| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | CURX | SPY | Δ |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2026 · ytd | −2.1% | +39.1% | +13.5% | −18.4% | −20.5% | −20.0% | −25.3% | — | — | — | — | — | −40.1% | +7.5% | −47.6pp |
| 2025 · part | — | — | — | — | — | — | — | — | −87.1% | −5.9% | −28.1% | −32.2% | −94.1% | +17.7% | −111.8pp |
A month’s return is its month-end close over the prior month-end’s, so a year’s figure is exactly the product of its twelve monthly cells — no separate annualisation, no resampling. A blank cell means no return exists for that month (the series had not started, or it is the very first month-end, which has nothing to be measured against). Blank is never zero.
Basis. CURX is on the total return (splits + dividends) basis, SPY on the total return (splits + dividends) basis. Both columns are on the same basis, so the Δ column is a like-for-like difference.
Partial years. A year marked “ytd” or “part” does not cover the full calendar year: either the stored history starts inside it (the figure is measured from the first month-end present) or the data stops before December (the figure runs through the last stored close). The row’s own tooltip names the exact boundary. It is never dressed up as a calendar-year return.
End of day. These are stored end-of-day closes, and the window is anchored on the LAST stored bar rather than today’s date — so the newest row ends where the data ends. Nothing here is intraday.
