PMP 2026 — Exam Cheat Sheet

PMP® 2026

Exam Cheat Sheet

The decision layer · read Parts 1 and 2 on exam morning

1 The answer-selection sequence

2 The trap catalog

3 Hard rules

4 Formulas and charts

5 Distinctions that decide items

6 Adaptive quick reference

7 Exam-day notes

How to Use This

This is not a summary of the three domain guides. It is the decision layer — the rules that actually pick the answer when two options both look right and you are tired at question 140.

Read Parts 1 and 2 the morning of the exam. They are the highest-yield pages here. Everything else is lookup.

Exam mechanics: 180 questions, 230 minutes, two 10-minute breaks (after Q60 and Q120). Roughly 77 seconds per question. Formats: multiple choice, multiple response, matching, hotspot, fill-in-the-blank, drag-and-drop, and graphic-based items.

Domains: People 33% (~59 items) · Process 41% (~74 items) · Business Environment 26% (~47 items). Roughly 40% predictive, 60% adaptive or hybrid.

Pacing checkpoints: Q60 by the first break, Q120 by the second. If you are behind, the fix is answering faster on the ones you know, not skipping.

PART 1 — The Answer-Selection Sequence

Run this in order. Most questions resolve by step 4.

1. Establish the world

Predictive, adaptive, or hybrid? If hybrid, which workstream does this event land in? The same event has opposite correct handling on either side.

Event	Predictive side	Adaptive side
New scope requested	Change request → CCB	Product owner reprioritises the backlog
Schedule pressure	Crash or fast-track	Reduce scope to MVP
Requirement unclear	Clarify and baseline	Refine in backlog, learn by iterating

2. Eliminate the disqualifiers

Concealment or misrepresentation — "only if raised," "without informing," "adjust the metrics," "delay disclosure until after," "downplay."
Dismissal — "regardless of impact on the relationship," options that brush off a stated concern.
Compliance traded away — anything that ranks a legal, safety, or regulatory requirement below schedule, cost, or a sponsor's preference.
Poison-pill qualifiers — read every option to its final word. with no cost ceiling · in the next release · informally · temporarily · regardless of · without approval.

3. Check for a hard boundary

These override everything else, including "assess first."

Trigger	Action
Harassment, discrimination, comparable conduct	Escalate immediately to HR/supervisor. No root-cause conversation, no ground rules, no mediation.
Acquisition, merger, major strategy shift	Escalate immediately to sponsor. The decision is above your authority.
Governance threshold breached	Follow the governance path. Threshold breach removes your discretion.
Blocker outside team control	Escalate — or daily loss exceeds sponsor tolerance, or team cannot resolve within 48 hours.

4. Ask what each option requires to already exist

The single highest-yield question on the exam. If an option depends on something that has not happened yet, it is not the answer — however sound it is as an eventual action.

Assess → decide/agree → formalise → execute → update documents. An option that skips a step is wrong even when its content is right.

5. Match the stem's literal verb

Stem asks for	Answer must
Align · resolve · reach agreement	Produce agreement. Analysis "only produces data — it does not align expectations directly."
Understand · determine cause · assess	Produce understanding. Facilitation is premature if there is nothing yet to discuss.
First · next	Be the earliest step whose prerequisites are already met.
Structured · systematic · proactive · innovative	Satisfy that literal word. Good practice that is not structured loses to a named structured method.
Prevent · avoid recurrence	Change the process or artifact — not add inspection or a reminder.

6. Prefer the option that includes all affected parties

Over any that engages one side, adjudicates between them, escalates past them, or documents instead of resolving.

PART 2 — The Trap Catalog

Nine failure shapes that decided real questions. Each is a wrong option that reads as reasonable.

1. Passive verbs

Real actions that stop one step short. Each critique below is verbatim from an official rationale.

Verb	Why it fails
Document · record	"An administrative step, not a leadership action."
Store · compile	"Does nothing to push that information out."
Submit · forward	"Passive and abdicates responsibility."
Monitor · observe	"Does not embed the practice into a controlled artifact."
Inform without acting	"Reflects poor responsiveness and weakens user trust."
Log as a risk	"Frames issues as threats, not systemic improvements."

2. The smuggled decision

Two options both begin with assess or evaluate and differ entirely in what follows. "Evaluate the risks and benefits of deploying the components *in the next release*" has already decided the timing. "Assess the components to determine their impact" stays open. Read the noun phrase after the verb.

