One floats in your beer; the other bends light through a drop of it. Both end up telling you a gravity, and the homebrew internet has settled into a tidy summary: the refractometer is convenient, the hydrometer is accurate, buy the hydrometer first. The convenience half is true and the accuracy half is folklore, and the reason nobody has noticed is that every page making the comparison leaves out the number that actually decides it.
What is the actual difference between them?
A hydrometer measures density directly. It sinks until it has displaced its own weight, so a denser liquid floats it higher, and density is what gravity means. Nothing is inferred.
A refractometer measures how much the liquid bends light, and converts that to a sugar concentration on the assumption that the dissolved solid is sucrose. Wort is not sucrose — it is maltose, maltotriose and dextrins — so the conversion is wrong by a few percent, in a way that is consistent for your instrument and correctable. That correction is the whole story of this page.
Which one is actually more accurate?
Read the spec sheets rather than the forums. A homebrew refractometer is sold with a stated tolerance: ±0.2 °Bx, with 0.2 °Bx scale divisions. At a normal 12.5 °Bx wort that band is ±0.8 gravity points. A glass hydrometer is ruled every 0.002 — two points between lines — and its makers publish that graduation rather than a tolerance, which is itself telling.
| Hydrometer | Refractometer | |
|---|---|---|
| Sample needed | ~150 mL, into a trial jar | Two or three drops |
| Published precision | Ruled every 0.002 (2 points) | ±0.2 °Bx (±0.8 points) |
| Reads a fermenting beer | Yes, directly | Only with a correction and the original reading |
| What you must calibrate | Offset, in distilled water | Zero, then the wort correction factor |
| Time per reading | Minutes — draw, cool, settle | Seconds |
| Survives being dropped | No | Usually |
| Typical price | About $10–20 | About $25–40 |
Precision figures are the manufacturers' published scale divisions and tolerance, converted to gravity points at 12.5 °Bx by this site's engine. Neither is a measured accuracy comparison — nobody publishes one. Prices are the street range at the time of writing; Homebrew Academy's comparison quotes $21.99 for a hydrometer and "under $40" for a refractometer, which is the same picture.
So on paper the refractometer is the finer instrument, and the folk ranking is backwards. What the hydrometer has is not precision but range: it keeps working once there is alcohol in the beer, and the refractometer does not. That is a categorical difference, and it is the one that should decide your purchase.
How do you actually use a refractometer?
Open the flap, put one or two drops on the glass prism, close the flap so the liquid spreads without bubbles, and look through the eyepiece at a light source. You read the Brix scale where the light and dark fields meet. Wipe the prism dry afterward — dried wort on the glass is the most common reason a refractometer starts reading strangely.
Two things the box will not tell you. Hot wort fogs the prism and reads badly, so let a drop sit ten or twenty seconds before reading it. And the "ATC" on the label does not do what most buyers assume — it stops a warm sample reading differently from a cool one, within a stated range. It is not a wort correction, and it is not an alcohol correction. Those are two separate problems, and they are the next two sections.
The 4% every comparison leaves out
Because wort is not sucrose, a refractometer reads it about 4% high. The fix is the wort correction factor: divide the Brix reading by it before converting. Most instruments land between 1.02 and 1.06, it is a property of your particular refractometer, and you find it once by reading the same cooled wort with both instruments.
Not one of the four ranking comparisons mentions it. Brew Your Own does not discuss it. DraftMag does not. Yeast Platform does not. Homebrew Academy tells you to "multiply by 4". So here is what the setting nobody mentions is worth, on the same reading:
| Reading | WCF 1.02 | WCF 1.04 | WCF 1.06 | Spread |
|---|---|---|---|---|
| 8 °Bx | 1.0312 | 1.0305 | 1.0300 | 1.2 pts |
| 12.5 °Bx | 1.0495 | 1.0485 | 1.0475 | 1.9 pts |
| 16 °Bx | 1.0641 | 1.0628 | 1.0615 | 2.5 pts |
| 20 °Bx | 1.0812 | 1.0796 | 1.0780 | 3.3 pts |
The same refractometer reading converted at each end of the usual WCF range, from this site's engine. Spread is the gap between the 1.02 and 1.06 columns in gravity points.
At ordinary strength it is about 2 points of original gravity — one hydrometer graduation, and always in the same direction, which is what makes it worth fixing. It widens with the beer: 3 points on a 20 °Bx barleywine. Two points of a fifty-point wort is four percent of your measured brewhouse efficiency, which is where a systematic error like this does its real damage — you spend a season chasing a crush problem you do not have.
Homebrew Academy's "multiply by 4 — 15 Brix is 1.060" is wrong twice, and the errors point in opposite directions. Converting 15 °Bx exactly but without the wort correction gives 1.0611, so the ÷4 rule reads about a point low. Correct for the wort properly and the answer is 1.0587, so skipping the WCF reads high. Net, the rule lands 1.3 points above the truth instead of 2.5. That accidental cancellation is why the rule of thumb has survived — and why it stops working on strong worts, where the ÷4 rule drifts and the WCF gap grows at the same time.
Run that 12.5 °Bx wort all the way to a finished beer at 1.010 and the entire WCF argument is 0.26 percentage points of ABV — 4.92% against 5.18%. If all you want is the number for the label, an uncalibrated refractometer is not lying to you in any way you would taste. It is your efficiency and your recipe scaling that pay for it.
When does a refractometer stop working?
The moment fermentation starts. Ethanol bends light too, and the instrument cannot tell which part of the reading is sugar and which is alcohol — so a fermenting beer reads far too high. It is not a calibration problem and no correction factor fixes it. What does work is knowing where the wort started: with the original Brix and the current one, a fitted formula backs out the alcohol. The refractometer correction calculator runs both published fits and shows where they disagree.
This is where the two most-read comparisons flatly contradict each other. Brew Your Own reports that brewers "stop using a refractometer after brew day and rely on a hydrometer for FG readings". Homebrew Academy recommends the reverse — hydrometer on brew day, then "use a refractometer during fermentation to avoid large sample waste". BYO has it right. The refractometer's easy, uncorrected accuracy exists only before yeast is pitched; recommending it for the fermentation window is recommending it for the one job it cannot do unaided.
So which should you buy first?
The hydrometer, and it is not close — but not for the reason the comparisons give. Buy it because it is the only one of the two that can answer the question you cannot avoid: has my beer finished? That needs a final gravity, days apart, on a beer full of alcohol. A refractometer cannot do it without a second instrument's reading to lean on, which means a refractometer alone is not a complete kit and a hydrometer alone is.
Add a refractometer when brew day itself starts to annoy you. It reads in seconds from two drops, which means you can check the mash, check pre-boil while there is still time to fix it, and confirm the OG — the readings you skip with a hydrometer because each one costs a cooled sample and ten minutes. Our hydrometer guide works out what those samples cost: four readings from a 150 mL trial jar is 600 mL, about two bottles of finished beer. A refractometer's four readings cost a few drops.
The pairing every experienced brewer converges on is not a compromise, it is each instrument doing what it is good at: refractometer up to pitching, hydrometer from pitching onward. And whichever you buy, the first thing to do with it is calibrate it — distilled water for the hydrometer's offset, a side-by-side wort reading for the refractometer's correction factor. Both take one brew day. Neither is mentioned by the pages telling you which to buy.