Risk: CVaR, IVaR, stress

What your portfolio loses in bad scenarios — and which holdings are responsible.

The Risk Analysis sub-tab of Portfolio is where you find out what your book actually loses when things go sideways. Three lenses, each answering a slightly different question.

CVaR — Conditional Value at Risk

The question: “On a really bad day, how much do I lose?”

CVaR is the average loss in the worst X% of historical scenarios. We typically show the 95% and 99% CVaR.

Reading	Meaning
95% CVaR = -3.2%	In the worst 5% of historical days, your portfolio averaged a -3.2% loss
99% CVaR = -6.8%	In the worst 1% of historical days, your portfolio averaged a -6.8% loss

CVaR is more honest than VaR (Value at Risk) because it doesn’t ignore the tail — it averages it.

How to use it: if your 99% CVaR is bigger than what you can stomach losing on a single day, your position sizes are too big.

IVaR — Incremental Value at Risk

The question: “Which holdings are responsible for my risk?”

IVaR breaks down your portfolio’s risk by holding. Each name gets a contribution score — how much of your CVaR is that position adding?

Reading	Meaning
NVDA: +18% IVaR	NVDA is contributing 18% of your portfolio’s downside risk
BRK.B: +2% IVaR	BRK.B is contributing 2% — disproportionately small risk for its weight

A position with a small weight but a large IVaR contribution is your risk problem — even if you don’t think of it that way.

How to use it: sort holdings by IVaR. If one name is 30%+ of your risk, ask whether that’s intentional.

Stress scenarios

The question: “What does my portfolio do in a specific historical event?”

Stress scenarios replay your current holdings through past crisis windows:

• 2008 GFC — Sep 2008–Mar 2009
• 2020 COVID crash — Feb–Mar 2020
• 2022 rate shock — full year 2022
• 2018 vol spike (Volmageddon) — early Feb 2018
• Custom date ranges — pick any window

Each scenario shows: total drawdown, time-to-recover, worst day, and per-holding contribution.

How to use it: these are “what would happen if X repeated tomorrow.” Useful for calibrating sizes before a known catalyst (FOMC, CPI, election).

Correlation matrix

Lives on the Performance tab but matters here too — if your IVaR shows two names contributing risk and the correlation matrix shows them at +0.92, you don’t have two positions, you have one.

How most people use it

• After every new position — check IVaR; did you just concentrate risk?
• Before known catalysts — run the stress scenarios most analogous to what you’re worried about
• Monthly — re-check 99% CVaR against your “I can stomach losing this in a day” number
• When markets feel calm — that’s when CVaR is most useful; the calm reading is your early warning that a bad day will hurt more than your gut thinks