Speedtype Lab
Product notes, worked examples, and reproducible typing guides
The Lab explains what Speedtype actually measures and how its features behave. Every article is written and reviewed by Saurabh against current source code, passage records, or automated tests. Claims are limited to the app's evidence: no universal WPM benchmark, guaranteed improvement, or invented user result is used.
Start here
Three product behaviors worth verifying
Why Speedtype Treats A and a as Different Keystrokes
See exactly how Speedtype grades capitalization, spaces, punctuation, and corrected input, with a case-mismatch example you can reproduce.
Read the documented example
Gross WPM, Net WPM, and Accuracy: Four Worked Results
Follow four calculations from attempted keystrokes to gross WPM, net WPM, accuracy, and unresolved errors using Speedtype formulas.
Read the documented example
Inside the Speedtype Race Bot: Pace, Progress, and Win Conditions
Understand the simulated race bot, its target WPM settings, pace variation, progress calculation, and the exact condition that ends a race.
Read the documented example
Complete library
Read the implementation before judging the number
Scoring articles include calculator-ready examples. Product notes explain exact start, progress, correction, and finish rules. Practice protocols separate personal observations from claims the project has not tested. Language and editorial notes describe where software checks end and human review begins.
Why Speedtype Treats A and a as Different Keystrokes
See exactly how Speedtype grades capitalization, spaces, punctuation, and corrected input, with a case-mismatch example you can reproduce.
Read the documented example
Gross WPM, Net WPM, and Accuracy: Four Worked Results
Follow four calculations from attempted keystrokes to gross WPM, net WPM, accuracy, and unresolved errors using Speedtype formulas.
Read the documented example
What Backspace and Corrections Change in a Speedtype Result
Learn how attempted keystrokes, unresolved errors, completed words, and the current word affect a Speedtype result after corrections.
Read the documented example
Inside the Speedtype Race Bot: Pace, Progress, and Win Conditions
Understand the simulated race bot, its target WPM settings, pace variation, progress calculation, and the exact condition that ends a race.
Read the documented example
How Speedtype Writes and Reviews English and Marathi Passages
A transparent account of passage authorship, minimum checks, difficulty labels, English review, and Marathi native-language approval.
Read the documented example
A Seven-Session Typing Baseline You Can Reproduce
Use a controlled seven-session protocol to establish your own typing baseline without relying on universal targets or guaranteed gains.
Read the documented example
Why English and Marathi WPM Should Be Compared Separately
See how script, input method, Unicode handling, and the five-character WPM unit affect fair comparisons between English and Marathi sessions.
Read the documented example
How to Read a Speedtype Result Card Without Chasing One Number
Interpret speed, gross WPM, accuracy, errors, keystrokes, and time together, then decide what your next controlled run should test.
Read the documented example
Editorial standard
What earns a place in the Lab
An article must answer a specific user question with current Speedtype evidence. It must state the tested behavior, provide a way to reproduce or calculate it, identify important limitations, and link to the maintained product reference. Old generic drafts and unsupported numerical promises are not published.
A source change can make an article stale, so each page carries a review date and evidence note. Confirmed corrections are reviewed before the page date changes. Reader feedback is moderated and never published without explicit permission.
Audit the project yourself
Read the formulas, creator information, passage workflow, and current privacy disclosures. Then reproduce a behavior in the app and report any mismatch with the smallest possible example.
Use the articles as documentation, then create your own evidence under controlled settings.