Methodology · Measures, scenarios and assumptions

The numbers behind the number.

Our promise is plain English on the surface and institutional discipline underneath. This page is the underneath — the measures we run on your wealth, the scenarios we test it against, and the assumptions behind every figure. Written to be read: a clear version for anyone, and a precise version for those who want the footnotes.

The short version
  • We answer two questions with numbers: are you rewarded for the risk you take, and how far could your wealth fall.
  • Downside is measured three ways — a normal market, a stressed market, and a real-world shock — using the same tail measure banks apply to their own trading books: Expected Shortfall at 97.5%.
  • We use the previous day's closing prices and replay real history — we assume no tidy bell curve, and we never predict.
  • We are open about what the method cannot do. The honest limits are on this page, not hidden in a footnote.
We measure — we never recommend Real history — never a forecast Read-only — nothing to sell you
What we measure

Two questions, answered in numbers.

Every measure on this page serves one of two questions. The first is about reward. The second is about safety. Neither is an opinion — both are arithmetic on your own holdings.

Question one

Are your returns worth the risk?

It is easy to know what a portfolio returned. It is much harder to know whether that return was fair payment for the risk taken to earn it — and whether it beat a fair comparison. That is what performance measurement answers.

Question two

How far could it fall?

Averages describe calm years. Risk is about the bad ones. We measure the size of a serious loss — not the chance of a small dip — across a normal market, a stressed one, and a genuine shock.

Performance · Are you rewarded for the risk?

Return, weighed against the risk it cost.

A 12% return sounds good until you learn what was risked to get it. We weigh return against risk, and against a fair comparison — then translate the result into plain green, amber or red.

01 · REWARD FOR RISK

Risk-adjusted return

How much return you earned for each unit of risk you carried — the institutional way to compare two portfolios fairly, rather than by headline return alone.

Precise: the Sharpe ratio (return per unit of total volatility) and the Sortino ratio (per unit of downside volatility), annualised, measured against a risk-free rate.
02 · A FAIR COMPARISON

Benchmark & peer-relative

Whether you beat a relevant market, and investors in a position like yours — not a flattering index chosen after the fact.

Precise: excess return versus benchmark (cumulative and annualised), and beta — how much your wealth moves when the market moves.
03 · A PLAIN VERDICT

The performance read

The measures are combined into a single, honest read of whether your approach is working — so you don't need to interpret a page of ratios.

Precise: a weighted score across return, risk-adjusted return, benchmark-relative return, drawdown and market exposure, resolved to green / amber / red.
Risk · How far could it fall?

Three ways of asking “how bad could it get?”

The three rows you see on the platform each come from a distinct, established measure. They rise in severity — an ordinary bad stretch, your own worst historical year, and a deliberate shock that hasn't happened yet.

The three-scenario ladderIllustrative
One measure per row — and where it comes from
The figures are illustrative. On the platform, each row is your own — computed on everything you hold, property included, and checked against the loss you said you'd accept.
Row · what it measuresRiskYour limit
Normal marketExpected Shortfall 97.5% · ~2yr of daily history −6.2%−10%
Stressed marketStressed ES · your worst 12-month window −9.4%−10%
Real-world shockForward-looking stress test · scenario shocks −22.0%−10%
Each row is watched daily against your limit. Cross a line and you're alerted — monitoring against a pre-agreed number on measured data, never a prediction that a fall is coming.
NORMAL MARKET

Expected Shortfall, 97.5%

The average loss on your worst days — not the chance of a loss, but its typical size when a bad day arrives. We read it from roughly two years of real daily history, ranking every day and averaging the tail.

Why 97.5%: it is the level global bank supervisors adopted for trading books, chosen because it captures the depth of tail losses, not just a threshold.
STRESSED MARKET

Stressed Expected Shortfall

The same tail measure, but recalibrated to the worst continuous stretch in your own book's history — often a period like 2020 or 2008 — so the figure reflects a genuinely hard year rather than a calm one.

Kept honest: the stressed window is re-selected periodically, and if your stressed and normal figures drift too close together we re-check straight away.
REAL-WORLD SHOCK

Forward-looking stress tests

Deliberate, hypothetical shocks — a sharp market fall, a currency move — applied to exactly what you hold today. These capture tail events that history hasn't shown yet, the way regulators stress banks.

Precise: defined percentage shocks per scenario, applied to each holding and summed across your whole balance sheet; only active scenarios are shown.
Assumptions · Stated, not buried

The choices behind every figure.

Every risk number rests on assumptions. Firms that hide them ask you to trust the output blind. Here are ours, in the open.

Which prices

Yesterday's closing prices

We value everything on the previous business day's end-of-day prices and exchange rates. It is the institutional convention: it uses only information that was genuinely available, so no figure benefits from hindsight.

Which model

Real history, not a bell curve

We replay actual market history rather than assuming losses follow a neat statistical curve — because real markets have fatter, uglier tails than the tidy models suggest. This is called historical simulation.

Which maths

Standard institutional conventions

252 trading days a year; returns linked the way compounding actually works; a risk-free rate for risk-adjusted measures. The same conventions used on institutional books, so the figures are comparable to a professional standard.

Which currency

Your currency, one balance sheet

You choose a reporting currency; everything you hold abroad is converted into it before any measurement, so the whole picture is stated on one consistent footing — including property, not just investments.

Which window

A look-back you can set

Performance is measured over a horizon you choose — typically one, three or five years. The tail measures use roughly two years of daily data; the stressed measure uses about one. Every report states the exact window it used.

The honest limits

What this method cannot tell you.

No measure sees everything. Naming the edges is the difference between a tool and a sales pitch — and it is how a risk manager actually thinks.

