Research

Risk conventions

Every risk figure on the app follows a specific convention — which unit, which scale, which simplifying assumption. This page lays out each one explicitly, so nothing is left to guess at.

On this page

Delta

Share-equivalent, not raw per-contract

Black-Scholes' own delta is a number between -1 and 1 per contract — how much the option's price moves per $1 move in the underlying. RiskAlgo instead reports position delta: that raw delta multiplied by the contract multiplier (100) and by however many contracts you hold. A Net Delta of +500 means the position currently behaves, for a small move, like owning 500 shares of stock — the same convention virtually every broker platform uses, chosen specifically because it's the one most traders already have intuition for.

Vega

Per 1 percentage point of volatility, not per 100

The raw mathematical Vega formula gives dollar price change per 1.00 (100 percentage points) of volatility — moving from, say, 20% to 120%, which never happens in practice. RiskAlgo scales this down by 100, matching standard market convention:

\[ \nu_{\text{displayed}} = \frac{S_0 \, \varphi(d_1) \sqrt{T}}{100} \]

So a Vega of $0.15 means the option's price moves about $0.15 if implied volatility shifts by one point — 20% to 21%, not 20% to 120%.

Theta

Per calendar day, not annualized

The standard Theta formula (see the math behind RiskAlgo) gives an annualized rate. RiskAlgo divides that by 365 before displaying it, so the number shown is "value lost per calendar day," matching what you'd actually see the position's mark-to-market drop by, day over day, all else held equal.

Delta notional and gamma in dollars

Dollars, so different stocks can be compared and added

Delta notional is position delta (in shares) times the stock price. It is the dollars of stock the position behaves like, and the amount you would trade to hedge it. Long is positive and short is negative.

\[ \text{Delta notional} = \Delta_{\text{shares}} \cdot S \]

Gamma in dollars per 1% move says how much that delta notional changes when the stock moves 1%. It uses position gamma (per $1, in shares), the price squared, and 1%:

\[ \Gamma_{\$} = \Gamma_{\text{shares}} \cdot S^{2} \cdot 0.01 \]

A positive figure means you get longer as the stock rises and shorter as it falls, so you would sell into rallies to stay hedged; a negative figure is the reverse. Raw gamma depends on the stock price, so it cannot be compared between a $20 stock and a $500 stock; gamma in dollars can, and it adds up across tickers.

Cross-ticker aggregation

Net totals blend tickers; the per-ticker breakdown doesn't

The "Aggregate Exposure" section's Net Delta, Vega, and Theta are summed across every open position, on every ticker. Netting is exactly correct within the same ticker — if you're +50 delta long and -30 delta short, both on AAPL, +20 net genuinely reflects offsetting risk on that name.

The blended total is less meaningful across different tickers. +50 delta in a $700 stock and +50 delta in a $20 stock read as identical in the summed total, despite representing very different dollar exposure per 1% move. The "Greeks by Ticker" section breaks the same figures out per underlying specifically to make this visible, rather than only showing the blended number.

Beta

Daily returns vs. SPY and QQQ, about one year

The Watchlist shows each ticker's beta to the S&P 500 (SPY) and to the Nasdaq-100 (QQQ). Beta is the covariance of the stock's daily log returns with the index's, divided by the index's variance, over the most recent 250 trading days. Closes are matched by calendar date, so a day missing from either series is skipped rather than misaligning the two, and dividend-adjusted closes are used where available. A beta of 1.5 means the stock has tended to move 1.5% for each 1% move in the index.

Beta is a statistical estimate from the past. It changes over time, and the screen shows the correlation next to it: a low correlation means the beta explains little of the stock's movement. With fewer than 60 matched days the app shows no beta rather than a poor estimate, and anywhere a beta is needed but unavailable it uses 1.00 and says so.

Stress test

Choose how the market move reaches each position

The Stress Test report applies a spot move (and an optional volatility shift in vol points) to your positions. You choose how the move is applied:

• Market: S&P 500 or Market: Nasdaq (the default is S&P 500). The slider is a move in SPY or QQQ, and each position is shocked by that move times its own beta to the index. A stock with beta 1.5 falls 15% when the index falls 10%.
• All tickers parallel move. Every underlying moves by the same percentage, whatever its beta.
• A single ticker. Only that ticker's positions are shocked, by the full percentage.

