Of all the methods in the forensic schedule analyst's toolkit, Collapsed As-Built (CAB) is the one most likely to generate heated arguments between experts. The method asks a deceptively simple question: what would have happened if this specific delay had not occurred? To answer it, the analyst starts with the as-built schedule of what actually happened, identifies the delays attributable to a specific party, and removes (collapses) them to produce a "but-for" schedule showing the project as it would have unfolded without those delays. The difference between the as-built completion date and the but-for completion date is the measured delay.
It is intuitive. It produces a clear number. And it has been heavily criticized by practitioners, courts, and AACE itself for the judgment-intensive steps buried inside that apparent simplicity. Understanding when CAB is appropriate, and when it is not, requires looking at how AACE RP 29R-03 treats it and where its weak points actually sit.
A Quick Orientation Within AACE RP 29R-03
CAB is a modeled, subtractive method. The RP classifies it under two MIPs depending on how it is executed:
- MIP 3.8 (Modeled / Subtractive / Single Simulation) performs a single collapse to produce one but-for schedule.
- MIP 3.9 (Modeled / Subtractive / Multiple Simulation) performs multiple collapses, typically one for each distinct delay event or party.
Both sit at the modeled end of the taxonomy, meaning the analyst is constructing a hypothetical schedule rather than observing contemporaneous data. That construction step is where the method's controversies originate.
What the Method Does
The central logic is subtraction. Start with what actually happened (the as-built schedule), identify the delay events to be collapsed (removed), remove them, recalculate the CPM network, and compare the resulting but-for completion date to the actual completion date.
Some variations:
- Single-simulation CAB (MIP 3.8) collapses all relevant delays attributable to one party in a single pass.
- Multiple-simulation CAB (MIP 3.9) runs separate collapses for different delay categories (owner-caused, contractor-caused, excusable non-compensable, force majeure), often to produce a more nuanced allocation.
- But-for analysis for the contractor's claim collapses only owner-caused delays and reports the difference as compensable delay.
- But-for analysis for the owner's defense collapses only contractor-caused delays to show the contractor would not have finished on time regardless.
The Procedure
The mechanics look simple but each step carries analytical weight:
- Develop a validated as-built schedule. This is the first and often largest challenge. True as-built data, with accurate logic ties reflecting how activities actually interacted, is rare on most projects. The analyst typically has to reconstruct it from daily reports, updates, correspondence, photographs, and interviews.
- Identify the delay events to be collapsed. Classify each by responsible party, cause, and timing.
- Model each delay event in the as-built schedule as an activity or series of activities with logic ties.
- Remove the delay events. This is the collapse step. In practice, it means deleting the modeled delay activities and adjusting activity durations and logic to reflect what would have happened without them.
- Recalculate the CPM network.
- Compare the collapsed (but-for) completion date to the as-built completion date. The difference is the measured delay attributable to the collapsed events.
When to Use It
CAB is sometimes the best available method when:
- Contemporaneous schedule updates are absent, unreliable, or manipulated to the point that observational methods (MIPs 3.3, 3.4, 3.5) cannot be trusted.
- The as-built record is relatively good, with reliable daily reports, documented events, and clear causation.
- The dispute focuses on a discrete set of clearly identifiable delays that can be isolated and modeled.
- A but-for analysis is explicitly required by contract or tribunal.
In other words, CAB is often the method of last resort when the contemporaneous record is weak but the as-built record is strong. That scenario exists, but it is less common than the volume of CAB reports in claims practice would suggest.
The Controversies
The criticisms of CAB are serious and worth understanding.
- As-built schedule reliability. A CAB result can only be as good as the as-built schedule it is built on. In practice, as-built schedules are reconstructions, and reconstructions involve judgment. Two competent analysts can produce materially different as-built schedules from the same project records.
- Logic reconstruction. The analyst has to assign logic ties to activities based on what actually happened. These ties are rarely documented contemporaneously and often have to be inferred. Different logic choices produce different but-for results.
- Delay identification and attribution. Deciding which events count as delays and who caused them is already contested in most disputes. CAB bakes these contested decisions into the analytical model.
- Collapse judgment. Removing a delay is not just deleting an activity. The analyst has to decide what the project would have done instead during that period. That counterfactual is speculative.
- Concurrency handling. Multiple-simulation CAB can address concurrency in some respects, but the treatment varies and is often contested.
- Manipulation exposure. Because each step involves judgment, CAB is more vulnerable than observational methods to consciously or unconsciously producing a preferred result.
Courts and tribunals have expressed skepticism about CAB in various jurisdictions, particularly when contemporaneous alternatives exist. Some have explicitly rejected or given limited weight to CAB results when the analyst could have used a contemporaneous method instead.
Caveats Worth Respecting
- Prefer contemporaneous methods when available. If the project has reliable contemporaneous updates, a windows or Half-Step approach is almost always more defensible than CAB.
- Document every judgment call. Every logic choice, duration estimate, and collapse decision should be explicitly stated and supported. A CAB report that does not expose its assumptions invites cross-examination.
- Use multiple simulations where possible. A single collapse asking "what if none of the owner delays happened?" is less defensible than a set of simulations testing different configurations.
- Validate against the as-built record. The but-for schedule should be plausible in light of what actually happened. If the but-for says the project would have finished three years early, the logic or durations are probably wrong.
- Be honest about the method's limits. Experts who overstate CAB's precision lose credibility. Those who acknowledge the method's judgment-intensive nature while defending their specific choices tend to do better.
Where It Sits in the Bigger Picture
CAB has a legitimate place in the forensic toolkit, but it sits at the end of the bench, not the front. When contemporaneous data is reliable, observational methods produce more defensible results. When it is not, CAB can fill the gap, but the analyst has to earn the result through rigorous documentation of every judgment call.
The practical takeaway is that CAB should be the method chosen because nothing better is available, not the method chosen because it produces the cleanest number. A but-for schedule is only as good as the analyst's honesty about what they had to assume to build it.
References
- AACE International, Recommended Practice No. 29R-03, Forensic Schedule Analysis (current edition). See MIP 3.8 and MIP 3.9, and the RP's discussion of modeled subtractive methods.
- AACE International, Professional Practice Guide No. 20 (PPG #20), Forensic Schedule Analysis (2nd Edition). Collected papers on collapsed as-built analysis, including critiques and defenses of the method.
- Zack, J.G. Jr. "But-For Schedules - Analysis and Defense," AACE Transactions.
- Society of Construction Law, Delay and Disruption Protocol (2nd Edition, 2017). Addresses as-built collapsed analysis within its broader methodology discussion.
- Popescu, A.I.; Avalon, A. "Retrospective As-Built Schedule Development," AACE Transactions.