top of page

Schedule risk analysis (QSRA): what a P80 date means for your programme

11 hours ago
14 min read

By Roman Bazelchuk | NEC Accredited Project Manager | APMG Project Planning and Control

Founder, NEC Planning Solutions Ltd


The request tends to arrive at a tender or a stage gate, in one line of a client's email: provide a schedule risk analysis and a P80 date. For the contractor's director the questions behind that line are commercial ones. What does the number mean? What does putting it in front of the client commit the business to? And can the programme that produced it carry that weight?


A schedule risk analysis, often called a QSRA, runs a programme many times over with realistic ranges on its durations and main risks. It reports how likely each finish date is. A P80 date is the date by which 80 per cent of those runs finish. It measures confidence. It is not a contract date.


On an NEC contract there is a second half to the question. Clause 31.2 of the NEC4 ECC already asks every programme submitted for acceptance to show time risk allowances, the Contractor's own provision for the risks it carries. A schedule risk analysis is the tool that can show whether those allowances are the right size and which risks they should cover. Much of the confusion about P80 dates on NEC work starts where the two are treated as separate subjects.


This article explains what a schedule risk analysis is, what P50 and P80 dates mean and what a P80 date does and does not commit a contractor to under the NEC4 ECC. It then shows how the analysis sizes time risk allowances without pricing the same risk twice, when the analysis is worth nothing and what the result should be used for. It builds on what counts as a reasonable time risk allowance rather than repeating it.



What is a schedule risk analysis?


A programme as it is normally built is deterministic. Every activity has one duration, the network gives one finish date and the date looks as certain as the arithmetic behind it. The work does not behave like that. Outputs vary, deliveries slip, approvals take longer and some of the risks on the register happen while others do not.


A schedule risk analysis replaces the single durations with ranges. A common form is a three-point estimate for each activity, an optimistic, a most likely and a pessimistic duration, agreed with the people who will do the work. The main risks are then added as events, each with a chance of happening and an effect on the activities it would hit: a late transformer delivery with a three in ten chance and a six-week effect, for example. Software then runs a Monte Carlo simulation. In each run it picks a duration for every activity from its range, decides whether each risk happens, recalculates the network and records the finish date. It repeats this many times.


The output is a spread of finish dates rather than one date. The US Government Accountability Office's guide to schedule assessment describes it as a probability distribution of possible completion dates and it is usually shown as an S-curve: the chance of finishing by each date, rising from nothing to certainty. The tools also report which risks and activities drive the spread, often as a tornado chart, which is the part a project team acts on.


The same method can be run on cost, where it is called a quantitative cost risk analysis, or QCRA. The two can be combined when the programme is resource-loaded. The rest of this article is about time.



What P50 and P80 mean


Each point on the S-curve is a confidence level. The P50 date is the date by which half the runs finished. The P80 date is the date by which 80 per cent finished, so on the model as built there is one chance in five of finishing later. Neither is best read as a forecast. Each is a statement about the model: the ranges, the risks and the logic that went into it.


The finish date on the programme itself tends to sit low on the curve and there is a structural reason for it. Where several paths of work have to finish before the next stage can start, the milestone waits for the latest of them, so, where the paths are independent of each other, the chances multiply. The Government Accountability Office's own example is two parallel paths, each with a 60 per cent chance of finishing on time, which together give the milestone a 36 per cent chance. A third such path takes it to about 22 per cent. This is merge bias and it is one reason the finish date of even a well-built programme tends to be later than the programme shows.


Figure 1 shows a typical result for a 52-week programme with a Completion Date in week 56, modelled on the risks the Contractor carries.


Schedule risk analysis S-curve for a 52-week programme on the Contractor's own risks, showing the chance of finishing by each week from week 48 to week 62. The curve rises from near zero to near 100 per cent. A diamond marks the programme's planned Completion at week 52, with about a 20 per cent chance. The P50 is week 55. A dashed line marks the Completion Date at week 56, where the chance is about 60 per cent. A red bar marks the remaining 40 per cent chance of finishing after it. The P80 is week 58.
Figure 1: a schedule risk analysis on the Contractor's own risks: the programme's own week 52 has about a one in five chance, the P50 is week 55, the Completion Date in week 56 has about 60 per cent, leaving about a 40 per cent chance of finishing late. The P80 is week 58.

