Most Reading Speed Tests Are Useless. Here Is Why.

Type “reading speed test” into a search engine and you will get a page with a block of text, a start button and a stop button. You read, you click, it tells you 412 words per minute, and you feel briefly pleased with yourself.

That number is worthless, and it is worthless for a reason worth understanding: nothing in that test checked whether you read the passage or merely looked at it.

You can score 900 on a test like that by dragging your eyes down the page without processing a word. The test cannot tell the difference. It is not measuring reading. It is measuring how fast you can move your gaze from the top of a box to the bottom.

Reading speed is a fraction, not a number

The moment you drop comprehension from the measurement, speed stops meaning anything, because speed and comprehension trade against each other. Push one up far enough and the other falls. A figure that ignores half of that trade is not a measurement, it is a stopwatch reading.

The honest unit is effective reading rate: your raw words per minute multiplied by the proportion of the material you actually retained.

This is why speed reading courses that advertise 1,000 words a minute are careful never to publish comprehension figures alongside. When researchers do measure both, the fast numbers collapse. Rates far above the normal range turn out to be skimming with a different name — and skimming is a legitimate, useful skill, but it is not reading, and it should not be scored as though it were.

What a test has to do to be worth taking

Three things, none of them optional.

It has to ask you questions afterwards. Not one question. Enough that guessing does not save you, and questions about the substance rather than a name you could have caught in peripheral vision.

It has to use text you have not seen. Familiar material inflates every score, which is why testing yourself on a passage you just read to warm up tells you nothing.

It has to tell you what it measured. A single number with no method behind it is a horoscope. You should be able to see the raw rate, the comprehension score and the product of the two.

Ours does all three. The reading speed test here measures you twice — once reading a passage the ordinary way, once with the words brought to you one at a time — and multiplies both results by the same comprehension score. Skimming earns nothing, because the questions come afterwards and there is no way back to the text.

The two-measurement part matters. One number tells you how fast you read. Two numbers tell you how much of your speed is being spent on the page rather than on the words, which is the only part you can actually do something about.

So what is a good number?

Lower than the internet has led you to believe.

For adults reading English silently, the measured average is around 238 words per minute for non-fiction and 260 for fiction. Reading aloud averages 183, because your speech rate becomes the ceiling. Proofreading sits near 200, since you are inspecting rather than comprehending. The typical range for competent adult readers is roughly 175 to 300.

If your test came back at 250, you are not slow. You are average, which is what average means. The full breakdown of where these figures come from is in the average reading speed.

What a number is genuinely good for is turning an abstract pile of reading into an amount of time. Six hundred pages stops being a vague dread and becomes four hours, or seven, and you can decide accordingly — that is what the reading time calculator is for.

What actually moves the number

Most advice on this points at the wrong thing.

Stopping your inner voice does not work. The advice to silence subvocalisation is everywhere, and the evidence does not support it. Suppressing your inner voice tends to damage comprehension rather than raise speed, because that inner voice is doing real work on anything harder than a menu. We went through it properly in the subvocalisation myth.

Reading in “chunks” mostly does not work either. Your useful visual span is far narrower than the courses claim. You cannot train your eye to take in three lines at once, because the acuity to resolve the words is not there outside a small window around the fixation point.

Removing wasted movement does work, because that time is genuinely being spent and genuinely recoverable. Two things eat it:

• Regressions. Between 10 and 20 per cent of an average adult’s eye movements go backwards, mostly without any decision to re-read. Regression in reading covers what triggers them.
• The return sweep. Every line ends with a long jump back to the start of the next one, and it undershoots often enough that a single line transition is usually three movements rather than one. The return sweep has the detail.

Neither of those is comprehension. It is navigation — the overhead of finding the next word on a page. And it is the one part of reading that can be removed outright rather than merely practised.

Removing the navigation

That is what RSVP does. The words come to you, one at a time, in one fixed position. There is no line to sweep back to, no place for your eyes to drift backwards to, and nothing to re-find after a pause.

What is left is the reading itself, and the speed you can hold turns out to be higher than the same person manages on a page — not because comprehension has been sacrificed, but because the eye movement overhead is gone.

Take the test and you will get both numbers: what you do on a page, and what you do without the page. The gap between them is the part that was never really reading.

Find out what your real number is

Two minutes, two measurements, and a comprehension score that neither of them can dodge. Nothing is uploaded and there is no account.

Take the reading speed test →

Then work out what it means for the book you are actually holding: how long will this take?

Also worth reading: The Average Reading Speed · Regression in Reading · The Return Sweep