A refractometer reads gravity by how a few drops of liquid bend light, and that works reliably on fresh, unfermented wort — but once fermentation starts, dissolved alcohol changes the liquid's refractive index in a way sugar alone doesn't, and a raw refractometer reading taken during or after fermentation overstates the true remaining gravity substantially if converted without correction.
A worked example
A batch with a known OG of 1.050 (12.4°Brix, using the same SG-to-Brix conversion this site treats as numerically equal to Plato), read mid-fermentation on a refractometer at 8.5°Brix, corrects to FG 1.0198 using the widely-published Sean Terrill correction formula — a cubic polynomial in original and apparent Brix, fit against real paired hydrometer/refractometer fermentation data. Read the raw 8.5°Brix as if it were a direct SG conversion instead, with no correction applied, and you'd land well above 1.030 — the correction exists specifically because that naive reading overstates remaining gravity once alcohol is present.
A second reading later in fermentation, at 6.5°Brix, corrects to FG 1.0129 — noticeably different from what a direct, uncorrected Brix-to-SG conversion of 6.5° would suggest, and a good illustration of why refractometer readings taken on fermenting beer need the correction applied every single time, not just once near the end.
Why the correction needs your original gravity too
Unlike hydrometer temperature correction, which only needs the current temperature, refractometer correction needs both the original gravity (as a Brix value) and the current apparent Brix reading — the formula is comparing how much the liquid's refractive index has shifted from its starting point, not evaluating the current reading in isolation. This means refractometer correction only works if you took an accurate OG reading (typically with a hydrometer, or a refractometer before fermentation started when no correction is needed yet) to correct against later.
When you don't need this at all
Before fermentation starts, a refractometer reads accurately with no correction needed — there's no alcohol present yet to distort the reading. The correction only becomes necessary once yeast is pitched and alcohol starts accumulating, and it matters more the further along fermentation has progressed, since more alcohol means a bigger gap between the raw reading and the true gravity.
The practical takeaway
If you're checking progress mid-fermentation with a refractometer, always apply the correction rather than reading the Brix value directly as if it were a hydrometer-equivalent SG — an uncorrected mid-fermentation refractometer reading can be off by a large enough margin to genuinely mislead you about whether fermentation is finished. Many brewers keep a hydrometer for a final, definitive FG check specifically because it needs no correction at all, using the refractometer only for quick progress checks along the way where an approximate, correctly-adjusted number is good enough.