See the factor bets hiding inside your portfolio.

Your returns, decomposed like a recipe

Instead of “up 14%,” Factor X-Ray tells you how much of your day-to-day behavior moved with the market itself and how much with tilts toward small or large companies, value or growth, profitable or unprofitable businesses, conservative or aggressive investors, and recent winners — the statistical recipe behind your results.

Real regression on published research factors

Exposures come from an ordinary least-squares regression against the daily factor returns Kenneth French publishes in his data library at Dartmouth — the same series used in academic work — not a rough proxy built from sector labels. You get a loading (beta) for each factor, the R² of the fit, and the annualized intercept: the part of your return the six factors leave unexplained.

Concentration you couldn’t see on a holdings list

“Diversified across 20 names” and “20 names loading on the same three factors” look identical in a portfolio table and completely different in reality. Factor X-Ray surfaces the difference so hidden concentration becomes a number.

How is the factor regression computed?

Five steps, all on data you can check. The same flow-adjusted daily return feeds the equity curve and the returns methodology, so this page and the performance pages read one series.

1. 1

   Your daily return, with deposits taken out

   Each day’s return is (value − external cash flow) ÷ previous value − 1. A $5,000 deposit into a $50,000 account is not a +10% day. If you record deposits and withdrawals, the value is your net worth; if you only record trades, it is position value measured against your buys and sells, with recorded dividends counted as income.

2. 2

   Six published factors, daily

   Market minus the risk-free rate, size (SMB), value (HML), profitability (RMW), investment (CMA) and momentum, from the Ken French Data Library, refreshed weekly. The library publishes with a lag of a month or two, so the regression window ends on its last available date.

3. 3

   Ordinary least squares, over all your history

   Your daily return minus that day’s risk-free rate is regressed on the six factors plus an intercept, using every day that has both a portfolio snapshot and a complete factor row. At least 60 overlapping days are required; below that the page tells you how many you have instead of printing loadings.

4. 4

   Loadings, fit and alpha

   You see six loadings, the R² (how much of your daily variation the factors explain), the number of observations and the window dates. Alpha is the daily intercept × 252. It is withheld if the account value ever went below zero — for example buys recorded without the deposits that paid for them — because an intercept fit across that is bookkeeping, not performance.

5. 5

   A plain-English read, by fixed rule

   A size, value, profitability, investment or momentum loading above +0.2 or below −0.2 counts as a tilt. A market loading above 1.2 reads as “more volatile than the market”, below 0.8 as “defensive”. The wording is deterministic — no AI writes it.

What does a factor X-Ray result look like?

Here is how a hypothetical result would be read, using the same thresholds the app applies. Read about why these tilts appear in are you secretly running a momentum fund?

This hypothetical book looks like a large-cap growth-and-momentum portfolio — a common shape for someone who has let winners run. If its R² were high and its alpha near zero, most of its day-to-day result would be explained by those three tilts rather than by anything specific to the stocks chosen.

Loadings are estimates with noise around them, and a short window makes them noisier. Treat them as indicative tilts, not precise exposures.