3. Vague framework language

"Strategic framework," "comprehensive methodology," "holistic approach" with no named mechanism. Interpret the option as what it literally says, not as the strongest version you can imagine. One live item offered "a strategic project management framework to identify sources of overall project risk" — it read like a risk breakdown structure and turned out to be PESTLE, and lost.

4. Tool versus governing process

One option offers a specific tool or data source; another offers the process that tool operates within. Where the stem emphasises rigour, fairness, or auditability, the process wins — the tool is only as objective as the framework applying it. Confirmed twice: audits lost to metrics-first; prior-contract KPIs lost to structured evaluation.

5. Right action, wrong container

Storing something reusable in the project's repository "limits access and prevents organisational learning." Check not only whether something gets stored, but where.

6. Assumes a future outcome as already true

Letting work continue while a certification is applied for "assumes award of certification." Watch for while · pending · expecting used to justify present action on an unapproved basis.

7. Legitimate tool, wrong scope

The employee handbook is real, and "a broad set of standards for the entire organisation" when the problem is team-specific. Company-wide policy proposals from a single project observation "exceed the project manager's authority."

8. Accommodation that undermines the requirement

Adapt the *method* when a stakeholder names a genuine constraint — a time zone, a channel preference, a format.

Hold the requirement and coach when they are declining something the project needs because they "don't see the value." That is an awareness gap. "Adapting immediately to stakeholder preferences may seem like a quick fix, but it undermines the project's needs."

9. Solving one problem when the stem named three

When a stem stacks several distinct concerns — skill gap and burnout and quality — and asks for a sustainable or comprehensive fix, the answer addresses multiple concerns through a mechanism that keeps operating. Options that solve only the most visible problem lose.

PART 3 — Hard Rules

These do not bend. If an option violates one, it is wrong regardless of how sensible the rest of it sounds.

Compliance and ethics

Compliance, safety, and ethics always outrank schedule, cost, and pressure — including sponsor pressure.
Compliance-driven changes still go through change control. Mandatory ≠ pre-approved.
You do not negotiate an exception to a legal requirement. Even attempting it "has ethical implications."
Over-delivery is noncompliance where the client operates an approval process. Doing more without approval means doing it wrong.
Budgeting for an ongoing fine is neither ethical nor cost-effective.

Reporting integrity

Report the true status, explain the variance, present a corrective plan. Yellow with a recovery plan beats green, beats red-and-halt, beats delayed disclosure.

"Accurate reporting should not be postponed while waiting for more details." Investigation runs in parallel with honest disclosure, not in front of it.

Never let bad news reach a stakeholder from someone else first. The failure is not the bad news — it is who delivered it.

Project manager authority

Yours	Not yours
Facilitating alignment between conflicting stakeholders	Deciding between them, or asking the sponsor to decide
Evaluating and recommending the delivery approach	Delegating that evaluation to the sponsor
Analysing a change and routing it to the CCB	Approving or rejecting a change request
Raising a disproportionate governance burden	Unilaterally skipping governance steps
Coordinating and communicating about a technical problem	Personally researching and fixing it — that is the team's job
Enforcing contract terms	Discussing a vendor's bankruptcy with them directly

Contracts

Verify against the agreement before acting. Conceding immediately "undermines contractual leverage and sets a negative precedent."
A verbal instruction is not a contract change.
Any absence of named contract resources is a violation — there is no acceptable duration to negotiate down to.
Where a defect traces to a contractor's own prior work that was accepted, responsibility may be shared — acceptance is itself a checkpoint that was passed.

Closure

Cancelled and failed projects still get full formal closure. Lessons from failures are especially valuable, not skippable.
Do not release resources before closure is complete.
Procurement closes before administrative closure.
The final acceptance review exists so that no open issues remain — that is legal protection, not paperwork.

PART 4 — Formulas and Charts

Earned value

EV always comes first. Cost uses AC, schedule uses PV. Subtract for a variance, divide for an index. Positive or above 1.0 is good.

	Formula	Reading
CV	EV − AC	Negative = over budget
SV	EV − PV	Negative = behind schedule
CPI	EV ÷ AC	< 1.0 = over budget
SPI	EV ÷ PV	< 1.0 = behind schedule
ETC	EAC − AC	Cost of remaining work
VAC	BAC − EAC	Negative = forecast over budget
TCPI	(BAC − EV) ÷ (BAC − AC)	Efficiency needed from here

Cost and schedule are independent axes. On budget tells you nothing about schedule. If a stem asks whether work is being delivered as planned, the answer is SV, no matter how healthy CV looks.