Bonds

Modelled by a close match, not priced bond-by-bond

Individual bonds are risk-modelled using a closely matched slice of the bond market rather than a full instrument-level pricing model. That is enough to size the risk well; it is not a substitute for bond-by-bond valuation, and we flag where it applies.

Currency

A revaluation and a shock — not yet its own risk factor

Exchange-rate moves are captured when we revalue foreign holdings and inside the stress tests, but currency is not yet tracked as a standalone risk in its own right. That is a deliberate, disclosed choice for launch.

Some assets

Alternatives sit outside the risk model at launch

Private holdings, collectibles and similar assets are not risk-modelled at launch — they lack the reliable daily prices the measures need. Where they matter to your total wealth, we say so plainly rather than pretend a number.

Property

Valued periodically, not priced live

Property enters your balance sheet at your most reliable valuation and is refreshed periodically. It does not move day to day like a listed share — and treating it as if it did would be false precision.

The deepest limit

History constrains — it does not predict

Every figure here is a measured range from real data, not a forecast. The next shock may be worse than any in the record. We measure so you can see the shape of the risk; we never claim to know the future.

Under the hood · The precise version

Every measure, in one line.

For readers who want the exact definitions. If this is your idea of a good footnote, this section is for you.

Expected Shortfall (97.5%)The average loss across the worst 2.5% of days — the depth of the tail, not merely the point where it begins. Read from ~500 trading days by historical simulation.
Value-at-RiskFor reference: the loss a bad day rarely exceeds at a set confidence. We lead on Expected Shortfall because VaR is silent about how bad the exceedances are.
Stressed ESThe same 97.5% Expected Shortfall, recalibrated over the ~250-day window that was historically worst for your book, re-selected on a periodic cadence.
Stress testsFixed percentage shocks per scenario applied to current holdings — an equity leg and a currency leg per position — aggregated across the whole balance sheet. Active scenarios only.
Sharpe / SortinoExcess return per unit of total volatility (Sharpe) and per unit of downside volatility (Sortino), annualised on 252 days.
BetaSensitivity of your returns to the benchmark's — roughly, how much you move when the market moves one percent.
Maximum drawdownThe worst peak-to-trough fall along the wealth path over the window, shown for your portfolio and relative to the benchmark.
Excess returnYour return minus a like-for-like benchmark's, both built the same way — cumulative and annualised.
Windows & cadencePerformance over your chosen look-back (1/3/5y); ES over ~2y of daily data; stressed window ~1y; annualisation factor 252; limits monitored daily; stressed window recalibrated periodically.

These are the measures used on institutional trading desks, applied to a private balance sheet. If you would like to test them against your own holdings, that is exactly what the free assessment is for.

Fair questions

The questions a careful reader asks.

What confidence level do you use, and why 97.5% rather than 99%?

We lead on Expected Shortfall at 97.5% — the level global bank supervisors adopted for trading books under the post-crisis market-risk rules, replacing the older 99% Value-at-Risk. The reason is that Expected Shortfall measures the average size of losses in the tail, so 97.5% Expected Shortfall is designed to be at least as demanding as 99% Value-at-Risk while telling you more about how bad a bad day actually is.

Do you predict crashes or tell me when to get out?

No. We measure and we monitor; we do not forecast. The daily check compares your measured risk against limits you set and alerts you the moment you breach one — that is oversight against a pre-agreed number, never a claim to see a fall coming. And we never tell you to buy or sell anything.

How current is the data behind my numbers?

We value your wealth on the previous business day's closing prices and exchange rates. It is the institutional convention because it uses only information that was actually available at the time — no figure quietly borrows from hindsight. Property is carried at your most reliable valuation and refreshed periodically.

Why model bonds with a proxy instead of pricing each one?

Reliable daily prices for individual bonds are patchy, so we risk-model each bond using a closely matched slice of the bond market that does have clean daily data. It sizes the risk well and keeps the whole picture consistent. It is not full bond-by-bond pricing, and we mark where it applies rather than hide it — see the honest limits above.

Does this include my property?

Yes. Most reviews quietly ignore the largest thing people own. We assess your whole balance sheet — investments, cash and property — as one picture, because that is the only honest way to measure your real downside. Property is valued periodically rather than priced live, which we state clearly.

Is any of this financial advice?

No. This is independent risk analysis and information: what your figures are, how they were measured, and what they mean. We never recommend specific investments, we never handle your money, and we are not authorised or regulated as an investment adviser. What you decide with the numbers is entirely yours.

Test it on your own wealth

The method is on the table. Bring your holdings.

A free 20-minute risk assessment with a senior risk specialist runs these measures on your own wealth — property included — and gives you the one number most investors never see: how far it could fall, against the loss you'd actually accept. No report, no pitch; just where you stand.

Read-only · One flat fee · Nothing to sell you
Book your free risk assessment →

Private Hedge provides independent risk analytics, education and information. Nothing on this page is financial advice, a personal recommendation, or an invitation to buy or sell any investment. Private Hedge Limited holds no custody of client assets, takes no commissions, and sells no products; any account connection is read-only. We are not authorised or regulated as an investment adviser, portfolio manager, or broker in the UK, EU, or US. The value of investments can fall as well as rise and your capital is at risk. Figures and screens shown are illustrative unless a source is given; past performance is not a reliable indicator of future results.

Sources & further reading
  1. Basel Committee on Banking Supervision, Minimum capital requirements for market risk (the Fundamental Review of the Trading Book), Bank for International Settlements — the standard that adopts a 97.5% Expected Shortfall for trading-book market risk. bis.org/bcbs/publ/d457.htm
  2. Bank for International Settlements, Explanatory note on the minimum capital requirements for market risk. bis.org/bcbs/publ/d457_note.pdf