If you have used a single-word reader, you may have noticed the words are not centred. One letter sits in a fixed spot and the rest of the word arranges itself around it. That spot has a name: the optimal recognition point, usually shortened to ORP.

Why not just centre the word

Readers do not identify a word letter by letter from the left. Word recognition is fastest when the gaze lands slightly left of the middle — for a five-letter word, around the second character. Centre the word instead and your eye has to re-centre on every word, which reintroduces exactly the movement the format exists to remove.

The practical consequence is that alignment matters more than speed. A reader that flashes words at 500 words per minute without aligning them is worse than one at 300 that does, because you are doing micro-saccades on every single word.

Where the idea comes from

The technique itself is old — rapid serial visual presentation goes back to reading research in the 1970s. The specific ORP alignment was popularised by Spritz, whose patent application covers the alignment method. The underlying bucket table — which letter to align for a word of a given length — appears in many open-source implementations, ours included.

The part that is better supported than ORP itself

Display duration can account for word length. In a study of readers with central vision loss, varying how long each word stayed on screen by its length improved reading rate by about a third. That result belongs to that group; it is not a predicted gain for every reader. A fixed duration treats “a” and “extraordinarily” as equal work.

The same logic extends to punctuation. A sentence boundary needs a longer gap than a word boundary, or sentences run together — and that costs comprehension for reasons unrelated to speed.

What ORP does not fix

Alignment reduces the need to move across a line, but no study here shows that ORP raises comprehension or establishes a universal WPM ceiling. In one French Spritz experiment, literal recall fell while visual fatigue rose. That finding does not test every RSVP design.

So ORP is best understood as a correctness requirement rather than a feature. If a reader does it, single-word reading works as intended. If it does not, the format is mostly noise. It is not the thing that makes you faster.

More on the mechanism in what RSVP reading actually does, and on the ceiling in does speed reading hurt comprehension.

References