Schedule analysis: DCMA health, Monte Carlo risk, earned schedule and compare
Grade a saved schedule against the DCMA 14-point checks, simulate the finish with Monte Carlo, measure time performance with Earned Schedule and compare any version with a baseline.
On this page
The Schedule analysis workspace asks four questions of a saved plan. Is the logic sound? How likely is the finish? Is the work keeping time against the baseline? What changed since the last version? Each answer comes from the same P6-validated engine that calculates the plan’s dates. Each result is saved with the version it describes, so you can come back to it in a month and see exactly what it said.
Open it from the Schedule tool’s Analysis button, or from /companies/{company_id}/projects/{project_id}/tools/schedule/analysis. It has four tabs: Health (DCMA 14), Risk (Monte Carlo), Earned schedule and Compare.
- Who
- Schedulers, project controls and PMs with Schedule admin
- Runs on
- The saved head version of the Planner schedule
- Produces
- Immutable, retained results with engine revision and inputs
- Connects to
- PlannerBaselinesInsights CPM pack
Run an analysis
- Save a plan in the Planner. Analyses run only on immutable saved versions, never on an unsaved draft. The page header shows the version it will use and the saved data date.
- Pick a tab and set its options. Health has none. Risk takes iterations, a distribution and a seed. Earned and Compare need a baseline or an earlier saved version. Health, Risk and Earned accept an optional data date; leave it blank to use the saved one.
- Run it. The result is calculated by the engine and retained with the saved version. The next time anyone opens the tab, the latest result for the current version is already there, with its calculation provenance: engine revision, parameters and when it ran.
- Drill in. Offenders, drivers and changed activities open in the Planner. Health offenders and Risk percentiles export to CSV.
Health: the DCMA 14-point assessment
The Health tab runs the fourteen DCMA checks and gives each one a status: pass, fail or skipped. It shows the measured value, the threshold applied, where that threshold comes from, the counts behind the value and up to 20 offenders. Offenders are activities, or relationships written as A1040->A1050 (SS). Filter by failed or skipped checks, open an offender in the Planner, or export every offender to CSV.
| # | Check | What it looks for |
|---|---|---|
| 1 | Logic | Activities missing a predecessor or successor (5% threshold) |
| 2 | Leads | Relationships with negative lag (0%) |
| 3 | Lags | Relationships carrying lag (5%) |
| 4 | Relationship types | At least 90% finish-to-start |
| 5 | Hard constraints | Constraints that override logic (5%) |
| 6 | High float | Total float above 44 working days (5%) |
| 7 | Negative float | Any activity with negative float (0%) |
| 8 | High duration | Durations above 44 working days (5%) |
| 9 | Invalid dates | Forecasts before the data date, actuals after it |
| 10 | Resources | Activities without resources (runs only when assignments exist) |
| 11 | Missed tasks | Baseline finishes missed by the data date (5%) |
| 12 | Critical path test | Whether a delay on the critical path moves the finish by the same amount |
| 13 | CPLI | Critical Path Length Index (0.95) |
| 14 | BEI | Baseline Execution Index (0.95) |
The grade is A at a score of 90 or more, B at 80, C at 70, D at 60, otherwise F. The score is the share of checks that could run and passed. Skipped checks are left out of the denominator, never counted as passes. A schedule with no relationships doesn’t get an A for having no bad relationships: checks 2 to 4 are skipped and check 1 fails.
Risk: Monte Carlo simulation of the finish
The Risk tab samples activity durations through the saved network’s logic, calendars and constraints, from 100 to 20,000 times. It returns:
- Finish-date percentiles: P10, P20, P50, P80, P90 and P95, plus the deterministic CPM finish and the probability of meeting it.
- An S-curve of cumulative probability from P5 to P95.
- Sensitivity drivers: the top activities, each with a criticality index (the share of iterations in which it was critical) and the correlation between its duration and the finish. The tornado chart sorts them by influence.
Choose PERT (beta) or triangular sampling. The seed defaults to 12345, so the same inputs reproduce the same answer. Leave the seed blank and one is generated and recorded, so you can rerun it exactly. Large runs go to a background job with progress and cancel. Completed activities stay at their recorded duration. Started activities vary only their remaining duration.
Earned schedule: SPI(t) against a baseline
Earned Schedule measures time performance in time units, not dollars. Pick the baseline (or an earlier saved version) to measure against, and the engine returns, in working days:
- PD, the baseline’s planned duration.
- AT, the actual time from the baseline start to the data date.
- ES, the point in the baseline when the work done to date was planned to be done.
