Extreme value theory — the tail is a different distribution
Extreme value theory models the maxima and tails directly instead of assuming a bell curve, which is why VaR built on a normal distribution understates exactly the losses that end firms.
Not yet scoped.
Why
The PoC fits a generalized Pareto tail to loss data and compares its 99.9% quantile against a Gaussian VaR to size the gap the normal assumption creates.
How it works
Not yet built.
Where it lands in Jayverse
Verex / DeFi: fit a generalized Pareto tail to settlement-loss or liquidation data instead of assuming Gaussian VaR. The normal assumption understates exactly the losses that would drain a market maker's or liquid-staking pool's reserve.
Auditor: publish the 99.9% quantile gap between the fitted tail and a naive Gaussian VaR. A standing risk-methodology number, the same "write the methodology down" instinct as the Kaiko-style rule.
Key expressions
Words and phrases from this page worth keeping, with the Korean meaning and the sentence they come from.
Expression
뜻 · 쓰이는 자리
bell curve
종형 곡선, 정규분포 모양 · 정규분포를 시각적으로 가리킬 때. "instead of assuming a bell curve"
understate
실제보다 적게 말하다, 과소평가하다 · 손실 규모를 실제보다 작게 나타낼 때. "understates exactly the losses that end firms"
end (a firm)
회사를 망하게 하다, 끝장내다 · 손실이 기업을 파산시킬 정도일 때. "the losses that end firms"
fit X to Y
X를 Y에 맞춰 적합시키다 · 통계 모델을 데이터에 적합시킬 때. "fits a generalized Pareto tail to loss data"
size (the gap)
차이를 가늠하다, 측정하다 · 두 값의 차이를 수치로 파악할 때. "to size the gap the normal assumption creates"
VaR
손실 위험액(Value at Risk) · 정규분포 가정 시 극단적 손실을 과소평가하게 되는 위험 지표. "why VaR built on a normal distribution understates exactly"
generalized Pareto (distribution)
일반화 파레토 분포(Generalized Pareto Distribution) · 극단값 이론에서 꼬리 분포를 직접 모델링할 때 쓰는 분포. "The PoC fits a generalized Pareto tail to loss"