thinking-margin-of-safety

1# Margin of Safety
2 
3Core rule: commit a buffered number when residual uncertainty plus breach cost can cause ruin or costly failure. Margin covers estimation error; it is not free slack for laziness.
4 
5## When to Use
6 
7- Capacity, timeouts, pool sizes, queue depths, storage, or SLA numbers under load uncertainty
8- Timeline or budget commitments where underestimation is costly
9- Architecture headroom when scaling or recovery is slow
10- Any public or production commitment where being wrong has asymmetric downside
11 
12## When NOT to Use
13 
14- Breach is detected immediately and fixed near-zero cost (fast auto-scale, live-tunable limit)—static fat buffers waste resources
15- `cost(buffer) > P(breach) × cost(breach)`—right-size or measure instead of maxing out
16- Uncertainty is eliminable by measurement or lookup—get the real number first
17- Stopping-criterion / search budget problems—use bounded rationality, not padding
18- Fully known parameters and low stakes where edge optimization is the goal
19 
20## Procedure
21 
221. **Point estimate without padding.** State the base requirement in explicit units (RPS, weeks, GB, ms) and confidence (high / medium / low).
232. **Uncertainty and breach cost.** List drivers (spike, growth, unknowns, dependency variance). State what fails if undershot: outage, missed launch, data loss, reputation—and whether failure is recoverable or ruinous.
24…