Student's t distribution
A bell-shaped curve with fatter tails than the normal. With few degrees of freedom big moves are far more likely; as ν grows it turns into the normal.
01The formula
- the density at t
- degrees of freedom: small means fat tails, large means close to normal
- the gamma function, which extends the factorial to non-whole numbers
- the value, measured in scale units like a z-score
02See it move
03Where it is used
- Risk managerFits a t with 3 to 6 degrees of freedom to daily returns so value at risk respects fat tails.
- Quant researcherUses it for t-tests on the mean return of a strategy with a short history.
- Options deskUses fat-tailed return models to explain why out-of-the-money puts trade above their Black–Scholes price.
04Why traders use it
Market returns have more extreme days than the normal predicts. The t distribution adds that with one extra number, ν, and also gives the right test when volatility is estimated from the data.
05Worked example
With 4 degrees of freedom, a move beyond 3.0 scale units either way has chance 3.994%. Under the normal it is only 0.270%: the t makes it 14.8× as likely.
06How it is built
- Take a standard normal and an independent chi-squared variable with degrees of freedom.
- Form : a normal divided by a noisy estimate of its own scale.
- When happens to be small, is large, which fattens the tails.
- Working out its density gives the formula above; as , and becomes normal.
07The common mistake
Matching the scale but not the variance. A t with ν degrees of freedom has variance ν/(ν − 2), above 1, so to fit a return series you must shrink the scale or you will overstate volatility.
08In Python and LaTeX
from scipy.stats import t, norm nu, k = 4, 3 print(round(2*t.sf(k, nu), 5), round(2*norm.sf(k), 5))
f(t) = \frac{\Gamma\!\left(\frac{\nu + 1}{2}\right)}{\sqrt{\nu\pi}\,\Gamma\!\left(\frac{\nu}{2}\right)} \left(1 + \frac{t^2}{\nu}\right)^{-\frac{\nu + 1}{2}}09Where it fits
Rather not code the maths yourself? Describe your trading idea and let AI build the bot.
Build a bot with AgenKitThe worked example and the model run the same code, so every number matches. Models simplify markets on purpose. This is education, not financial advice.