Example 1: 10 days at $4,000, crash to 7 days at $5,500
Slope = (5500 − 4000) ÷ (10 − 7) = 1500 ÷ 3 = $500 per day. These are the mini-calculator defaults. Each day you buy on this task costs $500, up to three days.
How to price one day of schedule compression and which critical-path task to crash first.
Crash time is the shortest duration a task can take with extra money. The cost slope is (crash cost − normal cost) ÷ (normal time − crash time). Crash the critical task with the smallest slope first, then recompute the path, because a different task may become critical.
Slope = (crash cost − normal cost) ÷ (normal time − crash time).
Crashing buys time. Overtime, a second crew, overnight shipping: cash in, days out, only on tasks that currently have zero slack. Fast-tracking (overlap) is a different lever and adds risk rather than invoice lines.
The mini-calculator computes one slope. The sequence of which tasks to crash is a loop: slope, crash a bit, recompute the critical path, repeat. Term: crashing.
slope = (crash cost − normal cost) ÷ (normal time − crash time)
Time saved cannot exceed normal time − crash time. Slope is extra cost per unit time. Lower slope on a critical task is the cheaper day.
Each example uses numbers you can load into the calculator above; the results match the tool exactly.
Slope = (5500 − 4000) ÷ (10 − 7) = 1500 ÷ 3 = $500 per day. These are the mini-calculator defaults. Each day you buy on this task costs $500, up to three days.
Task A (critical): slope $500/day, max save 3 days. Task B (critical): slope $300/day, max save 2 days. Crash B first (two days for $600), recompute; if A is still critical, buy days from A next. Do not crash a task with float until the path has moved onto it.
If the task has slack, shortening it does not move the end date. You spent money to create more float. Check float first on the slack guide.
Slope assumes a straight line between normal and crash. Real overtime is often a step: one extra person for a week, not a fraction of a day. Treat the slope as a ranking, then price the actual package.
Crashing buys a shorter duration with extra cost or people. Fast-tracking overlaps tasks that were finish-to-start and usually adds rework risk, not a line-item cost. A schedule buffer is extra time you already added; spending it is not crashing. Status language mixes the three. If a sponsor says “crash it”, ask which critical task, by how many days, at what slope, and what quality you are willing to lose. If they mean “start the next phase early”, that is fast-track. If they mean “use the two-week contingency”, that is buffer. The calculator on this page only ranks crash slopes. Sequence overlap belongs in the predecessor list on the critical path calculator.
A path is 20 days. You need 17. Task A is critical, normal 8 days / $4,000, crash 6 days / $7,000, slope $1,500 per day. Task B is critical, normal 6 / $3,000, crash 5 / $3,400, slope $400 per day. Crash B first for one day at $400. The path is now 19 days. Recompute: A is still critical, a new parallel path is now also critical. You cannot crash B again. Crash A by one day at $1,500. Stop and price the two packages rather than pretending you can buy 0.3 of a day. If a non-critical task had a $50 slope, ignoring it was correct: shortening it would not have moved the 20-day date.
Crash time is an input: the shortest duration with extra resources. What you calculate is the cost slope, extra cost divided by time saved, and then you apply it on the critical path.
(Crash cost − normal cost) ÷ (normal time − crash time). Example: $1,500 extra to save 3 days is $500 per day.
Not to move the end date. Only after they become critical, or if you have another reason (a person who must leave).
No. Fast-tracking overlaps tasks and adds risk. Crashing adds cost to shorten a task.
Only if the work can vanish, which it almost never can. A crash time of zero makes the slope undefined or infinite.
The critical path calculator marks zero-float tasks.
Written by Umesh Manral. Formulas checked against the site's test suite on 11 November 2026. Last updated 10 November 2026. Formulas and tests: methodology. Report a calculation error.