Negative Float, Zero Float, and the Wrong Activity Being Frozen

Negative total float = the amount by which the imposed finish date is earlier than the logic-driven finish date. It is not a rounding artifact and it is not an error in the calculation; it is the schedule telling you the target is infeasible under current logic.

Mechanics: the backward pass starts at the imposed finish (day 12) instead of the computed one (day 14). Every activity on the driving path inherits a late finish two days earlier than its early finish, so LF − EF = −2 for A, C, and D.

• A: ES 0, EF 5; LF 3, LS −2 → TF −2
• C: ES 5, EF 11; LF 9, LS 3 → TF −2
• D: ES 11, EF 14; LF 12, LS 9 → TF −2

The critical path is still A → C → D (14 days). What has changed is that the network now needs two days of compression or a two-day scope reduction on that path, because the float is not 'used up' - it is overdrawn.

Interpretation rule: negative float is a conflict flag, and the correct response is to resolve the conflict (compress, descope, or renegotiate the date) - not to hunt for the activity with the worst float number and blame it.

B is not critical, and freezing it is the wrong call.

Compute B properly:

• ES = 5 (after A), EF = 9
• LF = 9 (B's only successor is D, whose LS = 9), LS = 5
• Total float = LS − ES = 5 − 5 = 0
• Free float = ES(D) − EF(B) = 11 − 9 = 2

Why B shows zero total float without being critical: B's path is 12 days long, and the imposed finish is also day 12. B has no total float left against the contractual date - any slippage on B pushes past day 12 - but it has 2 days of free float, meaning it can slip 2 days without disturbing any successor. B is near-critical under the imposed date, not driving the unconstrained network.

Why the team lead's instruction is wrong on both counts:

1. It spends the budget where compression buys nothing for the 14-day logic. B sits on the 12-day path. Crashing B shortens a path that is already shorter than the critical path; the project finish does not move at all until the other path is dealt with.
2. It targets the low-uncertainty activity. B is well-understood and confident at 4 days. Buffers and attention belong on the high-variance, driving activities.

B's real status is a watch item: protect its 2 days of free float, but do not staff it.

Target the path with negative float: A → C → D.

1. Compress C or A. C (6 days) and A (5 days) are the two candidates. Crashing either by 2 days closes the gap; crashing B by 2 days changes nothing. Compare crash cost per day and risk on A versus C - the cheaper, lower-risk one wins.
2. Fast-track where logic permits. If A's back half and C's front half do not genuinely conflict, an SS with lag can recover days at near-zero cost. Verify the overlap is physically possible, not just diagrammatically possible.
3. Descope or renegotiate. If compression would cost more than the contractual penalty (or push risk past tolerance), the honest options are reducing scope on C or D, or returning to the customer with the negative-float evidence: the date is infeasible against the current logic, and here is what closing it costs.
4. Do not leave the negative float in the baseline. A submitted schedule showing −2 float on the driving path is an unmanaged conflict; resolve it or escalate it, but get it to zero or to a formally accepted variance.

What would change the answer:

• If the customer's day-12 date were a soft target rather than contractual, the plan could simply carry 2 days of negative float as an accepted risk and monitor it.
• If B's 4-day duration were actually uncertain (e.g., a new team), B's zero total float would move it into the risk conversation - but as a threat to the imposed date, not as the path to compress.
• If resource leveling were applied later, the critical path could shift entirely, so the compression decision should be re-validated after leveling, not before.