Historoll · Episode 01 companion

Hold Highmoor

Highmoor is an invented but typical 13th-century English stone castle: an outer wall with towers, and behind it an inner ward and a keep. Inside are 100 fighting men and 100 civilians, 4 months of food and one well. Outside, 3,000 attackers have just arrived. Change the castle, play one siege day by day, then run it 1,000 times.

The episode's what-ifs
Inside the walls
Rochester held out in 1215 with 95–140 men, Caerlaverock in 1300 with about 60. Food is stocked for the garrison plus 100 servants and families.
Months of full rations for the planned household. The episode assumes 4 months; nobody has good numbers for how much food castles actually stored.
Extra mouths the stores weren't planned for. At Château Gaillard in 1203 the townspeople crowded into the castle; Kenilworth held 1,200+ people in 1266.
Exeter's wells dried up in 1136. Nobody has numbers for how often wells failed; the episode assumes 0.2% a day.
Outside the walls
The episode assumes 3,000, a guess for a mid-sized royal army. Calais was besieged by up to about 32,000 at its peak (1346–47).

Day
—
Press play
 
Food
—
Defenders
—
Attackers
—
Outer wall
—
Hunger
—

This siege, day by day

Food left Defenders fit to fight Attackers Wall Fallen back to the keep

    How the simulation works

    Every simulated day the same things happen in order. Everyone inside eats. We check the well. Sickness spreads a little, faster when people are hungry. Outside, the attackers eat whatever supplies arrive and lose men to camp disease. The trebuchets throw stones of up to about 300 lb (136 kg) (Illumin, USC), the defenders repair what they can, the miners dig, and if there is a gap in the wall the attackers decide whether to charge. If the outer wall is stormed, the survivors fall back to the inner ward and keep and fight on. Then it starts again with new dice.

    The castle falls when fewer than 3 men are still fit to fight, or when the last line is stormed. The defenders win when the attackers drop below 4 men for every defender and below half the army they arrived with, too few to keep the siege going. This page runs a JavaScript copy of the episode's simulation (siege.py); with the episode's settings it gives the same results within the normal run-to-run spread.

    Guesses, stated plainly

    • Well: fails with a 0.2% chance a day; the cistern then lasts 12 days before thirst sets in. No source has failure rates.
    • Rations: cut to two-thirds once the stores fall below 30%.
    • Walls: each trebuchet knocks 0.25–0.75 points a day off a wall of 100, scaled by the square root of the share of the army still alive; a fed garrison repairs up to 0.6 points a day. A mine takes 14–44 days and starts on day 5–19. Each day of digging the defenders have a 0.3 ÷ (days the mine takes) chance of finding and collapsing it, so about a quarter of mines (26%) are found.
    • Assaults: with a breach open the attackers charge on 35% of days; any day there is a 1% chance of a ladder attack. A storming party is 30% of the army; each defender fit to fight counts as 3 attackers at a breach, 30 on an intact wall, and 1.6 times more in the inner ward.
    • Outer ward lost: 10–25% of the fighters, 20% of the civilians and 40% of the food go with it.
    • The attackers: get 65–100% of the food they need at the start, slowly less as foraging gets harder. Every siege rolls its own sanitation luck (1 in a typical siege, below 0.4 or above 2.6 in one siege in ten). Camp disease kills 0.12% of the army a day plus 0.12% × that luck, so about 0.25% a day in a typical siege; after day 30 the second part rises to 0.4% × luck, roughly doubling the toll. Hungry armies lose more.
    • Summer plague (what-if): from day 60 to day 150 the luck-scaled part is four times bigger, about three times as many deaths a day in a typical siege.

    Sources

    Garrisons and the case-study sieges

    Food, water and disease

    Armies, weapons and how sieges ended