29R-03Advanced22 min read

Forensic Schedule Analysis

When a project finishes late and the parties disagree about why, forensic schedule analysis reconstructs the schedule history to apportion responsibility. Learn the two axes, the nine MIP methods, the workhorses, and the hardest issue — concurrency. Built on AACE 29R-03.

What forensic schedule analysis is

When a project finishes late and the parties disagree about why, someone must reconstruct the history of the schedule to apportion responsibility. That is forensic schedule analysis — done after the fact, for a claim or a dispute.

Two axes organise every method

RP 29R-03 sorts the nine methods along two fundamental questions. Get these two ideas and the whole taxonomy falls into place.

AxisObservationalModelled
What it doesReads delay directly from the schedules as they existRuns a "what-if" simulation by changing the schedule
Question answeredWhat do the contemporaneous schedules show happened?What would have happened but for the delays?
Data demandGood baseline + regular updates / as-builtA reliable CPM you can defensibly manipulate

Within the modelled family there's a second split: additive methods insert delays into an as-planned schedule (the "impacted as-planned" idea, the same fragnet logic as a TIA), while subtractive methods remove delays from an as-built schedule to see what would have finished without them (the "collapsed as-built," or "but-for," idea).

The nine MIP methods

RP 29R-03 labels each method as a "Method Implementation Protocol" (MIP). You don't need to memorise the numbers — but you should recognise the families and the two or three workhorse methods within them.

MIPFamilyMethod
3.1Observational · Static · GrossAs-Planned vs As-Built
3.2Observational · Static · PeriodicAs-Planned vs As-Built, period subdivided
3.3Observational · Dynamic · Contemporaneous As-IsWindows (contemporaneous period) analysis
3.4Observational · Dynamic · Contemporaneous SplitWindows with split updates
3.5Observational · Dynamic · Modified/RecreatedWindows on recreated updates
3.6Modelled · Additive · Single BaseImpacted As-Planned (single)
3.7Modelled · Additive · Multiple BaseTime Impact Analysis (stepped)
3.8Modelled · Subtractive · Single SimulationCollapsed As-Built ("but-for")
3.9Modelled · Subtractive · Multiple BaseCollapsed As-Built, stepped

Method explorer

Tap a method to see what it does, where it sits in the taxonomy, and what it needs.

Method explorer

Explore

Tap a method to see what it does, where it sits in the taxonomy, and what it needs.

Observational · Static · MIP 3.1

As-Planned vs As-Built

Overlay the original plan against what actually happened and measure the gap. Simple and intuitive, but it shows that the project slipped, not which delay on which path caused it — and it ignores how the critical path shifted over time. A good first picture; weak on causation.

Concurrency — when both sides are late at once

The single thorniest question in forensic analysis is concurrency: two or more delays — one the owner's, one the contractor's — operating in the same period, each independently capable of delaying completion. Concurrency typically means the contractor gets an extension of time but not prolongation money for the concurrent period: time relief, because the owner-caused delay was real; but no compensation, because the contractor would have been late anyway.

Owner delay (late design)Contractor delay (resequence)concurrent windowEOT granted · prolongation money usually not
Concurrent delay in one window: an owner-caused and a contractor-caused delay overlap. The usual outcome is an extension of time but not prolongation money for the shared period.

The records pick the method

As with productivity claims, you don't choose a forensic method by preference — your records choose it for you. Reliable, regularly-updated contemporaneous schedules unlock the credible dynamic methods like windows analysis. Where updates are missing or manipulated, you're pushed toward weaker modelled methods that depend on assumptions the other side will attack. The integrity of the schedule record, built during the job, is what determines how strong your forensic analysis can be years later.

Ten things to remember

  1. Forensic schedule analysis is retrospective — it reconstructs why a finished project ran late.
  2. RP 29R-03 catalogues nine methods (MIPs) so analysts share one language.
  3. Observational reads the schedules; modelled runs a what-if.
  4. Modelled splits into additive (insert delays) and subtractive (remove delays).
  5. Windows analysis (MIP 3.3) is widely seen as most reliable when good updates exist.
  6. As-planned vs as-built shows slippage, not causation — a first picture, not a proof.
  7. Concurrency usually means EOT but not prolongation money for the shared period.
  8. Method choice changes the answer — and whether concurrency is even visible.
  9. Your records pick the method — schedule integrity built during the job decides what's possible.
  10. State the MIP, show inputs, make it repeatable — transparency beats a bigger unverifiable number.

Glossary

Collapsed As-Built
A subtractive ("but-for") method: remove delays from the as-built to see what would have finished.
Concurrency
Independent owner and contractor delays in the same period; usually time relief without prolongation money.
Forensic schedule analysis
Retrospective study of a project's schedules to determine the cause and responsibility of delay.
Impacted As-Planned
An additive modelled method: insert delays into the baseline to project their effect.
MIP
Method Implementation Protocol — RP 29R-03's label for each of the nine recognised methods.
Modelled method
Runs a what-if simulation by adding or removing delays from a CPM.
Observational method
Reads delay directly from the contemporaneous schedules without modelling.
Windows analysis
A dynamic observational method that tracks the critical path period by period.

Check your understanding

1How does forensic schedule analysis differ from a prospective TIA?
2The two primary axes of RP 29R-03 are:
3Which method is widely regarded as most reliable when good contemporaneous updates exist?
4Concurrent delay (owner and contractor in the same period) usually results in:
5"Your ___ pick the method" — RP 29R-03's point about choosing an analysis: