reciprocal: fourteen of them pair off two ways, so Cavalier points at Ninja and Ninja points back at Cavalier, as do Knight and Halbredier, Axe Fighter and Oni Savage, Mercenary and Samurai, Thief and Archer, Spellweaver and Dark Mage, and Pegasus Rider and Dragon Knight. The three that break the pattern break it completely, with no alternate listed at all.
One unit at a time: the class carries most of the number
A growth rate is a probability and nothing more. If a unit sits at 70 percent Strength, the game rolls for an extra point of Strength every time it levels, and it wins seven times in ten. The figure you actually roll comes from three places: the modifiers above, the unit's own line, and the class it is standing in, and the growth rates page is explicit that the class figure and the character figure are added together.
That makes the class tables the heaviest thing on the page. There is a base stat table of sixty-two classes with a movement column, a ceiling table of another sixty-two, a class growth table of sixty-three, and a first generation list of fifty-two named units with their personal percentages. Sixty of the classes appear in all three tables, which is enough to do the arithmetic for almost anything you field, and the gaps are worth knowing: the ceiling table has no Fox Spirit row, the base stat table has no Pegasus Knight row, and one row changes name between tables. Two naming sets run through our pages, one closer to the Japanese release and one to the English one, so a class or a unit you cannot find under one name is often sitting a few rows away under the other.
Changing class is cheap in levels and expensive in skills, which is the part that catches people. A re-class keeps the level it found you on, and instead of handing over the skill that matches that level, the game generally unlocks the lowest one you have missed. Arrive in Samurai at level ten and you get the level one skill first and the level ten skill on your next level up. Skip a base class entirely and its skills are simply gone: the page's own example takes an Apothecary through Mechanist to Elite Ninja and points out that nothing from the Ninja line ever arrives, so spending a few levels in each base class a unit owns is how you end up with a full skill list. Promotion is the other half of it. A Master Seal steps a basic class up into one of its derived classes, and the handful of special classes with nothing to promote into turn into promoted units on their own at level twenty-one.
Two units: the rank the other systems spend
Support has to be earned by two units at once, which is what makes it slow, and our max support page lists six ways to raise it. Pairing the two on the battlefield is the fastest, and it pays twice, since a paired unit gets a temporary stat boost on top of the relationship; the page adds that a pair which fights and wins together bonds more strongly. The other five open up at different points. The castle arrives in Chapter 6, and from then on you can walk up to whoever is standing around and speak to them, with the faces on duty changing from one day to the next. Handing over an accessory works whenever a character asks for one, with no chapter attached. Inviting a character who has a heart beside their name needs the private headquarters the page places early in the game, and the hot baths have to be installed in the base before anyone can use them. The support conversation is not a castle job at all: the game offers one after each battle. Two details save real time. A single accessory can be shuffled between characters rather than one being kept per character, and inviting a unit to your private quarters raises the chance of those post battle conversations.
The page also documents a grinding loop, and it is built on the fact that support does not care about damage. Take Challenges and Scout maps rather than story chapters, pair everyone up, and attack a unit that has no weapon equipped: the relationship rises on a hit that deals zero, and because the exchange awards no experience nobody creeps past the level you wanted them at. Run as many pairs as you can for ten to fifteen turns before somebody captures the castle and ends the map, then do it again.
Two ranks matter beyond the boost, and for those it is the growth rates page rather than the support one that sets out the rules. An A+ rank with a unit of the same gender lets you borrow that unit's class through a Buddy Seal, and an S rank with a unit of the opposite gender lets you borrow a spouse's through a Marriage Seal. Zero and Syarra are the two names that sit outside the gender rule there, a male avatar reaching S rank with the first and a female avatar with the second, and they are also the pair excluded from the same gender route to A+. The avatar gets a second exception of its own, because it cannot gain A+ supports at all and so uses the Buddy Seal on an A rank instead. That page also notes that every character carries its own list of who it can reach those ranks with, which is why the marriage page below is mostly lists.
Two units and a third: what a pair adds to the army
An S rank is where the pair stops being a pair. Our marriage page notes that the paralogue carrying the couple's child opens as soon as the rank lands, with no story progress attached, which is the change this game made to the system it inherited. Win the paralogue and the child joins as a unit, grown rather than young, and recruited without a promotion behind it, so a Child Seal may be wanted later to move it up a class.
There are twenty-one children on that page, in paralogues two through twenty-two, and each one has a fixed first parent plus a list of who the second can be. The lists are not the same length, and their shape tells you where the flexibility is. Kana's second parent is given as anyone at all. Four children carry a twenty name list, twelve carry fourteen names, and four carry only twelve. The four twenty name lists sit in paralogues three to six, behind Azura, Jakob, Silas and Kaze, while the four twelve name lists belong to the children of Ryoma, Takumi, Xander and Leo, which is where you have the least room to manoeuvre. What the child takes home is set by its parents: one class and one skill from each parent, a maximum stat modifier equal to both parents added together plus one, and the mother's hair colour, with Kana named as the exception. Classes arrive in a stated order, the child's own first, then one from the father and one from the mother, with a published priority chain deciding which of a parent's classes gets handed over, and unique classes such as Songstress never pass down.
The route caps this system harder than anything else in it. Counting the child entries on the recruitment page gives thirteen on Birthright and thirteen on Conquest, and five of those are the same five on both, which is exactly how a single path of either kind reaches thirteen and Revelation reaches all twenty-one. Only the Birthright list numbers those entries as paralogues; the Conquest and Revelation lists give the same children a chapter number instead, so match them by name rather than by label. The total lines up on three of our pages independently: twenty-one children on the marriage page, twenty-one child entries on the Revelation list, and twenty-one rows in the child growth table.
The whole roster: who you are allowed to build
Everything above assumes the unit is already standing in your camp, and that part is settled by the path rather than by you. Our recruitment page walks three lists: forty-two entries for Birthright, forty for Conquest and sixty-six for Revelation, with thirteen bosses handled separately. Fifteen names appear on both the Birthright and the Conquest list, so the two paths together account for sixty-seven, and Revelation is that set with two of them absent and one name, Fuuga, that neither single path offers. Whichever path you take, the army you will be pairing and re-classing for the rest of the run is decided by that purchase.
Three kinds of entry sit outside the ordinary chapter arrival. The thirteen bosses each need convincing in the chapter they appear in, which runs from Chapter 2 to Chapter 21 on that list. Each of the three lists holds two entries gated on a My Castle building at level three rather than on a chapter, one of them available from Chapter 19 onward and the other from Chapter 23 onward. Four names arrive that way across the three lists: Izana, Yukimura, Flora and Fuuga. The facility is not the same in every entry, a hot spring for Izana and for Fuuga, one of the Puppet buildings for Yukimura and a turret for Flora, and three of the six entries add that the unit turns up a battle after the upgrade, so read the entry for the path you are on before you spend on a facility. And the avatar's gender decides which of the maid and the butler joins in Chapter 2, because the other one waits until the end of Chapter 15.
Once you know who is in the army, the class chart, the seal locations and the per character builds are the next things worth reading, and all three sit on the Fire Emblem Fates guide alongside these five pages.