Sisu devlog #1: spaced repetition math

First real devlog entry: implementing the SM-2 scheduling algorithm and immediately disagreeing with it.

SM-2 assumes every card is independent. Finnish vocabulary isn’t — if I just failed talossa (“in the house”), showing me talosta (“from the house”) thirty seconds later is a freebie, and the algorithm shouldn’t reward it like a genuine recall.

interval(n) = interval(n-1) × easeFactor
easeFactor  = max(1.3, EF + (0.1 - (5-q) × (0.08 + (5-q) × 0.02)))

The tweak I’m testing: cards in the same word family share a cooldown — a correct answer within 10 minutes of a sibling review counts at half weight.

Next up: deck import from my course’s word lists.