September 6, 2026 · 9 min read
TL;DR: The exam picks an estimating technique from what the team has to estimate from. No experience of this kind of work means analogous, importing experience from elsewhere. A measured unit rate means parametric. Genuine uncertainty means three-point. A demand for accuracy and decomposable work means bottom-up, built by the people doing it.
A team is unsure whether its estimates are accurate because it has never built a product like this before. The options include three-point estimating, which averages an optimistic, most likely, and pessimistic figure and sounds exactly like what you reach for under uncertainty. It is not the keyed answer, and understanding why is most of what this topic tests.
Whichever one matches the raw material the stem gives you. Each technique consumes a different input, so the stem naming that input is the exam telling you the answer.
| Technique | What it consumes | The stem cue |
|---|---|---|
| Analogous | The actual result of a similar past piece of work | The team has no experience of this; someone knows a comparable job |
| Parametric | A measured unit rate multiplied by a quantity | Historical data establishing a rate per unit |
| Three-point | An optimistic, most likely, and pessimistic figure | Genuine uncertainty with a basis for all three points |
| Bottom-up | Task-level estimates from the people doing the work | Accuracy is prioritized and work can be decomposed |
| Relative sizing | A calibrated reference item | An agile team estimating in story points |
The team with no experience gets analogous estimating, drawing on the actual effort a similar feature took on a different project. The reasoning is that their difficulty is a lack of experience to estimate from, so the technique to reach for is the one that supplies experience from outside the team. It is coarse, and it is still the only option grounded in something that really happened rather than in the team's own guesswork.
Common trap: treating three-point estimating as the general-purpose accuracy technique. Averaging three numbers does not create information. If a team has no basis for its most likely figure, it has no basis for its optimistic or pessimistic figures either, and the average of three guesses is a confident-looking number with nothing underneath it. This bank offers three-point estimating as a distractor in exactly that situation, alongside t-shirt sizing and asking the product owner to estimate, and keys the technique that imports real experience instead.
When the stem prioritizes accuracy and the work is decomposable. Bottom-up estimates each task with the people who will do it, then rolls those up, which buys accuracy with effort.
A sponsor wanting accurate figures for budgeting gets bottom-up, and analogous estimating with scaling factors is offered as the trap. Analogous is legitimate elsewhere in this same bank, which is the point: the techniques are not ranked good to bad, they are matched to circumstances. Accuracy prioritized means bottom-up. Speed or an early rough order of magnitude means analogous. Reading which one the stem asked for matters more than knowing either definition.
Parametric estimating deserves a similar caution. It multiplies a measured rate by a quantity, so it needs data establishing that rate. Where a scenario offers no such rate, parametric gets ruled out however precise it sounds, and the same goes for function-point estimating without a function-point count.
Because convergence concentrates risk. A node with multiple predecessor activities can be delayed by any one of the paths feeding into it, so the chance of that activity starting late is higher than any single path suggests.
An auditor advising you to watch nodes showing path convergence, sometimes called sinks, is pointing at nodes in a network logic diagram that have multiple predecessor activities. The distractors are worth noting because they relocate the concept into the wrong diagram entirely: nodes in a work breakdown structure where branches rejoin, nodes in a conditional network where a path becomes deterministic, and nodes in a decision tree whose branches share a probability. Each borrows a real term from a different tool.
That sensitivity to which artifact a concept belongs to runs through this whole topic. The work breakdown structure decomposes deliverables, the network diagram sequences activities, and the schedule baseline is the approved version everything later gets measured against. Mixing them is a distractor pattern, not an accident.
Both compress a schedule that is already built, and the choice between them turns on whether you have money or tolerance for risk. Crashing adds resources and cost; fast tracking overlaps activities and accepts rework risk.
The bank keys this consistently. A project running under budget but behind schedule gets crashing, because the money is available and fast tracking would buy time by accepting rework risk instead, which is the wrong trade when budget exists. A constrained budget points the other way. One scenario offers a genuinely close miss: starting a night shift with the same co-located team is not crashing, because the team is not larger, it is merely awake longer. Bringing on additional teams across time zones is.
Because the two are swapped as literal options in at least six records here, it is worth reading the budget clause in every compression stem before choosing. The critical path post works through float and the compression trade-offs in more detail.
Which technique should you use when the team has never built anything like this? Analogous estimating, because it imports experience the team does not have. Three-point averaging of three baseless guesses adds confidence, not information.
What is the difference between analogous and parametric estimating? Analogous adapts one similar past result. Parametric multiplies a measured unit rate by a quantity and needs historical data to establish that rate.
When does the exam want bottom-up estimating? When accuracy is prioritized and the work decomposes. It is the most accurate and the most effortful.
What is path convergence? A node with multiple predecessor activities. It concentrates schedule risk, because a delay on any incoming path delays the converging activity.
The 121 schedule questions in PMP Practice sit inside a bank of 2,141 re-certified against PMBOK 8 and the July 2026 Exam Content Outline, and every wrong answer carries the reasoning that makes it wrong, explained rather than just marked. Start the free 20-question sample with no card and no signup.
Related: Critical Path Method and float · Cost baseline, budget, and reserves · Process domain study guide
Sources: PMI — PMP Examination Content Outline, 2026 (PDF) · PMI — PMBOK Guide standards
Which estimating technique should you use when the team has never built anything like this?
Analogous estimating, because it is the only technique that imports experience the team does not have. It takes what a comparable item actually cost elsewhere and adjusts for the differences. Three-point estimating feels like the accuracy answer, but averaging three guesses from a team with no basis for any of them produces a confident number with nothing underneath it.
What is the difference between analogous and parametric estimating?
Analogous estimating adapts the actual result of one similar past piece of work. Parametric estimating multiplies a measured unit rate by a quantity, so it needs historical data establishing that rate. If the scenario gives you no measured rate, parametric is not available no matter how precise it sounds.
When does the exam want bottom-up estimating?
When the stem prioritizes accuracy and the work can be decomposed. Bottom-up builds the estimate from task-level estimates made by the people doing the work and rolls them up, which is the most accurate approach available and also the most effort. If a stem stresses speed or an early rough figure, it is pointing at analogous instead.
What is path convergence in a network diagram?
A node with multiple predecessor activities, where several paths feed into one activity. It matters because it concentrates risk: a delay on any incoming path can delay the converging activity, which is why schedule risk clusters at those nodes.