A design summary, no code and no numbers changed. The full document is design/jobs.md: per variable, what players mean, why it matters in 11th-edition missions, a measurement recipe naming our fields, what's missing, confidence and the dial we already have. Rules from our own notes (design/rules-notes.md, design/mission-notes.md); community vocabulary in our own words; unit names only.
The picture in five lines
- Games are won on objectives, not kills. Every one of the 25 primaries pays from round 2 for a non-home objective; about 60% of the tactical secondaries pay for being somewhere with enough units, about 28% for kills, 11% for actions.
- Players build to a checklist: an answer to tanks, elite infantry, hordes and characters; a home guard and midboard holders; two or three cheap scorers; screens; delivery; leaders and the units that switch on the army's rule.
- Our Value is about 90% Kill. The popularity pointer (T-703) found the units players take more than we rank them draw 20% of Value from non-Kill lines, against 9% for the killers we overrate.
- 11th edition's terrain changes the sums: Hidden (seen only within 15" in terrain, when not firing) makes position, indirect fire and not shooting worth more; the 9" coherency cap makes a screen's worth per unit, not per model.
- Most variables can be built today from fields the ledger already writes; the weakest are the ones that need terrain, variance or a moving opponent.
The 30, by group (priority for T-704 in brackets)
| Group | Variables |
|---|---|
| Damage | anti-tank (2), anti-elite (3), horde clearing (4), character hunting (20), burst and trading up (1), reliability (29) |
| Delivery | speed and threat range (12), melee delivery (11), reserves (9), forward deployment (8), safe firing (27), transport (16) |
| Durability and deterrence | durability per point (5), hard to target (10), mitigation and recursion (23), reactive threat (28) |
| Board control and missions | OC per point (6), holding power (14), cheap activations (13), action-doer (21), skirmish and steal (26), screening (25), zoning and tarpits (30), VP denial (22) |
| Force multipliers | unlocks a squad (7), army-rule engine (15), command points (19), Battle-shock and debuffs (18) |
| Army building | slot competition (17), list dependence (24) |
Confidence: high for the damage classes, Burst, Soak, OC and speed; medium for delivery, holding, leaders, engines, CP, shock, VP denial and list dependence; low for hiding, safe firing, reactive threat, reliability and zoning.
The hard ones, and their proxies
- Screening: units × the 8" arrival bubble around a unit at most 9" long (about 345 square inches, 13% of the table, whatever the squad size) × the field's share of points arriving from reserves or charging; at least one enemy activation to remove. Can't see placement, terrain gaps, or whether anything behind needs screening.
- Zoning and tarpits: threat area × Burst; enemy melee activations needed to remove it × its OC. Jailing barely exists in 11e (T-684). Can't see the opponent's choices.
- Unlocking a squad: the joined form's Value minus the two solo Values, as its own column (not folded into Value, as the five earlier attempts did). Can't see which squad a reviewer meant.
- Army-rule engines: one more carrier's marginal coverage, as synapseAdd does for Synapse; budgets credited by share; once-per-army on one carrier. Can't see a list built around the engine.
- Hiding: a share of enemy fire avoided, from not firing × a dial for standing in terrain × the field's fire from beyond 15". Can't see real tables (T-699).
- Reliability: P(finish its best target in one activation) from repeated dice runs; no variance exists in the ledger.
Interactions worth naming
Delivery × melee damage, leader × squad, transport × rider, screen × backline value, durability × OC, not shooting × Hidden, reach × OC, Burst × Soak; plus saturation: a unit's rank in its army for its job, since a list needs one or two answers per job. A handful of named products, never every pair.
What's next
T-704 takes variables 1 to 25 (20 buildable today from existing fields, five needing a small new field or script). T-705's tags map onto the variable keys (design/jobs.md section 5). A second 11th-edition list per army (T-706), real score sheets, and T-699's terrain would each raise the confidence of the weakest variables.