03 — Exam Pattern and Strategy

Assessment #2 — Technical, 45 minutes Read this once a week out from the exam, then again the night before.

1. Confirmed mechanics

Mechanic	Finding	Tag	Source
Duration	45 minutes	`[CONFIRMED]`	Official job page + PrepInsta ×2
Question count	45	`[LIKELY]`	PrepInsta ×2; candidate counts sum to 45
Pace	60 seconds per question	`[INFERRED]`	arithmetic
Timing model	Pooled across the whole technical assessment, not per-sub-area timers	`[LIKELY]`	"45 mins total, shared"
Navigation within a section	Allowed	`[LIKELY]`	"intra-sectional navigation is allowed"
Navigation between sections	Allowed	`[REPORTED]`	"Inter Sectional Navigation is allowed"
Negative marking	None	`[CONFIRMED]`	3 independent aggregators
Cutoff structure	Overall technical cutoff plus a sub-sectional cutoff per sub-area	`[LIKELY]`	PrepInsta ×2
Cutoff value	Not published. Any specific % you read online is invented	`[CONFIRMED]`	aggregators say so themselves
Elimination	Yes — candidate calls it "the most elimination-heavy round"	`[LIKELY]`	REV-001 + PrepInsta
Pen and paper	Not allowed	`[REPORTED]`	REV-001, confirmed on follow-up
Calculator	No evidence either way	`[UNVERIFIED — model inference]`	assume none
Platform	One mention of "Lucid"; uncorroborated. Whether a Safe Exam Browser is used was asked in the thread and never answered	`[REPORTED]`	REV-003
All OA rounds same day	Assessments #1–#3 run back to back; communication and interview are separate dates	`[REPORTED]`	REV-001

The three that change how you sit the paper

No negative marking + free navigation. There is never a reason to leave a blank, and never a reason to grind on a hard question the first time you see it.
Sub-sectional cutoffs. You cannot trade a strong Section B against a zero in Section A. Breadth beats depth. This is why the tier list has no Tier 4.
No pen and paper. Every pseudocode trace happens in your head. This single constraint should reshape how you practise — see §5.

2. Time allocation

45 minutes, 45 questions. Sections are not equally dense, so do not spend equal time.

Order	Section	Qs	Budget	Sec/q	Why this budget
1st	A — MS Office	7	6 min	51 s	Mostly recall. You either know Ctrl+M or you don't — deciding takes seconds
2nd	C — Net/Sec/Cloud	14	12 min	51 s	Definition-heavy. Scenario wrappers are long to read but quick to answer
3rd	D — Computer Science	14	13 min	56 s	SQL and OOP output questions need a little more thought
4th	B — Pseudocode	10	13 min	78 s	The only section where a trace genuinely takes over a minute
—	Reserve	—	1 min	—	Final sweep for unanswered questions

The principle: buy time for Section B by being fast in Section A. Section A is 16% of the paper but should consume 13% of the clock. Section B is 22% of the paper and should consume 29%.

The 60-second rule

Every question gets at most 60 seconds on the first pass. At 60 seconds you either answer or flag and move. There is no third option. A pseudocode question you have stared at for three minutes has cost you two other questions you could have answered outright — and with sub-sectional cutoffs, those two might have been in a different sub-area you now fail.

3. Question-order strategy

Three-pass method

Pass 1 — Harvest (target 28 min). Go start to finish. Answer only what you can settle in under 60 seconds. Flag everything else and move immediately. Do not read a flagged question twice. Expect to answer 30–34 of 45.

Pass 2 — Work (target 14 min). Return to flagged questions, hardest-section-last: finish Section A and C flags first (they resolve quickly), then D, then B. Pseudocode traces are last because they are the most expensive per mark.

Pass 3 — Sweep (target 2 min). Scan for anything unanswered. Fill in every blank. No negative marking means a blank is strictly worse than a guess.

When to guess and when to skip

Situation	Action
You can eliminate 2 of 4 options	Guess now. 50% is a good return on 20 seconds
You can eliminate 1 of 4	Flag it, guess in Pass 3
You recognise the topic but not the answer	Flag it — Pass 2 often unlocks it after seeing a related question
Pseudocode, and you have not finished the trace at 60 s	Flag it. Do not sink more time on the first pass
"All of the mentioned options" is an option	In this corpus it is correct more often than chance. When two options are individually defensible, it is the percentage play
Option set is `1 and 2 / 2 and 3 / 1,2 and 3 / 1 and 3`	Evaluate the statements, not the options. Decide each of 1, 2, 3 true/false, then pick. Trying to read the option list first is how people lose a minute

Shortcuts specific to this paper

Pseudocode: read the guard before the body. In 4 of 16 observed local traces, the first condition was false and the entire body was dead code. Evaluating one condition answers the whole question. Always evaluate the outermost condition first.
(a+b+c) > (c+b+a) is always FALSE. Same operands, both sides equal, > fails on equality. This exact device appeared twice in the local corpus. Recognise it and skip the arithmetic entirely.
Section A image questions: look at the image first, then the stem. The stem is usually a template ("apply the formula in the formula bar"); the image carries all the information.

4. Common failure modes

Each one below was reported by a candidate or is directly visible in the corpus.

