Your personal baseline is the reading speed you can expect from yourself on a normal day with a given kind of text. To find it, take three to five timed tests on different passages across more than one day, keep only the attempts where you understood the passage, and use the median. A single test result is a sample, not a baseline.

Why one result is not enough

Run the same reader through five passages and you will get five different numbers. Several things move the result, and none of them is a change in the reader:

  • The passage. Short, common words go by faster than long, rare ones. Brysbaert (2019) reports a higher adult average for fiction (260 WPM) than for non-fiction (238 WPM) and attributes the gap largely to word length.
  • What you already know. A topic you are familiar with needs less working out.
  • Your state. Tired, distracted or rushed readers read differently from rested ones.
  • The timing. A second lost before you start or after you finish is a visible share of a 90-second reading.
  • The pace you chose. Being timed tempts people to hurry, and Rayner et al. (2016) conclude that speed gained that way is paid for in comprehension.

A baseline is what is left when these influences are averaged out. You cannot remove them from one attempt, but you can stop any single one from deciding the answer.

The protocol

This is a practical procedure, not a standardised assessment. It is designed to be quick and to resist flattering results.

  1. Take three to five attempts on the reading speed test. Five is better than three.
  2. Use a different passage each time. The test has five, one per category: general non-fiction, everyday science, history, narrative and technology. Never repeat a passage, because the second reading is faster for reasons unrelated to your speed.
  3. Spread the attempts over at least two days, ideally at different times of day.
  4. Read at the pace you would use if nobody were timing you. Click Start Reading when you are ready and Finished Reading at the last word.
  5. For this protocol, keep attempts with at least 3 of the 4 comprehension questions correct. This is a practical cut-off, not a validated assessment threshold. A lower score may reflect rushing, difficulty, distraction or a misunderstood question; do not use it alone to decide how much of the passage you understood.
  6. Take the median of the attempts that count. Round it to the nearest 5 or 10 WPM.

The test can keep your last five results in your browser’s local storage, which is enough to hold one full round. If you want a longer record, write the numbers down.

If fewer than three attempts qualify, do not settle for the ones you have. Wait a day and add attempts by the manual method described in how to measure your reading speed accurately, using fresh text.

Why the median and not the average

The median is the middle value once the results are sorted. With an odd number of results it is the one in the centre; with an even number it is the mean of the two central values.

It is the better summary here because it ignores how extreme an unusual attempt was. Take five qualifying results of 210, 220, 225, 240 and 300 WPM, where the 300 came from a passage on a subject the reader knew well.

  • Mean: (210 + 220 + 225 + 240 + 300) ÷ 5 = 1,195 ÷ 5 = 239 WPM
  • Median: 225 WPM

The mean has been pulled up to 239 by one easy passage, and four of the five attempts are below it. The median, 225, sits among the results the reader actually produces. With only a handful of measurements, one odd value is likely, and the median is the summary that survives it.

A worked example

Here are five made-up attempts by one reader over three days. Each rate is words × 60 ÷ seconds.

Attempt Passage Words Time Raw WPM Correct Counts?
1 General non-fiction 384 96 sec 240 4 of 4 Yes
2 Everyday science 330 90 sec 220 3 of 4 Yes
3 History 405 108 sec 225 4 of 4 Yes
4 Narrative 440 88 sec 300 2 of 4 No
5 Technology 350 100 sec 210 3 of 4 Yes

Checking the first row: 384 × 60 = 23,040, and 23,040 ÷ 96 = 240 WPM. The others follow the same way: 19,800 ÷ 90 = 220, 24,300 ÷ 108 = 225, 26,400 ÷ 88 = 300 and 21,000 ÷ 100 = 210.

Step 1: apply the comprehension cut-off. Attempt 4 is the fastest by a wide margin, but only two of four answers were right. Its comprehension-adjusted score on the test would be 300 × 0.5 = 150. That is a site-defined comparison metric, not a measure of words understood or proof that half the reading was wasted. This attempt is dropped under the protocol’s chosen rule.

Step 2: sort what remains. 210, 220, 225, 240.

Step 3: take the median. There are four values, so the median is the mean of the middle two: (220 + 225) ÷ 2 = 222.5 WPM.

Step 4: round. The baseline is about 220 WPM.

Rounding down to 220 is deliberate. The measurements are not precise enough to justify 222.5, and a slightly conservative figure produces estimates you are more likely to meet.

Notice what the protocol changes. Reporting the best attempt would have given 300 WPM, despite its weaker quiz result. Averaging all five without the cut-off would have given 239 WPM. The provisional baseline of 220 is lower than both and is a more cautious planning choice. Actual reading sessions will show whether it fits.

Keep separate baselines for easy and dense material

A single figure hides the fact that you do not read everything the same way. Carver (1992) describes reading as several distinct processes with different typical rates, and reading to learn is markedly slower than ordinary reading in his account. Your own numbers will show the same split.

Two baselines cover most needs:

  • Everyday baseline: general articles, email, narrative, familiar subjects. The protocol above gives you this.
  • Dense baseline: textbooks, technical documentation, contracts, academic papers, anything in a field that is new to you.

The site’s passages are written for a general reader, so they will not tell you your speed on a pharmacology textbook. For the dense baseline, use the manual method on the real material: take three sections you have not read, get the word count of each, time yourself, and check your understanding by summarising each section from memory. Take the median as before.

Expect the dense figure to be clearly lower. The size of the gap is personal and depends on the subject, which is why it is worth measuring instead of guessing. Our guide to reading speed by text type describes how different kinds of material behave.

Using the baseline in the calculator

From the result screen, the speed test can send a result to the reading time calculator as your tested speed. That passes one attempt, so once you have worked out a median, enter the baseline as a custom WPM value instead.

Continuing the example, with an everyday baseline of 220 WPM and, say, a dense baseline of 160 WPM measured separately, a 4,400-word document comes out as:

  • 4,400 ÷ 220 = 20 minutes as everyday reading
  • 4,400 ÷ 160 = 27.5 minutes, or 27 min 30 sec, as dense reading
  • 4,400 ÷ 238 = 18.49 minutes, or 18 min 29 sec, at the population benchmark

For this reader the benchmark would understate the time in both cases, by a minute and a half for easy material and by nine minutes for dense material. That difference is the whole point of having a baseline. The division itself is explained in how reading time is calculated.

A baseline still describes uninterrupted reading. It does not include note-taking, re-reading, looking things up or breaks. Forgetting that allowance is one of the errors covered in our guide to common reading time calculator mistakes.

When to re-test

A baseline is not permanent. Repeat the protocol when:

  • your estimates are consistently off, for example you regularly take a quarter longer than the calculator says
  • you start reading a new kind of material, such as a new subject, profession or language
  • you change how you read, for instance from paper to a tablet
  • you have spent several months reading much more than before
  • your eyesight, health or working conditions have changed

Otherwise, checking once or twice a year is a reasonable rule of thumb. Because the test has five passages and a passage should not be repeated while you still remember it, leave a long gap between rounds or use the manual method with new text.

Expect small movements. A baseline that shifts from 220 to 230 WPM could reflect ordinary variation; this informal method has no established margin of error. Compare several new results with your old ones, not with the 238 WPM benchmark, which describes a population and was never a target for any one reader.