# Reading a ship properly

How to read a statline so it tells you what the ship actually does. Worked
through three hulls my own archetype table lumped together as "Wall" —
**H.Barkentine**, **M.Chlor** and **Shark** — which turn out to be three
completely different ships.

The summary metrics in [README.md](README.md) (`dmg/1kPR`, `EHP/1kPR`, `dpt`)
are averages. Averages are the wrong tool for shields, and this file is mostly
about why.

---

## The three, at face value

| | H.Barkentine | M.Chlor | Shark |
|---|---|---|---|
| Race / class | Collective Cruiser | Miner Cruiser | Marauder Cruiser |
| PR / BR / hull | 3,246 / 15 / 5,900 | 1,656 / 16 / 5,000 | 2,052 / 16 / 6,250 |
| Weapon | **1,300 M** | **500 E** | **250 E + 250 K** |
| Shields E/K/M/C | 0.20 / 0.20 / 0.25 / 0.30 | 0 / 0.30 / 0.10 / 0.10 | −0.20 / 0 / **0.40** / −0.30 |
| Range / LR / RF | 7 / 0 / **0** | 6 / 0 / 1 | 4 / 0 / 1 |
| Capture | **100** | 0 | 0 |
| dmg/1kPR · EHP/1kPR | 339 · 2,346 | 298 · 3,488 | 227 · 3,104 |

Read as a row of numbers, Barkentine's shields look the weakest of the three —
nothing above 0.30, no strength anywhere. That reading is exactly backwards.

---

## 1. A shield row has two properties, and the row shows neither

What matters is the **mean under the actual weapon mix**, and the **spread
across opponents**. Computed against all 95 armed hulls:

| | mean incoming | min | max | **spread** |
|---|---|---|---|---|
| H.Barkentine | 0.77 | 0.70 | 0.80 | **0.10** |
| M.Chlor | 0.87 | 0.70 | 1.00 | 0.30 |
| Shark | 0.98 | 0.60 | 1.30 | **0.70** |

**Barkentine has the tightest incoming range of any useful hull in the game.**
Nothing on the roster gets an angle on it — every attacker, of every race, does
between 70% and 80%. Broken out by attacking race the range is narrower still,
0.75 to 0.79. It has **no exploitable hole**, and mild-looking flat shields are
precisely how you get that.

Shark is the opposite. Its 0.98 mean reads "no shields worth mentioning", but it
is a **missile tank with two holes**: 0.60 against missile, 1.20 against energy,
1.30 against chem. The mean is a number that never actually occurs in a fight.

The practical rule: **a flat shield row buys consistency; a spiky one buys a
matchup.** Both are legitimate, and the mean tells you which you have only if
you also look at the spread.

### Which races exploit which hole

| Race | E | K | M | C | shoots mostly |
|---|---|---|---|---|---|
| Neutral | 18% | 37% | 21% | 24% | kinetic |
| Terran | 21% | 41% | 38% | 0% | kinetic |
| Marauder | 26% | 29% | **45%** | 0% | missile |
| Guardian | 29% | 26% | 29% | 16% | even |
| Collective | **39%** | 20% | 21% | 20% | energy |
| Viral | 3% | 33% | 20% | **44%** | chem |
| Miner | 29% | **50%** | 19% | 1% | kinetic |

Incoming multiplier by attacker race:

| | Neutral | Terran | Marauder | Guardian | Collective | Viral | Miner |
|---|---|---|---|---|---|---|---|
| H.Barkentine | 0.77 | 0.78 | 0.78 | 0.77 | 0.77 | 0.75 | 0.79 |
| M.Chlor | 0.84 | 0.84 | 0.87 | 0.88 | **0.90** | 0.84 | 0.83 |
| Shark | 1.02 | **0.89** | **0.87** | 0.99 | 1.05 | **1.06** | 0.99 |

Shark's 0.40 missile shield is aimed squarely at Marauder (45% missile) and
Terran (38%) — and Shark *is* a Marauder hull, so it is partly a mirror answer.
Against Viral (44% chem, straight into Shark's −0.30) it takes 1.06.

