Attrition versus Spike
Two encounters can have identical expected damage output and produce completely different games. The difference is variance, and it matters more than the average.
01The same number, two completely different fights
Consider two monsters. The first deals two points of damage every round, without fail. The second swings once for ten points but only lands one time in five. Over five rounds, both deal an expected ten points. A spreadsheet cannot tell them apart. A table can.
Key distinction
- 01Attrition encounter — damage is steady and predictable; players manage resources across time
- 02Spike encounter — damage is infrequent but large; outcome is shock and reaction rather than planning
The grinding attacker is a resource problem. Characters bleed slowly; healers and recovery abilities drain at a predictable rate; the referee can tune the encounter like a dial. Players know roughly how long they can sustain it, and decision-making opens up — push forward, fall back, expend that charge now or later. The fight becomes a resource budget with skin on it.
The spiking attacker is something else entirely. Four rounds of nothing, then a character drops — possibly dies — in a single exchange. The expected value is identical, but the experience is shock and luck rather than management and choice. High variance compresses decisions: before the spike, there is nothing to decide; after it, the party is reacting to an emergency instead of executing a plan.
Twelve separate d1 attacks (read: automatic one-point-per-segment damage) are pure attrition — the mean and the outcome converge.
This is not a design flaw. It is a design lever. Horror games reach for spikes precisely because they want helplessness and surprise. Grind-heavy dungeon crawls reach for attrition because resource depletion over many rooms is the whole point. The error is deploying one when you mean the other, or never examining which you are actually using.
The test to pull out
- 01"Best plausible run vs. worst plausible run" — a referee quick-check: imagine the best plausible run of rolls and the worst, and compare the two.
02What the dice are actually doing
Damage variance lives in the die size and the attack frequency together. A single d12 attack hits the table once per round and produces wild swings. Twelve separate d1 attacks (read: automatic one-point-per-segment damage) are pure attrition — the mean and the outcome converge. Every decision to use a big die or a small die, to attack once per round or four times, is a choice about where on the attrition-spike spectrum a creature sits. Most game designers know this instinctively; fewer state it explicitly in the monster's stat block or encounter notes.
The practical test for a referee is fast: imagine the worst plausible run of rolls and the best plausible run. If the spread between those two futures is narrow, the encounter is attrition. If one future ends a character and the other barely scratches the party, the encounter is a spike — regardless of what the averages promise. Design around the spread, not the mean, and the session plays the way it is supposed to feel.