EAC — selection is the whole question

The scenario says	Use
The variance was atypical and will not recur	EAC = AC + (BAC − EV)
Current cost performance will continue	EAC = BAC ÷ CPI
Both cost and schedule performance will continue	EAC = AC + [(BAC − EV) ÷ (CPI × SPI)]
The original estimate was fundamentally flawed	EAC = AC + a new bottom-up ETC

The arithmetic is trivial. The wrong options are the other formulas' correct results, so a clean-looking number does not confirm you chose right.

Reading charts

Chart	The one thing to get right
Burndown	Tracks work remaining. Actual line above target = behind. Higher is worse.
Burnup	Tracks work completed. Actual above target = ahead. Higher is better. A stepping total-scope line = scope added.
Why burnup over burndown	Burndown cannot distinguish "we added scope" from "we are slow." Burnup's separate scope line can.
Velocity	Stabilises after several sprints. Do not forecast from early-sprint velocity.
CFD — band widening	Work accumulating at that stage = bottleneck. Review WIP limits to relieve it.
CFD — flat band	Items stuck: resource constraint, technical blocker, or poor prioritisation.
CFD — zigzag	Coordination failure — unclear roles, poor communication between stages.
Lead vs cycle time	Lead = entire lifecycle including queue. Cycle = active work only. The gap is waiting time.
Pareto	Act on the vital few highest-frequency categories. Spreading effort evenly dilutes focus.
Control chart	Rule of seven: seven consecutive points one side of the mean = investigate, even inside the limits. Spec limits come from the customer; control limits come from the process.

Reserves

	Contingency	Management
Covers	Known risks, already assessed	Unknown risks
Position	Inside the cost baseline	Outside the baseline
To use	PM authority	Needs management approval

Reserves are for identified risks, not a discretionary pool. Newly discovered scope is a baseline change regardless of which pot the money comes from — and available funds never substitute for change control.

PART 5 — Distinctions That Decide Items

Pair	The discriminator
Qualitative vs quantitative risk	Broad prioritisation of all risks, first · deep modelling of the few that screening flagged
Risk vs issue vs blocker	Future and uncertain · present and realised · present and obstructing but still avoidable
QA vs QC	Process, prevention, audits · product, inspection, testing
Validate Scope vs Control Quality	Customer acceptance · internal correctness. QC comes first.
Quality vs grade	Meets requirements · category of features. Low grade can be fine; low quality never
Scope creep vs gold plating	External uncontrolled change · internal unrequested additions
Crashing vs fast-tracking	Add resources, costs more · overlap activities, riskier
Levelling vs smoothing	May move the end date · stays within float
Register vs repository	Raw, this project, living · validated, organisational, authoritative
Explicit vs tacit knowledge	Documents and repositories · conversation, pairing, mentoring, shadowing
Output vs outcome vs value	Delivered · used · worth what it cost
Leading vs lagging indicator	Predicts before it happens (team engagement) · confirms after (defect density)
Criteria vs metrics	What you want · the instrument that measures whether you got it
Collaborate vs compromise	Win-win problem solving · both sides give something up. Collaborate always wins
Mentoring vs coaching	Showing the way · drawing the answer out of them
Fixed price vs cost reimbursable	Seller carries cost risk · buyer carries it
Input vs output-based contract	Client specifies inputs and carries calculation risk · provider does
Squad vs individual resourcing	Agile requests intact teams with known velocity · predictive requests named people
Awareness gap vs resistance	"I don't see why" → coach · "I had no say" → involve

Sequences worth knowing cold

Change control: review with the team and assess impact → then formal request → CCB decides → communicate → implement → update all affected documents.
Risk materialises: review the register → does a planned response exist? → verify it fits actual conditions → update if not → implement. A planned response is not automatically the right response.
Value: value proposition (narrative) → value breakdown structure (quantified components) → quality control (verifies attributes achieved).
Culture assessment: collect (surveys, interviews) → interpret (cultural and political awareness tools) → act. Forwarding raw data skips the step that is your job.
Closure: formal acceptance → verify closure criteria → transition and knowledge transfer → retrospective → administrative closure and resource release.
Conflict: diagnose severity first. Leas levels 1–3, let the team resolve it. Level 4 (communication broken down) and above, intervene.

