Total float is the amount of time an activity can slip without delaying the project finish. It's computed as Late Start − Early Start, which always gives the same answer as Late Finish − Early Finish. Drag the slider below to see it happen live.
Total float vs free float
Total float is measured against the whole project finish — how much X can slip before the project finish date moves. Free float is narrower: how much X can slip before it delays its own immediate successor, ignoring the rest of the network. An activity can have zero free float (its very next successor is tight against it) while still carrying total float, if float reappears further down the chain.
Float belongs to the path, not the activity
Float isn't really a property of one activity — it's shared across every activity on that non-critical path. If one activity on a float path uses up float (runs long), every other activity sharing that same path loses the same amount of float, because it's the same spare time being measured at different points along one chain.
Negative float
When an activity's Late Start comes before its Early Start, total float goes negative — drag the slider past 10 in the visualiser above and watch it happen. Negative float means the network, as logically linked, cannot actually finish by the date it's being measured against; something upstream needs to be resequenced, compressed, or the target date itself needs to move.
High float as a warning sign
The DCMA 14-point assessment flags any activity with more than 44 working days of total float as suspect — not because float is bad, but because an activity with that much slack is usually a sign of missing logic (nothing meaningfully constrains it) rather than a genuinely low-priority task.
Why the float type setting matters
P6's project-level float calculation setting changes how float is computed across relationship types, which is why two schedulers can open the same XER, apply different settings, and get different float numbers for the same activities. When comparing a contractor's schedule against your own recompute — or comparing float across two organisations' schedules — mismatched float settings are a common, easy-to-miss source of "wrong" numbers that are actually just computed differently.