In every mode each option is repriced with the shocked spot and any vol shift, and the same point shift applies to every leg regardless of moneyness or expiry. Days to expiry are held at today's value, so time decay is not modeled.

The beta modes are a one-factor model: they capture market-driven moves only, with no stock-specific moves, and betas estimated in calm periods tend to understate how closely stocks fall together in a crash. The report lists the beta used for each ticker. The stress test is a scenario, not a forecast.

VaR methodology

5-day horizon, realized (not implied) volatility

The VaR Report measures each ticker's own risk independently, using that ticker's realized (historical) volatility rather than an option's implied volatility — VaR is conventionally a realized-vol measure. The horizon is fixed at 5 trading days (a standard short-term risk window), regardless of the position's actual time to expiry. Both the Pricer's own 95%/99% VaR and this report use the same underlying lognormal-distribution approach, just over different horizons and different volatility inputs — the Pricer uses time-to-expiry and implied vol (what you're pricing with), the report uses a fixed 5-day window and realized vol (what actually happened historically).

On top of the per-ticker figures, the report shows the whole portfolio's 5-day, 99% VaR under three models, using the same position values and option repricing:

• S&P 500 market model and Nasdaq market model. The index (SPY or QQQ) moves by its own realized volatility and every position moves by its beta to that index. These capture market-driven risk only; stock-specific moves are not included, so they are usually smaller than the sum of the ticker figures.
• All tickers move together. Every ticker moves by the same number of its own standard deviations (perfect correlation). This is the least diversified case and usually the largest figure.

Each ticker row also shows its notional delta, the share-equivalent delta times the live price. A loss beyond a VaR figure remains possible; VaR does not say how large it would be, and all of these are estimates from past data.

Assignment risk heuristic

In-the-money + within 7 days — a flag, not a prediction

The Assignment Risk report flags any short option leg that is currently in-the-money and within 7 calendar days of expiration. This is a simplified heuristic based on when early assignment becomes more likely — as remaining time value shrinks, or ahead of an ex-dividend date for calls — not an actual probability model or a guarantee. Assignment can happen outside these conditions, and being flagged doesn't mean assignment is certain. Confirm your real exposure with your broker, especially close to expiration.

Risk limits

Advisory caps on the size of an exposure

A limit caps the absolute size of one exposure, so -$30,000 of delta notional counts the same as +$30,000. The exposures are delta notional, gamma in dollars per 1% move, vega (per vol point), theta (per day), estimated margin, and scenario loss. Scenario loss is the worse of the loss if every stock in scope falls by the chosen percentage or rises by it, with no change in volatility and no betas.

A limit can apply to the whole portfolio, to every ticker, to every sector, to one ticker or to one sector. A limit on one ticker or sector replaces the general limit for that name when both measure the same exposure. Usage is shown as green below 80%, orange from 80% to 100%, and red above 100%. Sectors come from Yahoo Finance and can be corrected in the editor; a ticker with no sector is left out of sector limits. Limits are for your own discipline: the app never blocks or changes a trade.

Gamma exposure (GEX)

An estimate from open interest, not a measurement

GEX estimates how many dollars of stock option dealers must trade to stay hedged when the stock moves 1%. For each strike and expiry within 60 days:

\[ \text{GEX} = \pm\, \Gamma \cdot \text{OI} \cdot 100 \cdot S^{2} \cdot 0.01 \]

\(\Gamma\) is the Black-Scholes gamma using the option’s implied volatility from the chain, OI is its open interest, and the sign is positive for calls and negative for puts. That convention assumes dealers are long the calls and short the puts. Nobody can see who is really on which side, so the sign is an assumption.

Net GEX above zero suggests dealers buy dips and sell rallies, which tends to calm the stock down. Net GEX below zero suggests they sell drops and buy rallies, which tends to make moves bigger. The gamma flip is the price where net GEX, recomputed at other stock prices with each implied volatility held fixed, crosses zero. The call wall and put wall are the strikes with the largest call and put GEX.

Open interest is the previous close’s figure from Yahoo Finance, which is unofficial and sometimes missing for thinly traded strikes. Options are priced as European with no dividend yield, and options expiring today are left out. Treat GEX as context, not a forecast.

This page documents the specific conventions and simplifications behind RiskAlgo's own figures — it isn't financial advice, and every methodology described here is a model, not a guarantee of real-world behavior. See RiskAlgo's Terms of Use for the full disclaimer.