Government typing exams, compared

Each row links to a practice page that explains that exam in depth. “Not stated” means the document we read does not say; it does not mean the rule does not exist.

Two rows need a warning. UPSSSC publishes its notices through a page we could not read, so its figures come from exam news coverage and are marked as reported. High Court is not one exam: every High Court writes its own rules, and the row shows the two courts whose documents we read.

Key depressions (SSC). SSC states its requirements in words per minute and in key depressions per hour, using five key depressions per word: 35 WPM is about 10,500 KDPH. The CGL data entry test does not state a speed at all — it gives a passage of about 2,000 key depressions and 15 minutes, which is 8,000 KDPH, or about 27 WPM, if you are to finish it. The WPM vs KPH guide converts between the units.

Words minus a penalty (Railways). RRB counts every word you type, ignores mistakes up to 5% of those words, then deducts ten words for each mistake beyond that, before dividing by the time. Half mistakes count as half. It also refuses to evaluate a script with fewer than 300 English or 250 Hindi words.

Speed plus a mistake ceiling (Delhi High Court). The 2020 Junior Judicial Assistant test counted words as characters with spaces divided by five, and passed a candidate who reached 35 WPM in 10 minutes with mistakes no more than 3% of the words typed.

The difference matters. Here is one 10-minute performance — 400 words, or 40 WPM — scored both ways as the mistakes rise:

RRB forgives the first 20 mistakes completely, then punishes hard: at 40 mistakes the same typing is worth 20 WPM. The Delhi rule never touches your speed, but a 13th mistake fails you outright. The same typist can pass one exam and fail the other with identical work.

SSC publishes its evaluation guidelines, and the RRB and West Central Railway notices use almost the same list, so this is the closest thing to a common standard. Full mistakes:

• Each omitted word or figure
• Each wrong word or figure substituted for the right one
• Each word or figure added that is not in the passage
• Each spelling error — a letter repeated, added, left out or swapped (“speling”, “seplling”)
• A repeated word (“I shall shall be”)
• An incomplete, half-typed word

Half mistakes:

• Spacing: no space between words, or a space inside a word
• Wrong capitalisation (English only)
• Punctuation added, left out or changed
• Two words transposed (“hope I” for “I hope”)
• A paragraph started with spaces instead of the Tab key

Two consequences are easy to miss. A single wrong letter makes the whole word a full mistake, so there is no partial credit for nearly-right words. And formatting counts: SSC wants one space after each punctuation mark and the Tab key at the start of each paragraph, and charges half a mistake for each paragraph indented with spaces.

This is the most common question and the one most often answered wrongly online. The documents disagree with each other:

• SSC: its instructions to candidates say they are not required to repeat the passage once finished, “however they are allowed to revise and correct their mistakes and inaccuracies, if any, during the prescribed time.”
• RRB: “Use of editing tools for correcting the mistakes of the typed matter is not permitted.” The West Central Railway notice adds that the test runs without editing tools or spell check.
• CPCT, DSSSB, High Courts: the documents we read do not say.

Our practice tests follow the documents: Backspace starts on for SSC CGL and CHSL and off for Railway NTPC. Every test has a switch, so you can practise the strict way whatever your exam. Even where correction is allowed, it costs time — fixing a word you typed three words ago means deleting three good words too.

Hindi speeds are consistently 5 WPM below English — 30 against 35 for SSC CHSL and DSSSB, 25 against 30 for RRB and CPCT. The font rule varies far more:

• SSC CHSL: Mangal only; any other font means non-qualification.
• RRB NTPC: Krutidev or Mangal.
• CPCT: scored as Hindi (Unicode); the portal shows Inscript and Remington (GAIL) charts.
• DSSSB (2019) and Rajasthan High Court (2017): Kruti Dev.

Mangal is a Unicode font and Krutidev is not, and they are typed on different layouts. If your exam says Mangal, practise on the Mangal Inscript test; if it allows Krutidev and that is what you know, use the Krutidev test. The Hindi typing hub explains the difference.

The Delhi High Court published the full roll-number-wise result of its 2020 Junior Judicial Assistant typing test, held in March 2022, with each candidate’s words typed, speed, mistakes and allowance. We read all 846 legible rows (absent and exempt candidates excluded):

54 candidates typed at 45 WPM or more and still failed. One typed 52.36 WPM — half as fast again as required — with 23 mistakes against an allowance of 16. The median passing candidate typed 45.3 WPM; the median failing candidate, 37.5. Speed got almost everyone over the line. Accuracy decided who passed.

1. Read your own notice first. Note the speed, duration, language, font and any rule on corrections. Everything else follows from those five facts.
2. Practise at the full duration. Sustaining 35 WPM for 10 or 15 minutes is a different task from a one-minute burst, so finish your preparation with full-length runs. Our exam pages default to the duration in each notice where it states one.
3. Train accuracy before speed. Aim for under 2% mistakes at your current speed, then push the speed. The Delhi results show why.
4. Practise the formatting. One space after punctuation, Tab for paragraphs, capitals exactly as printed. These are half mistakes that add up.
5. Decide your Backspace habit to match the exam — correct as you go for SSC, never for RRB.

If your raw speed is the problem, the how to type faster guide and the touch-typing course come first.

Every rule is quoted from the recruiting body’s own document, except UPSSSC, which is marked as reported. RRB and CPCT documents are cited through the Internet Archive copies we read because the official sites refuse automated access. The Delhi High Court figures are our count of the published result sheet; the mistake-scoring table is arithmetic from the published formulas.