If you've ever stared at two consecutive schedule updates and tried to answer the question "why did the finish date move?" only to find that the update contains actual progress and a pile of logic changes, duration tweaks, and resequenced work - you already understand the problem the Half-Step was designed to solve.
Delay analysts call this the "mixed update" problem. Progress and plan revisions get baked into the same file, and once they're baked together, teasing them apart by eyeballing the schedule is painful, slow, and rarely defensible in a claim setting. The Half-Step technique, formalized under AACE International's Recommended Practice 29R-03, Forensic Schedule Analysis, gives analysts a structured way to pull those two threads apart.
A Quick Orientation to AACE RP 29R-03
AACE RP 29R-03 is the industry's taxonomic reference for forensic schedule analysis. Rather than endorsing a single "best" method, it catalogs nine Method Implementation Protocols (MIPs) grouped by whether they are observational or modeled, retrospective or prospective, and static or dynamic. The RP is explicit that it is a reference framework, not a prescriptive standard of practice — the analyst still has to select the method that fits the project, the contract, and the dispute.
The Half-Step lives inside MIP 3.4 - Observational / Dynamic / Contemporaneous Split (Bifurcated). Its close cousin is MIP 3.3, the straight "Contemporaneous As-Is" period analysis, which takes updates exactly as submitted and measures variance between them. MIP 3.4 layers an extra analytical step on top of 3.3, and that extra step is the Half-Step.
What the Half-Step Actually Does
In plain terms, the Half-Step is a bifurcation procedure. It splits a single schedule update into two conceptual pieces:
- The progress component - actual work that occurred between data dates: activities started, percent complete changes, actual finishes, and the resulting recalculation of the critical path.
- The revision component - everything else the scheduler changed: added or deleted activities, modified relationships, shortened or lengthened remaining durations, new constraints, calendar changes, and resequencing of remaining work.
By isolating the first from the second, the analyst can finally answer the forensic question that a raw update-to-update comparison cannot: Did the finish date slip because the work went slowly, or because the plan was rewritten?
That distinction matters enormously in claims. A project can appear to "recover" lost time in a monthly update not because any work sped up, but because the scheduler compressed downstream durations or changed logic. Without the Half-Step, that kind of paper recovery is easy to miss.
The Procedure, Step by Step
The mechanics are straightforward, even if the bookkeeping is tedious by hand:
- Start with the Beginning-of-Period (BOP) schedule - i.e., the previous accepted update.
- Make a working copy of the BOP schedule.
- Apply only the progress data from the End-of-Period (EOP) update into that copy: actual starts, actual finishes, remaining duration changes attributable to progress, percent complete. Leave logic, non-progress durations, added/deleted activities, and constraints untouched.
- Recalculate the copy using the EOP data date. This is the Half-Step schedule (sometimes written Schedule A½ when comparing Schedule A to Schedule B).
- Compare three schedules: BOP vs. Half-Step shows the impact of pure progress. Half-Step vs. EOP shows the impact of non-progress revisions. BOP vs. EOP (the traditional MIP 3.3 comparison) is now decomposable into those two discrete causes.
AACE RP 29R-03 notes that MIP 3.3 and MIP 3.4 are not mutually exclusive; the Half-Step can be applied selectively to windows within a larger 3.3 analysis, particularly where scope changes or means-and-methods shifts are suspected delay drivers.
Why Analysts Reach for It
A few situations make MIP 3.4 especially useful:
- Schedules with frequent revisions. Projects where the scheduler routinely reshapes logic or compresses remaining durations between updates are prime candidates. A straight 3.3 comparison in that environment hides as much as it reveals.
- Disputes over mitigation and acceleration. If a contractor claims to have mitigated delay, the Half-Step can show whether the "mitigation" was real performance or simply optimistic re-planning.
- Identifying schedule manipulation. The method doesn't presume bad faith, but it does make revision-driven movement visible. That alone changes negotiation dynamics.
- Supporting causation. Because the method ties movement of the critical path to a specific category of change (progress vs. revision), it strengthens the causation narrative that any forensic opinion has to carry.
Caveats Worth Respecting
The Half-Step is not a silver bullet, and RP 29R-03 is candid about its limitations. Two come up repeatedly:
- Update quality drives result quality. The technique depends on the contemporaneous updates being reasonably accurate in both reported progress and in underlying means and methods. Garbage in, garbage out.
- Heavy use of date constraints makes life hard. When schedulers have liberally applied constraints, extracting meaningful forward-pass behavior from the Half-Step becomes difficult and sometimes unreliable.
There's also a practical concern: doing this by hand across two or three years of monthly updates is grueling. Each period requires identifying every change, classifying it as progress or revision, applying only the progress subset, and recalculating. It's why purpose-built tools (Steelray Delay Analyzer and Plannex are two examples in the market) have emerged to automate the mechanical work and let the analyst focus on judgment calls - critical path behavior, delay driver attribution, and responsibility.
Where It Sits in the Bigger Picture
The Half-Step isn't a replacement for Time Impact Analysis, Windows Analysis, or Collapsed As-Built methods. It's a refinement of contemporaneous period analysis; a way to make observational retrospective methods more surgical. When the dispute hinges on why the critical path moved rather than simply that it moved, MIP 3.4 earns its keep.
For practitioners, the takeaway is this: if your analysis is producing numbers that don't square with the story the project records tell, and your schedule updates contain substantial non-progress revisions, you may be looking at a case where the Half-Step is the right tool. And if you're writing an expert report, citing MIP 3.4 of RP 29R-03 explicitly rather than describing the method generically, gives the opinion the grounding that triers of fact increasingly expect.
References
- AACE International, Recommended Practice No. 29R-03, Forensic Schedule Analysis (2011, as revised). See in particular MIP 3.3 and MIP 3.4 in Section 3, and the taxonomy discussion in Section 1.
- AACE International, Professional Practice Guide No. 20 (PPG #20), Forensic Schedule Analysis (2nd Edition). Collected papers on RP 29R-03 methodology, including:
- Freas, R.M.; Nagata, M.F.; Carson, C.W. "Using the Half-Step Technique in Retrospectively Analyzing Schedule Delay," Cost Engineering Journal, 2018.
- Nagata, M.F.; et al. "Making Changes to the Schedule When Performing MIPs 3.3 and 3.4."
- Gonzales, A.; Schuldes, J.; et al. "Contemporaneous 'As-Is' Period Schedule Delay Methodology — Forensic Delay Methodology, MIP 3.3," AACE Transactions, 2018.
- Livengood, J.C. "The New AACE Recommended Practice on Forensic Schedule Analysis — Choosing the Right Methodology," Construction Claims Advisor, 2008.
- Lowe, J.S.; Nagata, M.F. "A Synthesis of Forensic Schedule Analysis Techniques," AACE Transactions, 2008.