Read that way, the figure answers the director's question directly. The programme's own date of week 52 has about a one in five chance. The Completion Date of week 56 has about a 60 per cent chance, which leaves about two chances in five of finishing late on the risks the Contractor carries. The P80 date of week 58 is two weeks beyond the contract date. That is on the Contractor's risks alone: a P80 for the client would come from a run on every risk and would normally be later. A client asking for a P80 is asking how much headroom the plan really has. The confidence level to work to is a choice for the business that owns the risk. The Government Accountability Office's guide gives the example of an organisation picking the date it is 70 per cent confident of reaching.



What a P80 date commits you to under an NEC contract


Under clause 30.1 of the NEC4 ECC the Contractor has to achieve Completion no later than the Completion Date and clause 30.3 does the same for the Condition attached to each Key Date. That is the promise. Where Option X7 applies, delay damages run from the Completion Date. A P80 date is not part of that promise unless the business offers it as one of those dates. A Scope can make the analysis a deliverable, but producing a P80 is a different obligation from achieving one.


The programme carries two dates that matter here. Planned Completion is the date the Contractor's programme forecasts. The Completion Date is the date in the contract. The gap between them is terminal float, which the terminal float guide explains is not the same thing as a time risk allowance. A schedule risk analysis tells the Contractor whether that gap is enough and at what confidence.


At tender the analysis does its most useful work. Where the client leaves the bidder to state the Completion Date in Contract Data part two, the analysis shows what confidence the offered date carries before anyone signs. Where the client fixes the date in part one, it shows how much of the risk of delay damages the business is accepting and how much time-related cost the price should carry against it. Either way the decision is made with a number in front of the director rather than a feeling, which is the same argument the tender programme rests on.


During delivery a client may want the analysis repeated. The place for that requirement is the Scope. The last item of clause 31.2 brings in other information the Scope requires a programme to show, so a requirement written there is part of what the Project Manager checks under clause 31.3. A request outside the contract is a service the Contractor can price.



Schedule risk analysis and the NEC time risk allowance


Clause 31.2 asks each programme submitted for acceptance to show provision for float and time risk allowances. NEC's user guide for the ECC treats the allowances as the Contractor's own, part of planning realistically for the risks it carries. Practitioners writing on the NEC website in December 2024 added that an allowance needs a reasoned basis that shows it is realistic, should be measurable and verifiable rather than a generic percentage and that a programme which leaves allowances out should not be accepted because it does not show what the contract requires.


That is where a schedule risk analysis earns its place on an NEC job. The ranges agreed for each activity and the risks mapped to it are the record that makes an allowance measurable. The analysis shows where the time risk sits and how much the programme as a whole needs, so the allowances can be placed on the activities that carry the risk, at a size the record supports.


The analysis has to respect the contract's split of risk, or it will price the same risk twice. A time risk allowance covers the Contractor's risks. Some of the risks on a typical register belong to the Client under the NEC4 ECC, as compensation events. When one of those happens, the delay is assessed under clause 63.5 against the Accepted Programme current at the dividing date. Table 1 sets out how common schedule risks fall.


Risk in the model

Who carries it under the NEC4 ECC

In the time risk allowance?

Outputs, productivity and the Contractor's own supply chain

The Contractor

Yes

Weather that does not meet the 60.1(13) test, including anything up to the once-in-ten-years level

The Contractor, because only weather meeting that test is a compensation event

Yes

A weather measurement, recorded within a calendar month, before the Completion Date for the whole of the works and at the place the Contract Data names, that the weather data show to occur on average less often than once in ten years

The Client, as a compensation event under 60.1(13), assessed on the difference only

No

Physical conditions within the Site, other than weather, that an experienced contractor would have judged at the Contract Date to have such a small chance of occurring that it would have been unreasonable to allow for them

The Client, as a compensation event under 60.1(12), assessed only on the difference from the conditions it would have been reasonable to allow for

No, except the part it would have been reasonable to allow for

The Client late with something it is to provide, against the date on the Accepted Programme

The Client, as a compensation event under 60.1(3)

No

A late reply from the Project Manager or Supervisor, outside the time the contract requires

The Client, as a compensation event under 60.1(6)

No


Table 1: who carries common schedule risks under the NEC4 ECC. The main list of compensation events is in clause 60.1. Some main and secondary Options add others and the Contract Data can state additional events.


