Pick two states off the ForeKaster board. Your browser derives each side's strength from the sourced spine — same rules as the desk's own tool: real personnel figures only, parse refusals listed instead of guessed — and runs the deterministic FogSim engine locally. Same seed, same run, forever. This is a model being exercised in public, never a claim about any conflict.
SIMULATED — this page represents a model, not a conflict picture. Lanchester square law with symmetric effectiveness; the only asymmetry is documented personnel. Outcomes are properties of the toy, not forecasts.
BLUE
RED
Only states whose personnel figure parses under the desk's published rules are selectable — … of …. The refused remainder is listed at the bottom of each menu with the refusal reason, disabled: an unparsed figure is a gap on the record, not a zero.
seed
Personnel is the only force figure this desk has a source for. Armour, aircraft, missile volume, air defence and readiness are not on the record — so this panel does not pretend to supply them. You type them, and every field you touch is stamped USER in the run artifact while the personnel figure stays SOURCED. A wargame that cannot tell you which of its numbers came from a document is a wargame that launders invention as intelligence.
This panel runs gk-multidomain/0.1 — not the sealed fogsim-lanchester/1.0 engine above. Sealed campaigns depend on that engine being frozen, so this one is a separate, unsealed sandbox. Nothing run here is eligible for a sealed campaign until it is ported to fogsim.py and re-verified.
| parameter | BLUE | RED | provenance | what it does in the model |
|---|
seed
Any wargame can produce a winner. The question that matters is whether the winner is a property of the forces or a property of your guesses. Which input decides this? re-runs the whole model with each parameter alone moved ±25%, on fixed seeds, and reports which single number moves the outcome. When one invented field flips the result, the run is telling you about that field and nothing else — a conditional zero-point sitting inside your own map, printed instead of hidden.
Strengths are recomputed in this tab from the spine's Factbook personnel strings by the same parser the desk ships: first-figure anchored, year forms stripped, range low-ends inheriting the high end's scale, anything under 500 or over 5,000,000 refused rather than published. The engine is a byte-faithful port of fogsim.py's run_model — 64-bit LCG, Lanchester square, break at 55% — so any run here re-executes from (scenario, rules_version, seed) alone, which is the property the sealed campaigns stand on. Effectiveness is a fixed constant on both sides by design: no readiness, no equipment, no doctrine. That honesty is the feature.