M.Chlor's energy shield is **0** — a clean hole, worst against Collective, the
39%-energy race.

---

## 2. Durability is hull per PR, not shields

Decomposing EHP:

| | hull/PR | mean incoming | EHP/1kPR | **share from shields** |
|---|---|---|---|---|
| H.Barkentine | 1.82 | 0.775 | 2,346 | 29% |
| M.Chlor | 3.02 | 0.866 | 3,488 | 16% |
| Shark | 3.05 | 0.981 | 3,104 | **2%** |

Roster median hull/PR is about 0.9. All three are durable because they carry
**two to three times the normal hull for their price** — Shark's shields
contribute essentially nothing (2%) to why it survives.

So "is this a tank?" is answered by hull/PR. The shield row answers a different
question: "who can get through it?"

---

## 3. Damage types are not interchangeable

The roster is not shielded evenly. Mean shield by type, across all 97 hulls:

| Type | mean shield | effective damage per raw point | ships with a negative shield |
|---|---|---|---|
| Energy | 0.001 | **0.996** | 43 (44%) |
| Chem | 0.049 | 0.961 | 38 (39%) |
| Kinetic | 0.062 | 0.921 | 34 (35%) |
| Missile | **0.172** | **0.835** | 30 (31%) |

**Missile is the most-defended type and energy the least.** A point of energy
weapon lands ~19% harder than a point of missile, on average.

That makes raw weapon totals across types incomparable — and it means
Barkentine's 1,300 missile is *not* worth 2.6× M.Chlor's 500 energy. Its real
edge is weapon **per PR**:

| | weapon | PR | weapon/PR |
|---|---|---|---|
| H.Barkentine | 1,300 | 3,246 | **0.400** |
| M.Chlor | 500 | 1,656 | 0.302 |
| Shark | 500 | 2,052 | 0.244 |

---

## 4. Offensive consistency is the mirror of defensive consistency

Barkentine is the most consistent ship in the game to be shot at, and one of the
least consistent to shoot with. Output multiplier across the roster:

| | min | median | max | spread | targets it does <0.3× to |
|---|---|---|---|---|---|
| H.Barkentine | **0.00** | 0.75 | 1.55 | 1.55 | 5 |
| M.Chlor | 0.01 | 1.00 | 1.80 | 1.79 | 4 |
| Shark | 0.01 | 0.97 | 1.52 | 1.51 | 2 |

A single-type weapon is maximally exposed to a single shield value. Barkentine
pays for its flat defence with an all-or-nothing offence.

### The counter map

Output multiplier into the high-shield capital hulls:

| | K.Hun-Zen | K.Hun-Li | G.Fluorite | L.Topaz | G.Quartz | Tourmaline | G.Diamond |
|---|---|---|---|---|---|---|---|
| H.Barkentine | **0.75** | **1.00** | 0.70 | **0.15** | **1.40** | 1.20 | 0.75 |
| M.Chlor | **0.01** | 0.20 | 1.20 | 0.65 | 0.50 | 1.10 | 0.20 |
| Shark | **0.01** | 0.20 | 0.95 | 0.93 | 0.90 | 0.82 | 0.85 |

- **K.Hun-Zen carries +0.99 energy and +0.99 kinetic.** Any pure-E or E/K
  attacker does **1%** damage. M.Chlor and Shark are completely blanked;
  Barkentine, on missile, does 75%.
- **L.Topaz carries +0.85 missile** and blanks Barkentine to 15% — the exact
  reverse.
- **G.Quartz carries −0.40 missile**, so Barkentine hits it for 140%.

Confirmed in simulation (each as defender at equal PR, attacker's % PR lost):
Barkentine takes 60.5% off K.Hun-Zen and wipes Tourmaline outright (100%), while
M.Chlor and Shark take 0.8% and 0.6% off K.Hun-Zen respectively.