Put the Client's risks into the allowances and the programme grows by time the contract deals with separately, through compensation events. The tender programme looks slow and the Project Manager has grounds to ask what the allowances are for. Leave the Contractor's own risks out and nothing protects the date when they happen. The cleanest way to keep the two apart is to run the analysis twice: once on the Contractor's risks to size the allowances and once on every risk to show the director what the date really carries.


The same split applies inside a quotation. Under clause 63.8 the assessment of a compensation event allows for the cost and time risks that have a significant chance of occurring and are not compensation events themselves, which is a small schedule risk question of its own.



When a schedule risk analysis is worth nothing


A schedule risk analysis is only as good as the programme and the ranges behind it. The result looks equally authoritative either way. Four failures are behind many of them.


Logic that cannot carry the risk


The simulation moves activities by pushing them through the logic. Where an activity has no successor, its overrun reaches nothing. Where a hard constraint holds an activity in place, the model cannot move it at all and the risk disappears from the result. The Government Accountability Office's guide asks for complete logic, all of the work and as few constraints as possible, each with a reason. The Primavera P6 guide covers why constraints distort a programme even before a risk model touches it. The five signs of a programme that has lost its integrity are the checks to run first.


Ranges set in the room


Ranges agreed in a workshop are exposed to pressure to agree with the room. The Government Accountability Office's guide asks for the experts to be interviewed, with their views kept anonymous and no one in authority in the room. It also warns that workshops can bring pressure to conform. A range that starts from the duration already on the programme also inherits whatever optimism that duration carries. A range that the record can support starts from what similar work actually achieved, from the risk register and from the people who will do the work. It is challenged before it is entered. A range nobody can explain is a guess with a decimal point.


Risks that move together, modelled apart


Activities that rely on the same crew or the same assumption, or are hit by the same risk, such as one supplier or the weather, tend to overrun together. Modelled as if each one were independent, one activity's overrun is offset by another's underrun within the same run, so the spread narrows and the P80 date comes out earlier than it should. Good practice links them, so that a bad month for one is a bad month for the others.


The wrong programme


Crossrail's head of programme risk, writing in 2018, described running the analysis on the Level 1 schedule because it was the only one that showed the logic between projects. The same paper records that the choice of schedule is vital. A summary programme that drops the points where paths merge also drops the merge bias. The Government Accountability Office's guide accepts a summary where time or resources rule out the full programme, but warns that one condensed too far hides long activities, hand-offs, how float is spread and the points where paths merge.



NEC Planning Solutions runs schedule risk analysis on Primavera P6 programmes: a check that the programme is fit to model, a session to agree the ranges and the risks, the analysis itself and a short report giving the P50 and P80 dates, the risks that drive them and the time risk allowances they support. Every fee is fixed and agreed in writing before any chargeable work. Where the programme needs repair first, a programme review starts from £1,250 plus VAT.




What the result is for


At tender, the result sets three things. The Completion Date the business offers or accepts, the time-related cost the price carries and the time risk allowances the programme shows. Each of them can then be explained at clarification from the record behind it.


In delivery, the value is in the drivers rather than the date. The risks that move the P80 most should already be early warnings, since clause 15.1 requires one as soon as the Contractor or the Project Manager becomes aware of a matter that could delay Completion. Each needs a place on the Early Warning Register and a named owner for the mitigation. Crossrail ran a schedule review with senior directors every four weeks, used tornado charts to show the key risk drivers and reported a summary of the analysis to its sponsors every six months. The Government Accountability Office's guide makes the same point in its own terms: the results are best treated as inputs to managing the project rather than as forecasts.


