A grognard exercise, not a scientific model. What that means

War journal

Cases where play or the model throws up something the rules do not cover, and the ruling reached. 2 so far.

Shown as it stands in the war journal's repository, at 8de4441. Open cases are argued on its discussion board.

The journal records cases where play or the model throws up something the rules do not cover, and the ruling reached. Each case says what happened, what the question is, the argument, and the rule that follows.

Rulings

Case Date Where it came up Proposed ruling Status
1. Canada holds Sweet Grass, the US holds Lethbridge 18 Nov 2026 Border-battle prototype Lethbridge decides the win. Add supply. A US win with a base lost is marginal, not decisive. Open: Poll 1, Poll 2
2. JTF 2 sits in Sweet Grass for twelve days 19 Nov 2026 Border-battle prototype, JTF 2 variant None yet. What does a special-operations detachment do with twelve days behind the line, and how does the American public react? Open: estimates pending

Taking part

Open cases are argued on the journal's discussion board. Each open case has a poll and a thread. Vote, and say why: the reasons matter more than the count. Kevin decides, and the ruling records what the discussion said and where it changed the outcome.

Who writes this

Kevin Duska sets the direction and decides each case. The cases are written up with Claude, an AI system made by Anthropic, which acts as the model judge: it proposes a ruling and argues for it. Where AI agents were used to simulate press coverage or memes, the case says so, says how many, and says what they were told.

What this is not

It is a simulation and a game. Nothing here is a prediction, a plan or advice. The press coverage and memes are fiction, name no real person or outlet, and are marked as fiction.