# Race Knowledge Base — index, turn economy, archetypes

Working reference for PvP balance. One file per race, Neutral through Aspha
Miners. NPC races (7+) and the Corruption event are out of scope.

> **Read [Combat-Mechanics.md](Combat-Mechanics.md) first.** It carries the
> stack/flanking/capture rules and corrects several mechanics stated wrongly in
> the earlier balance work — including firing order, what "flanking" actually
> is, the Special/Strafez class that was excluded from my analysis, and the
> capture system I had missed entirely.
>
> **Then [Ship-Reading-Guide.md](Ship-Reading-Guide.md)** for how to read an
> individual statline. The summary metrics below rank a hundred hulls quickly
> but mislead on any hull you actually care about — shield *spread* matters more
> than shield mean, durability comes from hull/PR rather than shields, and
> damage types are not interchangeable. Worked through H.Barkentine, M.Chlor and
> Shark, three ships this file's archetype table wrongly calls the same thing.
>
> **And [High-PR-Dynamics.md](High-PR-Dynamics.md) before applying any of it at
> the top end.** Measured on twelve real 90M–270M PR fleets, a stack's own
> stats correlate with its survival at **r ≈ 0.07**; the PR of the stack it is
> paired against correlates at **−0.56**. Battles there are ten near-independent
> duels — 72% of stacks end at 0% or 100% casualties — and stack *sizing* beats
> ship selection. Everything below is scoped to who wins a given pairing, not
> who wins a high-PR battle.

| File | Race | One-line identity |
|---|---|---|
| [0-Neutral.md](0-Neutral.md) | Neutral | Universally available filler. Cheap per PR, worst per turn. |
| [1-Terran.md](1-Terran.md) | Terran | Deepest roster, generalist, no bad matchups and no great ones. |
| [2-Marauder.md](2-Marauder.md) | Marauder | Best damage per PR in the game. Fragile, slow to accumulate. |
| [3-Guardian.md](3-Guardian.md) | Guardian | Fastest build economy. Weakest per PR. Capital-hull race. |
| [4-Collective.md](4-Collective.md) | Collective | Second-best tempo, widest class spread, no extremes. |
| [5-Viral.md](5-Viral.md) | Viral | Four native hulls. Its real roster is what it steals. |
| [6-Aspha-Miners.md](6-Aspha-Miners.md) | Aspha Miners | Largest roster, bulk and attrition, high internal variance. |

---

## 1. Class means nothing

`ship_type.class` (Fighter, Corvette, … Leviathan) is a **display label
indicating rough size**. It is not a balance bracket. Nothing in
`f_com_attack3.cfm` reads class except the starbase upkeep multiplier and the
(broken) reverse-eligibility check.

Comparing a ship "against its class mean" is therefore meaningless — a Guardian
Corvette (G.Amethyst, 18,018 PR) and a Terran Corvette (T.Maru, 51 PR) are 353×
apart in unit size and share nothing but a word. **All comparisons in these
files are against the whole buildable roster, or within a race.**

## 2. The turn economy

Turns regenerate **1 per 5 seconds**, capped at a **90-turn bank**.

| | |
|---|---|
| Regen rate | 720 turns/hour, 17,280/day |
| Bank cap | 90 turns — refills in 7.5 minutes |
| Implication | A player who logs in *N* times a day and spends the bank gets **90N turns**. A player sitting on the build page continuously gets 17,280. |

`ship_type.reqturn` is the **build rate (BR)** — units produced *per turn*, not
turns required. The `buildmod` multiplier on it (`application.mod<server>_51`)
is a defunct server effect; treat it as 1.

Three derived metrics do the work in these files:

| Metric | Formula | What it answers |
|---|---|---|
| **ppt** | unit PR × BR | How fast this ship grows your power score |
| **dpt** | effective damage × BR | How fast this ship grows your *combat capability* |
| **bank90** | ppt × 90 | PR added by one full 90-turn bank |

`effective damage` is mean damage after shields against the whole buildable
roster, so it accounts for weapon-type matchups rather than raw weapon totals.

### Why this matters

`ppt` spans **5,100 → 101,858** across the roster — a 20× spread. One full
90-turn bank buys 459k PR of C.Aries or 9.2M PR of G.Diamond. Two ships that
look equivalent at a fixed PR checkpoint can be four hours apart in reality.

