| name | macedonian-phalanx-advisor |
| description | Formation-geometry advisor built from Taylor, The Macedonian Phalanx (2020). Computes phalanx frontage, depth, men-to-hold-a-line, sarissa-hedge projection (how many ranks' points reach the enemy), unit breakdown (file -> syntagma -> phalanx), and frontal density vs a Roman line; explains how the phalanx fought and its limits. Use when designing or adjudicating ancient pike/spear formations (wargame rules, scenario setup, what-if deployments) or reasoning about phalanx tactics. |
| allowed-tools | ["Read","Grep","Bash"] |
| argument-hint | ["a formation to compute","a tactics question","or \"selftest\""] |
Macedonian Phalanx Formation Advisor
Source: Taylor, The Macedonian Phalanx: Equipment, Organization and Tactics (2020) · Model: the ancient tacticians (Asclepiodotus/Aelian/Arrian) + Polybius 18.28–32
This skill computes phalanx formation geometry, it does not just describe it. The script scripts/advise.py is the source of truth — its constants live in data/model.json, each cited to the sources. Don't do the formation math by hand; run the script. It validates against the book's own numbers (advise.py selftest).
Companion skill: macedonian-logistics-advisor (Engels) models how the same army moved and ate; this one models how it fought. Same army, same era, two halves of the system.
How to use this skill
- Lay out a formation — give men + depth + order, run
advise.py frontage. Report frontage, depth, files, and feasibility caveats.
- Fit a line — given a frontage to hold, run
advise.py line --length-m L.
- The spear hedge —
advise.py projection --sarissa-cubits 14 → how many ranks' points reach and how far.
- Order of battle —
advise.py unit --men N decomposes into the doubling hierarchy; --name syntagma looks one up.
- Vs Rome —
advise.py density → Macedonians-per-Roman and sarissa points faced.
- Tactics question — read
reference/constraints.md, cite the rule. Numbers at a glance: cheatsheet.md.
The advisory procedure
Gather inputs FIRST, then invoke the script.
Step A — The formation
Ask for: men, depth (ranks; standard 16, early/hoplite 8, doubled 32), and order (open 4 cubits/march, close 2 cubits/battle, synaspismos 1 cubit/locked-shields).
Step B — Run it
python3 scripts/advise.py frontage --men 4096 --depth 16 --order close
python3 scripts/advise.py line --length-m 1000 --depth 16 --order close
python3 scripts/advise.py projection --sarissa-cubits 14 --grip polybius
python3 scripts/advise.py unit --men 9000
python3 scripts/advise.py density --order close
Step C — Interpret
Frontage = (men ÷ depth) × spacing; a 16,000-man phalanx at close order is a 1,000 m × 16 m line. Report the feasibility warnings: only ~5 ranks reach (deeper = push/morale, not reach); the line must stay continuous on flat ground; flanks/rear are fatal. For the hedge, give the per-rank projection and the count of reaching ranks.
Core tactics (how it fought, in seven rules)
Full version with mechanisms + citations in reference/constraints.md. In brief:
- Mass over skill — the formation is the combat atom; the sarissa is driven by the file's weight, not arm-thrust. Once the block breaks, the phalangite is the worst heavy infantryman.
- The hedge — ~5 ranks' points project per front-man (descending 10/8/6/4/2 cubits), giving ~10 sarissae per opposing Roman.
- Depth ≠ reach — only the first ~5 of 16 ranks fight; the rest buy push, morale, and replacements. >16 ranks was thought useless for pushing.
- Flanks / rear / gaps are fatal — one continuous line, no reserves, faces only forward; any gap (terrain, casualties, uneven advance) or flank attack decides the battle (Cynoscephalae, Pydna).
- Three orders — open 4 cubits (march), close 2 cubits (battle), synaspismos 1 cubit (locked-shields defence). The mass advance is the win condition.
- Combined arms — hypaspists/peltasts as the hinge, cavalry on the wings; the phalanx is the anvil, not a standalone army (Magnesia: a phalanx "denuded of cavalry" is surrounded and shot down).
- Command is coarse — drill is atomic at file level but army control is wing-level; a sub-commander's initiative can open a fatal gap.
Index
| Need | File |
|---|
| Compute a formation | scripts/advise.py (+ data/model.json) |
| At-a-glance tables | cheatsheet.md |
| The tactics, with mechanisms + citations | reference/constraints.md |
| Chapter-by-chapter | reference/chapters.md |
| Term definitions | glossary.md |
| Sources, disputes, what's faithful | PROVENANCE.md |
Scope & limits
Faithful to Taylor's text (a clean EPUB — see PROVENANCE.md); the numbers are the tacticians'/Polybius's idealized figures, judged by eye in reality, so treat outputs as formation geometry (±), not survey data. It covers infantry formation/depth/frontage/sarissa-reach/unit structure and phalanx tactics. It does not model cavalry organisation, the legion in detail, casualties, or combat resolution. An agent-consulted design/planning tool, not a runtime engine.