#	Failure mode	Evidence	How to avoid it
1	Burning the clock on one pseudocode trace	REV-001: pseudocode was the hardest section, 10–12 line snippets, no paper	Enforce the 60-second rule mechanically. Flag and move
2	Trying to trace without paper and losing your place mid-way	REV-001: "No pen and paper allowed, so you need to visualize carefully"	Practise traces mentally from day one (§5). Practising with paper builds the wrong reflex
3	Reading `&` as `&&` (or as `+`)	9 of 19 local pseudocode items hinge on a bitwise op; `3&4=0` and `10&1=0` both collapse to zero	Drill the AND/XOR truth tables until they are automatic. When you see `&`, convert to binary
4	Using a variable's old value after it was reassigned	4 local items reassign a variable before a later condition reads it	Trace top-to-bottom and physically re-state the current value at each line
5	Assuming an outer-true `if` means the body runs	Item 68: outer true, inner `if(a>b)` false, nothing changes. The published key got this wrong	Evaluate every guard, not just the first
6	Neglecting a sub-area entirely	Sub-sectional cutoffs	Even in a 7-hour panic plan, touch all four sections
7	Preparing from PrepInsta only	Aggregators describe a 3-section paper; the current one has 4	Prepare Section D. It is 31% of the paper and absent from every aggregator syllabus
8	Skipping MS Office as "too easy to study"	Candidate rated it "easy medium" — but it is 7 questions with a sub-sectional cutoff	4 hours on Section A is worth more than 4 more hours on pseudocode
9	Ignoring previous-year questions	REV-002: 22–31% were verbatim repeats	Work every question bank in this library. Repeats are free marks
10	Panicking at the Networking section	REV-002 called it the hardest; REV-001 reports scenario-based OSI questions	Scenario wrappers hide simple recall. Strip the story, find the concept
11	Leaving blanks	No negative marking	Pass 3 exists solely to prevent this
12	Mis-reading multi-statement options under time pressure	The `1 and 2 / 2 and 3 / …` format is heavily used in the local deck	Evaluate statements first, options second

5. How to practise, given no pen and paper

This deserves its own section because it is the constraint most candidates prepare wrong.

Do not practise pseudocode with a notepad. You will build a reflex you cannot use in the exam.

Instead:

Trace out loud. Say each variable's value as it changes. Speech is a working memory aid that transfers to silent subvocalising in the exam.
Track at most three variables. Almost every observed snippet uses exactly three (a, b, c or p, q, r). Hold them as a mental triple and update it line by line.
Convert to binary immediately. When you see & or ^, convert both operands to 4-bit binary before doing anything else. 3 & 4 → 011 & 100 → 000 → 0.
Check the guard first, always. Before tracing a body, evaluate whether it runs. This alone resolves ~20% of questions in seconds.
Time yourself at 60 seconds from the first practice session, not the last.

6. Pre-exam checklist

Hardware

[ ] Laptop/desktop fully charged and plugged in
[ ] Webcam working and unobstructed (Assessment #1 is proctored)
[ ] Microphone tested — REV-002 stresses using one audio input and not changing it
[ ] Mouse working; a dead trackpad mid-exam costs minutes
[ ] Close every other application (RAM and pop-ups both hurt)

Internet

[ ] Primary connection tested at the exam location
[ ] Backup: mobile hotspot, already paired and tested
[ ] Do not switch networks mid-assessment — REV-002 lists this as a top "don't"
[ ] Ask housemates not to stream during your slot

Environment

[ ] Quiet room, door closed, phone silent and out of reach
[ ] Desk clear — no paper, and none needed
[ ] Good lighting on your face for the proctored rounds
[ ] Water within reach; no bathroom break planned

Software & access

[ ] Complete the Mock Assessment (20 min, mandatory) at least a day early, not on exam day
[ ] Browser updated; disable extensions and pop-up blockers
[ ] Assessment link bookmarked and opened once beforehand
[ ] Login credentials confirmed working

Documents

[ ] Government photo ID
[ ] Accenture/NEO registration ID
[ ] Resume copy accessible

The night before

[ ] Read 91-CHEATSHEET.md once — do not learn anything new
[ ] Sleep. A tired trace is a wrong trace

7. Day-of timeline

Time	Action
T-90 min	Eat properly. Not heavy
T-60 min	Final `91-CHEATSHEET.md` read. Ports, OSI layers, operator precedence, cloud models. Read, don't study
T-40 min	Set up the room. Close applications. Start the hotspot as a hot standby
T-30 min	Bathroom. Water on the desk
T-20 min	Open the assessment portal. Confirm it loads
T-10 min	Sit down. Breathe. Do not open notes again
T-0	Mock Assessment (20 min) — hardware/software/internet check
	Assessment #1 — Behavioral + Cognitive, 40 min. Personality (no elimination) then gamified puzzles (is eliminating). Bubble game: ~15 s/question, 25 questions. Practise decimals — REV-002 saw them
	Assessment #2 — Technical, 45 min. Run the three-pass method. Watch the clock at the 28-minute mark; if Pass 1 is not done, start flagging harder
	Assessment #3 — Coding, 60 min, 3 questions. Java / Python / C# (C++ was unavailable to both local candidates despite the official page listing it). Java users: comment out the `throw new Exception(...)` line in the template or your code will not run
After	Communication assessment arrives separately, then interviews ~20–25 days later

8. Scoring targets

No published cutoff exists, so these are set to make the cutoff irrelevant.

Target	Overall	Section A	Section B	Section C	Section D
Safe	32+/45 (71%)	5/7	7/10	10/14	10/14
Comfortable	27/45 (60%)	4/7	6/10	8/14	9/14
Minimum viable	23/45 (51%)	4/7	4/10	7/14	8/14

Calibration point: REV-002 self-reported 32–34 on the technical round and progressed all the way to interview. That is the only real datapoint available, and it lands squarely in the "Safe" band. Aim there.

Note the minimum viable row never drops below 50% in any sub-area — that is the sub-sectional cutoff constraint expressed as a score. An overall 23/45 built from 7/7 in Section A and 2/10 in Section B is a likely rejection even though the total looks passable.