### And against fez

| | S.Runner | L.Runner | Queen | King | L.Fodder |
|---|---|---|---|---|---|
| H.Barkentine | **0.000** | 0.750 | **0.000** | 0.100 | 0.500 |
| M.Chlor | 1.000 | 1.000 | 1.000 | 1.150 | 1.100 |
| Shark | **1.475** | **1.475** | **1.475** | 1.150 | 1.425 |

The Runners and the Queen carry **+0.9995 missile** shields. **Barkentine cannot
kill fez** — it does five hundredths of one percent. Shark, on kinetic into a
−0.95 kinetic shield, hits them at 1.475×. Given how much of the meta runs
through fez positioning ([Combat-Mechanics.md](Combat-Mechanics.md) §3), that
single fact separates the two ships more than any summary metric does.

---

## 5. Return fire is worth far less than the arithmetic suggests

`afire = 1`, `dfire = 2`, and return fire lands at half damage — so a wall that
gets attacked twice a wave should deal `1 + 2×0.5 = 2.0×` its base. That reads
like a 100% bonus for `return_fire = 1`.

It is not, because **return fire only triggers when the attacker has
`longrange = 0`** — and **60 of 95 armed hulls (63%) are longrange**. An RF wall
gets nothing back from most of the roster.

Measured over a ten-attacker panel (average attacker % PR lost, with return fire
on vs forced off):

| | RF on | RF off | worth |
|---|---|---|---|
| M.Chlor | 40.5 | 38.1 | **+6%** |
| Shark | 30.5 | 28.8 | **+6%** |
| H.Barkentine | 42.9 | 42.9 | n/a — `RF = 0` |

So `return_fire` is a real but minor stat, and `long_range` on the *attacker* is
what decides whether it exists at all.

Note Barkentine has `RF = 0`: it never retaliates. It is a purely passive
absorber, which is consistent with what it is actually for —

---

## 6. Barkentine is a capture platform, not a wall

`capture = 100`. Every Collective `H.*` hull takes **100% of the eligible ships
it kills** ([Combat-Mechanics.md](Combat-Mechanics.md) §4). Its 339 dmg/1kPR is
not the point; it is a durable, uncounterable, high-range hull whose job is to
land kills that convert into your own fleet.

Judging it on damage per PR — as my archetype table did — measures the one thing
it is not trying to do. This is the clearest case in the roster of a summary
metric actively misleading.

---

## 7. What the three actually are

| | H.Barkentine | M.Chlor | Shark |
|---|---|---|---|
| **Role** | uncounterable capture platform | cheap bulk hull | anti-missile / anti-fez tank |
| Defensive shape | flat, no hole (spread 0.10) | one hole: energy | spiky: missile-proof, E/C-soft |
| Offensive shape | all-or-nothing missile | even, blanked by E/K shields | consistent, blanked by E/K shields |
| Beats | K.Hun line, G.Quartz, Tourmaline | mid-shield bulk | Marauder/Terran missile fleets, **all fez** |
| Blanked by | L.Topaz, **all fez** | K.Hun line | K.Hun line |
| Real durability source | hull/PR 1.82 + shields (29%) | hull/PR 3.02 | hull/PR 3.05 (shields ~0) |

Three ships, one archetype label, almost nothing in common.

---

## Checklist for reading any hull

1. **hull/PR** — is it actually a tank? (roster median ≈ 0.9)
2. **Shield spread**, not just the mean — flat row = no counter; spiky row = a
   matchup and a hole. Check which races shoot the type it is soft to.
3. **Weapon type** — energy lands hardest on average, missile softest; totals
   are not comparable across types.
4. **Weapon concentration** — single-type means an all-or-nothing offence, and a
   9× swing between best and worst target is normal.
5. **`long_range` on the attacker**, not `return_fire` on the target, decides
   whether retaliation happens at all.
6. **`capture`** — a non-zero value can be the whole point of the ship.
7. **Build rate** — see [README.md](README.md) §2; a hull that is 20% worse and
   builds twice as fast is usually the better buy.

The summary tables elsewhere in these files remain useful for ranking a hundred
hulls quickly. They should not be the last step for any hull you actually care
about.