The DCMA 14-point assessment, check by check

The DCMA 14-point assessment is the most widely used quick test of a CPM schedule's construction. It was written by the US Defense Contract Management Agency to screen contractor schedules, and it has spread well beyond defence work because it is short, objective and easy to repeat every month.

It is a screen, not a verdict. A schedule can pass all fourteen checks and still be wrong about the work; a schedule that fails several is almost certainly unreliable as a tool for forecasting or for proving delay. The value is in knowing what each check is looking for.

The fourteen checks

Thresholds below are the ones conventionally applied. Most are measured against incomplete activities only, because completed work can no longer be fixed.

1. Logic. Incomplete activities with no predecessor or no successor. Threshold: no more than 5%. Open ends let dates float free of the network, so delay in one place never reaches the finish date.

2. Leads. Relationships with negative lag. Threshold: none. A lead says a successor starts before its predecessor finishes by a fixed amount, which usually hides a missing activity.

3. Lags. Relationships with positive lag. Threshold: no more than 5%. Lags are durations without an activity to own them; a long lag can absorb or hide delay.

4. Relationship types. Share of finish-to-start relationships. Threshold: at least 90%. Start-to-start and finish-to-finish links are legitimate, but heavy use makes the critical path hard to follow and easy to manipulate.

5. Hard constraints. Constraints that fix a date regardless of logic, such as must-start-on or must-finish-on. Threshold: no more than 5%. A hard constraint overrides the network, so the schedule stops showing the effect of delay.

6. High float. Incomplete activities with more than 44 working days of total float. Threshold: no more than 5%. Large float usually means missing successors rather than genuine flexibility.

7. Negative float. Incomplete activities with total float below zero. Threshold: none. Negative float means the schedule already cannot meet a date it is constrained to; it needs a recovery plan or a change to the date.

8. High duration. Incomplete activities longer than 44 working days. Threshold: no more than 5%. Long activities are hard to status and hide where progress is really being lost.

9. Invalid dates. Forecast dates before the data date, or actual dates after it. Threshold: none. Either way, the update has not been statused correctly.

10. Resources. Activities with no resources or cost loaded. Usually reported for information rather than against a threshold, where the contract requires a loaded schedule.

11. Missed tasks. Activities that the baseline said would be finished by the data date and are not. Threshold: no more than 5%. A rising number is an early signal of slippage the finish date does not show yet.

12. Critical path test. Add a large delay to an activity on the critical path and check that the project finish moves by the same amount. If it does not, the critical path is broken by a constraint, a lag or missing logic.

13. Critical path length index (CPLI). The remaining critical path length plus the total float to the finish, divided by the remaining critical path length. Threshold: at least 0.95. Below 1.0 the project is behind its required finish; well below, it is unlikely to recover.

14. Baseline execution index (BEI). Activities actually completed, divided by activities the baseline said would be completed by now. Threshold: at least 0.95. It measures whether the work is being done at the rate the plan assumed.

Reading the results together

The checks interact. Missing logic (1) produces high float (6) and breaks the critical path test (12). Hard constraints (5) produce negative float (7). A schedule that fails the first group of checks will also misreport the last two, because CPLI and BEI depend on a critical path and a baseline that can be trusted.

That is why the order matters when fixing a schedule: repair logic and constraints first, then read float, then the indices.

Where Schedule Cracker fits

Schedule Cracker runs the fourteen checks on any update you upload, alongside the rest of its rule set, and each result shows the counts behind it. Running them on each update in turn shows how the schedule is trending, which is usually more telling than any single month.

Questions

Is the DCMA 14-point assessment a contract requirement? Only where the contract or the owner's specification says so. Many owners use it, or a variant of it, as an acceptance check for baseline and update submissions.

Why 44 working days? It is roughly two months of working time, the conventional limit for an activity or a float value before it is treated as too coarse to manage.

Does passing all fourteen checks make a schedule correct? No. The checks test how the schedule is built, not whether its durations and logic describe the real work. They are a floor, not a guarantee.