Matching Character by Character

In address matching and data comparison you hold a copy against the original and check it one character at a time. Letters, digits, marks (full stops, commas, hyphens, slashes) and the order of the parts all count. A copy is the same only if every one of them agrees.

What counts as a difference

Kind of change Original Wrong copy
a letter changed Apartments Apartmants
a letter added or dropped Sitaram Sitaraam
two letters swapped Raigarh Riagarh
a digit changed or two swapped 492013 492031
a mark added or dropped Sector-4 Sector 4
a short form Road Rd.
two parts swapped Ward 7, Sector 2 Sector 2, Ward 7

In this course a wrong copy always differs in something you can see on the page. No copy differs only in its spacing or only in its capital letters.

The checking order

  • Numbers first. The house or flat number, the ward or sector number and the PIN. Two swapped digits are the quickest difference to catch.
  • Short words and marks next. No. or No, Road or Rd., a hyphen in Phase-2.
  • Long words last, letter by letter. Names of people, buildings, colonies and towns. Read them in sounds (Ma-hesh-wa-ri) so your eye cannot jump a letter.

Another way that works well is to split the line at its commas and compare it part by part, ticking each part as it passes.

The question shapes

  • One exact copy among four. Strike out each copy as soon as you find its difference. The one left standing still gets a full read before you mark it.
  • Numbered copies, "which is/are identical". Check every copy. The answer can be one copy, two, three or none of them.
  • "Which two are identical to each other". Take one part at a time, the house number first and then the PIN, and see which entries agree on it. An entry that stands alone on any part cannot be in the pair. Give the last two a full check against each other.
  • "NOT the same". Exactly one copy has a difference, and it can be anywhere in the line.
  • Pairs in two columns. Before you start, see whether the question counts the pairs that are the same or the pairs that differ. Read long numbers in groups of three: 650 823 149.

[SSC Tip] Keep your pen on the part of the copy that your eye is reading in the original. The pen stops the eye from sliding past a part it has not checked.

Records: Date and City Matching

A records question gives a small table of people with a date, a city and sometimes a department, and asks for the one person, or the number of people, who fit two or three conditions at once.

The method

  • Read the question before the table, and write its conditions on your sheet, for example: Stores, March, Bhopal.
  • Start with the condition that cuts the table most, often the city or the department, and tick the rows that pass it.
  • Check the other conditions on the ticked rows only.
  • For "how many", test every row against every condition. Do not stop at the first two that fit.
  • For "which entry agrees with the records", take the name in each option to its own row and read the other details there.
  • When the question leaves one person out ("apart from Ravi"), leave that row out of your answer.

Reading dates

  • Dates are written DD-MM-YYYY: day, then month, then year. The month is always the middle pair, so 04-09-2018 is 4 September, not 9 April.
  • "Born in March" means 03 in the middle. "Joined in 2019" means the last four digits are 2019. "October 1999" needs both.
  • Most tables are built from near misses: the day and month swapped (11-09 against 09-11), one digit changed (19 against 09), the right date in the wrong city, or the right city and date in the wrong department.

For example: the question asks who joined in June 2018 in Bhopal. A row reading 06-11-2018 and Bhopal fails, because its month is 11, November. A row reading 18-06-2018 and Indore fails on the city. A row reading 18-06-2018 and Bhopal passes.

Two habits

Keep your finger on the row you are checking, and read across it from left to right. Eyes that jump between rows pick up one person's date and another person's city.

When a count comes out as zero, or as every row in the table, check one more time. A table usually has at least one row that fits and several that just miss.

Indexing and Filing

Names by surname

  • The surname is the last word of the name. Sort by the surname first, letter by letter, with capitals ignored.
  • The first letter where two surnames differ decides the order: Mehra comes before Mehta, because R comes before T.
  • A surname that is the start of a longer one comes first: Sen before Sengupta, Shah before Shahi.
  • Names with the same surname go by the first name: Arun Mehta before Zoya Mehta.
  • The first name never jumps the surname. Zoya Mehra comes before Arun Mehta, because Mehra comes before Mehta.

For example: Chetan Joshi, Bela Jain, Anil Joshi, Diya Jaiswal. The surnames Jain and Jaiswal share JAI and part at the fourth letter (N before S), and Joshi comes after both because O comes after A. The two Joshis go by first name. So the index reads Bela Jain, Diya Jaiswal, Anil Joshi, Chetan Joshi.

Files and codes

  • Sort by the letters first, as in a dictionary: AB before AC, and AC before B.
  • Inside the same letters, sort by the number taken as a whole number: K-9, K-12, K-104.
  • Reading the digits as text gives K-104, K-12, K-9, because 1 comes before 9. That order is wrong for filing.
  • A quick test: when the numbers have no zeros in front, the one with fewer digits is the smaller number.

The method

  • Group the entries by the first letter where they differ, or by their letter code.
  • Sort inside each group, then write the full order as entry numbers on your sheet.
  • Only then look at the options. Two options often differ in one swapped pair, so check that pair again.
  • For "which comes third" or "which comes last", sort only as far as you need. For the last place, start from the end of the alphabet.

The stem tells you how the entries are ordered. Read it, and follow what it says rather than what you expect.

Codes, Roll Numbers and Timing

Reading a code by its rule

  • Everything you need is in the stem. Read the rule once, in full, and note the parts with their lengths, for example: centre 3 digits, shift 1 digit, serial 5 digits.
  • On your rough sheet, mark each code where its parts meet: 286/1/40375.
  • Check only the parts the question asks about, in the order it asks them.
  • Use the rule exactly as it is printed. Do not guess a pattern, and do no arithmetic on the digits.

Common mistakes

  • The centre digits in another order: 325 where the question wants 352.
  • The wanted digits sitting in another part of the code, such as a seat number that happens to match the centre you are looking for.
  • The same venue number in a different city. GWL-05 and SGR-05 are different venues, because a venue number belongs to its city.
  • The right centre with the wrong shift, or the right shift at the wrong centre.

For a count, go through every code in the list. For "which two", group the codes by the parts the question names and look for the group of two.

Time targets

  • address matching: about 20–30 seconds
  • data comparison: about 30–40 seconds
  • date and city matching: about 30–45 seconds
  • indexing: about 30–40 seconds
  • code classification: about 30–40 seconds

The Reasoning section of Tier 1 has 25 questions and its own locked 15-minute clock, 36 seconds a question on average. Address questions sit well under that average. The other types run close to it, and a long table can take up to 45 seconds. If a question is still open when its time target is up, mark it, move on and come back at the end.

Checks for the last week

  • Numbers first. House number, ward and PIN before any word.
  • Every copy. In "which is/are identical", check all the copies. The answer may be none of them.
  • Month in the middle. 03-07-1997 is 3 July.
  • Every row. For a count, test every row against every condition.
  • Surname first. Then the first name, letter by letter, with the shorter start first.
  • Whole numbers. K-9 before K-12 before K-104.
  • Rule, not pattern. Read a code only by the rule in the stem.