Scale matters less than habit. Crossrail's Level 1 schedule typically held 4,000 to 5,000 current and future activities, against the 200 to 1,000 the same report gives as typical of analyses on other major projects. A contractor's programme for a single package is far smaller and the analysis is correspondingly quicker. What it needs is a programme fit to model and someone who will act on what it shows.



The view from the desk


My position is that a contractor should run its own schedule risk analysis before it offers or accepts a Completion Date, not after the client asks for a P80. Left until the client asks, the analysis tends to arrive after the price is set or the date is signed, when it can only report on a commitment already made. Run before the bid goes in, the same work decides the commitment.


When a client asks for a P80, I would send the curve and its drivers, not the date alone. A date on its own gets copied into reports and loses its assumptions. Months later it comes back as the date the contractor promised. The curve keeps the conditions attached. It also turns the talk from how confident the contractor is to what each party can do about the drivers. Some of those will sit with the Client.



Download the schedule risk analysis checklist


A two-page check to run before a P80 date leaves the business: whether the programme is fit to model, the ranges and the risks, the split between the Contractor's and the Client's risks, the result and what it commits you to under the NEC4 ECC. Direct download, no sign-up.


Preview of the two-page Schedule Risk Analysis Checklist: whether the programme is fit to model, ranges and risks, splitting the risks by owner under the NEC4 ECC, reading the result and the check before a P80 date goes to the client.



Schedule risk analysis: frequently asked questions


A schedule risk analysis, or QSRA, puts realistic ranges on the durations in a programme and adds the main risks, each with a chance of happening and an effect. Software then runs the programme many times in a Monte Carlo simulation and records the finish date each time. The result is a spread of possible finish dates, usually shown as an S-curve. It also shows the risks that drive it.

The P80 date is the date by which 80 per cent of the simulated runs finished, leaving one chance in five of finishing later. The P50 is the date half the runs reached. Both describe the model and the risks put into it, so they are best read as measures of confidence rather than forecasts. Neither becomes a contract date unless the business offers it as one.

No. The NEC4 ECC does not require one. Clause 31.2 requires each programme submitted for acceptance to show time risk allowances. A schedule risk analysis is a good way to size them. A client that wants an analysis can ask for it in the tender documents or the Scope. Where the Scope requires it with each programme, it becomes part of what the Project Manager checks under clause 31.3.

The time risk allowance is the Contractor's own provision for the risks it carries, shown on the programme under clause 31.2. A schedule risk analysis measures the spread those risks create. Run on the Contractor's risks alone, it shows where the allowances belong and how large they should be. Risks the contract treats as compensation events stay out of the allowances, because the contract deals with them separately. For weather and physical conditions, that is only the part beyond what the Contractor carries.

Neither is a contract date unless the business offers it as one. Under the NEC4 ECC the commitment is the Completion Date and any Key Dates. Which confidence to work to is a commercial decision: a later date is safer but carries more time-related cost, an earlier one can make a bid more competitive but carries more chance of delay damages where Option X7 applies. The analysis puts a number on that choice before it is made.



About the author


Roman Bazelchuk, founder of NEC Planning Solutions

Roman Bazelchuk is the Founder of NEC Planning Solutions Ltd, a UK project planning and controls consultancy supporting contractors with NEC programme compliance, compensation event assessments and live project controls. He is an NEC Accredited Project Manager and holds the APMG Project Planning and Control qualification, with a BEng in Mechanical Engineering and postgraduate training in Planning and Control.


NEC Planning Solutions provides contract-aware planning support through a QA-governed delivery model, helping project teams keep programmes accepted, current and commercially useful from tender through to live delivery.




Asked for a P80 date and not sure what it commits you to?


If a client has asked for a schedule risk analysis, if the programme behind a tender has never been tested against its risks, or if the time risk allowances on an NEC programme are a percentage nobody can defend, NEC Planning Solutions checks the programme, runs the analysis on the native P6 file and turns the result into allowances and a date the business can stand behind. Where the programme needs work first, programme review and recovery is the place to start.


You don't need to send anything first. We reply to every enquiry within one working day. The scope and a fixed fee follow in writing within 48 hours of the first call.



bottom of page