Build rate is almost perfectly inverse to unit size (r = −0.982 on log scale),
but **not quite** — ppt still rises with unit PR (r = +0.63). Capital hulls
reach a given PR faster. That is the central trade in the game.

## 3. The two framings, and why they disagree

Everything depends on whether you hold **PR** or **turns** constant.

Combined = attack win% + defence repel%, simulated, random 6-stack fleets,
**Special/Strafez hulls included**:

| Race | Equal PR | Equal turns | Average |
|---|---|---|---|
| Guardian | 94.8 | **121.2** *(first)* | **108.0** |
| Miner | **109.4** *(first)* | 104.6 | 107.0 |
| Marauder | 101.3 | 105.7 | 103.5 |
| Terran | 96.2 | 99.7 | 97.9 |
| Viral | 106.2 | 85.1 *(last)* | 95.7 |
| Collective | 92.2 *(last)* | 83.7 | 87.9 *(last)* |
| | sd 6.2 | sd 12.8 | sd 7.0 |

Excluding the Special hulls — as earlier drafts did — moves this a lot: Viral
reads 117.8 at equal PR instead of 106.2, Guardian 130.9 at equal turns instead
of 121.2, and equal-PR sd rises from 6.2 to 9.3. The fez line is a genuine
equaliser because every race can field it. Both tables are in
[../Ship-Balance-Analysis.md](../Ship-Balance-Analysis.md) §5.

Neither framing is "the truth":

- **Equal PR is what matchmaking enforces.** You fight people near your power,
  so at any given bracket damage per PR decides the fight.
- **Equal turns is what the player actually spends.** Build rate decides how
  fast you enter a bracket and — more importantly — **how fast you replace
  losses after a battle**.

The PR advantage is partly self-neutralising: building power faster moves you
into a harder bracket rather than letting you keep an edge. What does *not*
self-neutralise is **recovery speed**. A race with 1.7× the ppt rebuilds a
shattered fleet in 60% of the time and can attack more often. That is a real,
permanent advantage that an equal-PR analysis cannot see.

**Rule of thumb used throughout these files:** judge combat strength at equal
PR, judge sustainability at equal turns, and never call a ship weak on one axis
without checking the other.

## 4. Archetypes

Derived from percentile position across all 91 buildable player ships on damage
per PR, EHP per PR, ppt, dpt and unit PR.

| Archetype | n | Signature | Median dmg/1kPR · EHP/1kPR · ppt · dpt |
|---|---|---|---|
| **Elite** | 15 | Top-third damage per PR *and* top-third dpt. Good at a checkpoint and fast to field. | 973 · 867 · 42,000 · 35,032 |
| **Glass cannon** | 13 | High damage per PR, bottom-third durability. Trades survivability for output. | 715 · 796 · 27,520 · 22,218 |
| **Tempo** | 7 | High dpt without elite per-PR damage. Converts turns into force fast. | 530 · 1,274 · 60,420 · 32,438 |
| **Growth engine** | 13 | Huge unit PR, BR 1–3, top-third ppt, below-median damage per PR. Score accumulators. | 377 · 2,064 · 80,316 · 26,875 |
| **Line** | 21 | Mid on everything. The default body of a fleet. | 511 · 1,397 · 34,176 · 17,592 |
| **Wall** | 15 | Top-third EHP per PR, bottom damage. Cheap hull that absorbs pairings. | 298 · 2,693 · 32,832 · 8,408 |
| **Swarm filler** | 7 | Tiny unit PR, BR ≥ 100. Maximum granularity. | 632 · 1,300 · 36,025 · 22,202 |
| **Dead weight** | 1 | Bottom-third on damage, durability *and* dpt. | 375 · 1,172 · 29,835 · 11,175 |

### Archetype mix by race

| Race | Elite | Glass | Tempo | Growth | Line | Wall | Swarm | Dead |
|---|---|---|---|---|---|---|---|---|
| Neutral | 1 | 1 | 0 | 0 | 4 | **6** | 1 | 0 |
| Terran | 1 | 4 | 1 | 2 | 3 | 3 | 2 | 0 |
| Marauder | **6** | 3 | 0 | 0 | 3 | 2 | 0 | 0 |
| Guardian | 2 | 2 | 2 | **6** | 2 | 0 | 0 | 0 |
| Collective | 2 | 1 | **3** | 1 | 2 | 2 | 2 | 0 |
| Viral | 0 | 1 | 0 | 1 | 2 | 0 | 0 | 0 |
| Miner | 3 | 1 | 1 | 3 | **5** | 2 | 2 | 1 |

The mix *is* the race identity. Marauder is six Elite hulls and no Growth
engines; Guardian is six Growth engines and no Walls. That is a deliberate
opposition, not an accident.