PART 6 — Adaptive Quick Reference

Roughly 60% of the exam is adaptive or hybrid.

Roles

Product owner — owns the vision, the backlog, and prioritisation. In agile the vision is theirs, not the PM's.
Scrum master — facilitates, removes impediments, coaches the process.
Team — self-organising, decides how the work gets done.

Ceremonies — and protecting them

Ceremony	Purpose
Sprint planning	What we commit to this iteration
Daily standup	Coordination and blockers — not status reporting to management
Sprint review	Demonstrate the increment, gather stakeholder feedback
Retrospective	Process quality — how we work. Applies to predictive projects too.

A reasonable question asked in the wrong ceremony gets redirected, not answered. Protecting a ceremony's purpose outranks the convenience of handling a request on the spot.

Estimating and decomposition

Epic → feature → story → task.
INVEST: Independent · Negotiable · Valuable · Estimable · Small (half a day to two days) · Testable.
Story points measure complexity. Velocity is points completed per sprint. Capacity must subtract standing non-project work (support, break-fix).
Definition of done — the shared completion standard. Acceptance criteria — what this specific story must satisfy.

Frameworks — one line each

	Defining mechanism
Scrum	Time-boxed sprints, 1–4 weeks. Work is pushed at sprint planning.
Kanban	No fixed iterations. Continuous flow, WIP limits. Work is pulled.
Scrumban	Kanban board with sprints. No predefined roles. Planning triggered by WIP dropping below a threshold.
Lean	Pull system, eliminate waste, maximise value.
XP	Test-first, continuous integration, refactoring, pair programming.
FDD	Feature-list driven, domain-driven design. Large complex projects.
DSDM	Six phases from feasibility study through post-project.
DA	A toolkit, not a prescription. Tailored to context.
SAFe	Enterprise scale. Program increment planning, agile release trains.

Lean waste — TIMWOODS

Transportation (handoffs lose information) · Inventory (unfinished work) · Motion (project switching) · Waiting · Overproduction (features nobody needs) · Overprocessing (analysing before it is needed) · Defects · Skills underutilised — including knowledge held by one person with no way to share it.

The agile-is-not-a-silver-bullet rules

Agile is not a risk management approach. Iterating does not manage risk automatically; retrospectives and standups are opportunities that must be used deliberately.
Value is realised at release, not at sprint completion. A finished increment sitting undeployed is an output, not value.
Never choose the approach by team comfort or precedent. If the team lacks the skills the project needs, train them — do not change the approach to fit them.

PART 7 — Exam-Day Notes

The mindset that scores

You are a servant leader on an empowered team, not a controller.
Collaboration is foundational, not one option among several. Options that bypass it "bypass the collaborative spirit foundational to effective project management."
Assume you have authority to act and a supportive organisation unless the stem says otherwise.
The answer is almost never do nothing, wait and see, or hope it improves.
Both premature yes and premature no are wrong when neither has been assessed.

When two options both look right

Which one is first? Check prerequisites.
Which one matches the literal verb in the stem?
Which one includes all affected parties?
Which one changes the artifact or process, rather than adding activity around it?
Which one has a qualifier in its final clause that quietly disqualifies it?

Pacing

77 seconds per question is the budget. Most items you know take 30–45 seconds, which banks time for the hard ones.
Long stems: read the last sentence first, then the options, then the stem. You often only need two facts from four sentences.
Flag and move on rather than stalling. A flagged item you return to fresh is worth more than three minutes of grinding.
Take both breaks. Hour three is where accuracy drops, and ten minutes of reset is worth more than ten minutes of questions.

Your specific watch-items

Burndown direction. Higher is worse. It is the easiest thing to invert under pressure.
EAC formula selection. Read the scenario wording; the wrong answers are the other formulas.
Hybrid workstream identification. Establish which side the event landed on before anything else.
The accommodation trap. "They prefer email updates" is not a constraint if the reason is "we don't see the value."
Match the source to the knowledge type. Numbers live in repositories; judgement lives in people.
Read to the final word. More items turn on a trailing qualifier than on domain knowledge.

One last thing. Where your instinct and a plausible-sounding softer option disagree, the procedural, systematic, artifact-based answer has been right more often than the warmer consultative one. Trust the process answer.

Page