The engine also returns SPI(t) = ES ÷ AT (below 1.00 is behind), SV(t) = ES − AT and the forecast duration IEAC(t) = PD ÷ SPI(t), which assumes current performance continues. The workspace shows them with duration-weighted percent complete, the remaining planned work and the forecast slip. A result stored before the engine published these indices has them worked out from ES and AT on the page, so old and new results read the same. See SPI(t).
Compare: what changed and why the finish moved
Compare puts the current saved version beside a baseline or any earlier saved version. Both are calculated on their own stored data dates. You get:
- Finish movement in calendar days and the engine’s plain-language explanation of why the finish moved.
- Criticality changes: activities that became critical and activities that left the critical path.
- Activity date and float deltas for every matched activity, with its change kinds: added, removed, renamed, duration changed, date shifted, relogicked, constraint changed, calendar changed, progress added or criticality changed. Filter to changed, added, removed, slipped, float lost or criticality.
- Relationships added and removed, from the authored version review.
Activities are matched by id, by code, or by name and WBS. The result says which key it used and lists any matches it couldn’t make one-to-one.
Mistake: Publishing a letter grade without its basis
In the app: Every health result carries the screening-indicator statement, and the workspace shows it beside the grade.
Mistake: Reading P50 as the most likely finish of a padded plan
In the app: Risk shows the deterministic finish and the probability of meeting it next to every percentile.
Mistake: Comparing two exports with different data dates as if they were the same update
In the app: Compare calculates each version on its own stored data date and reports the match key it used.
Mistake: Counting a skipped check as a pass
In the app: Skipped checks are excluded from the score and listed separately with the missing input.
Before you send an analysis to an owner or reviewer
- The analysis ran on the version you’re submitting (check the version badge and provenance).
- The data date is the status cutoff you mean, not a later export date.
- Health is reported with its basis statement and the skipped checks.
- Risk states the iterations, distribution, seed and estimates source, with digests of the estimates and the risk model.
- Compare names the baseline it measured against.
Use it through the API
Every analysis is one call: POST /rest/v1.0/projects/{project_id}/schedule/calculations with operation set to health, risk, earned or compare. Send expected_version_id (the current head) and, for earned and compare, baseline_version_id. The Schedule API reference documents each result shape with an example, and the Schedule API guide covers permissions and version checks.
{ "expected_version_id": 42, "operation": "risk",
"risk": { "iterations": 5000, "distribution": "pert", "seed": 12345 } }
Related
- Planning and CPM: build and calculate the plan you analyse.
- Baselines and versions: name the baseline you measure against.
- CPM analytics in Insights: DCMA score, float and finish variance across every project.
- Glossary: DCMA 14-point assessment, SPI(t), Monte Carlo P80, float.
Keep exploring
Related across os.construction
- DocsScheduling: CPM planning, analysis and evidence in one workspaceA native critical path scheduling workspace with a P6-validated engine, P6-style layouts and views, Line of Balance and takt planning, Last Planner weekly work plans, activity discussion and a change feed, P6 import options, DCMA health, three-point risk, a reviewed update cycle, portfolio programs, reports, baselines, versions, resources, costs and period performance.
- GlossaryMonte Carlo P80The finish date that 80% of simulated schedule outcomes meet or beat in a Monte Carlo schedule risk analysis.
- GlossarySchedule BaselineA preserved reference plan used to compare later progress and forecast dates against the agreed planning basis.
- DocsP6 compatibility: settings, import, export and what round-tripsHow closely os.construction matches Primavera P6. Which scheduling settings the engine honours, which P6 calculated fields it works out, what crosses an XER or P6 XML file in both directions, the measured agreement with P6's own dates and costs, and the limits that remain.
- BlogHow to calculate liquidated damages exposure and carry it in your forecastLiquidated damages exposure is the days between your forecast finish and the adjusted contract date, times the daily rate, for each milestone that carries LDs, limited by any cap. The hard part is using honest inputs: approved extensions only, a forecast from the current schedule, pending requests and risk shown as separate scenarios. Here is the method and a worked example with two milestones.
- BlogEarned value vs percent complete: one measures progress, the other tells you what it meansPercent complete says how much of the work is done. Earned value turns that progress into the budgeted value of the work done, so it can be compared with the plan and with what was spent. Here is how the two relate, when to use each, a worked example and the mistakes that make every job look on budget.
Was this helpful?
Run it on your own project.
Start a 14-day free trial. Want a say in what ships next? Join the founding program.
14-day free trial. Pricing for paid plans isn't set yet. Talk to us. Already have an account? Sign in →