> **Archetypes are a sorting aid, not a description.** They key on averaged
> `dmg/1kPR` and `EHP/1kPR`, which collapses genuinely different ships. The
> "Wall" bucket alone contains H.Barkentine (an uncounterable capture platform
> with `capture = 100`), M.Chlor (cheap bulk hull with an energy hole) and Shark
> (an anti-missile tank that hits fez at 1.475× while Barkentine cannot scratch
> them). See [Ship-Reading-Guide.md](Ship-Reading-Guide.md).

## 5. Cross-race summary

Medians of each race's own buildable ships. Rank 1 = best.

| Race | n | ppt | dpt | dmg/1kPR | EHP/1kPR | bank90 | Top dpt hull |
|---|---|---|---|---|---|---|---|
| Neutral | 13 | 27,900 ⁷ | 10,149 ⁷ | 398 ⁷ | 1,853 ¹ | 2.51M | Light fighter-drone (34,537) |
| Terran | 16 | 42,786 ⁴ | 22,421 ⁴ | 611 ³ | 1,434 ² | 3.85M | Tyr (42,714) |
| Marauder | 14 | 38,016 ⁶ | 27,424 ³ | **717 ¹** | 1,092 ⁷ | 3.42M | Angel (67,739) |
| Guardian | 14 | **63,805 ¹** | **29,603 ¹** | 499 ⁵ | 1,396 ⁴ | **5.74M** | Tourmaline (50,811) |
| Collective | 13 | 40,500 ⁵ | 29,530 ² | 530 ⁴ | 1,300 ⁵ | 3.65M | H.Overlord (53,162) |
| Viral | 4 | 53,892 ² | 22,018 ⁶ | 618 ² | 1,224 ⁶ | 4.85M | V.Borrelly (22,218) |
| Miner | 17 | 47,628 ³ | 22,202 ⁵ | 457 ⁶ | 1,417 ³ | 4.29M | M.Lakko (64,111) |

**Neutral is the only race that is bottom-ranked on every axis that predicts
winning** (ppt, dpt, damage per PR). Its single first place is EHP per PR, which
has no measured relationship to winning battles (r = −0.04, see
`../Ship-Balance-Analysis.md` §2.1). That makes Neutral the clearest genuine
weakness in the game, not Guardian.

## 6. Roster notes and scope

- **G.Livid is a mission reward, not buildable.** It has `upkeep = 0` and BR 1
  (ppt 13,013, less than a quarter of `G.Livid (r)` at 50,052). It is excluded
  from all roster statistics in these files. **`G.Livid (r)` is the Miner
  fighter that matters.** G.Livid is also the only `upkeep = 0` player hull.
- **The 13 `UW.*` Neutral hulls have BR 0** and cannot be built by anyone.
  Excluded throughout.
- **Scouts (class 21) are unarmed** and never fielded in combat. Excluded.
- Buildable player roster: **97 hulls** — 91 of class 1–10, plus the 6
  Special/Strafez hulls (all Neutral).

> **Scope caveat.** The per-race medians, archetypes and the §4 archetype table
> below are computed on the **class 1–10 pool only** — they predate the Special
> correction and have not been recomputed. The race-balance table in §3 *does*
> include Special hulls. [0-Neutral.md](0-Neutral.md) reports both. Where a
> figure could be either, it is labelled.

## 7. Metric reference

| Term | Meaning |
|---|---|
| BR | `ship_type.reqturn` — units built per turn |
| ppt | PR per turn = unit PR × BR |
| dpt | Effective damage per turn = effective damage × BR |
| bank90 | PR from one full 90-turn bank = ppt × 90 |
| dmg/1kPR | Effective damage per 1,000 PR of fleet — decides fights at a fixed bracket |
| EHP/1kPR | Hull ÷ incoming multiplier, per 1,000 PR — **does not predict wins** |
| Effective damage | Mean post-shield damage vs the whole buildable roster |
| Incoming multiplier | Mean damage taken vs the roster's weapon mix (E/K/M/C ≈ 29/31/29/11) |