Person-seconds, not seconds
A rostrum that takes four people forty seconds is 160 person-seconds of work. Six crew can in principle clear 160 person-seconds in 27 seconds — but not this piece, because it needs four of them simultaneously and the other two have nothing to do with it. That is why the calculator reports a range rather than a number.
The lower bound assumes perfect parallelism: everybody busy every second, no waiting, no collisions. The upper bound assumes the shift is a queue: the whole crew handles one piece, then the next. Real shifts land between them, usually closer to the bottom for lists of many small pieces and closer to the top for lists dominated by one big awkward thing that everybody has to be present for.
The travel is often bigger than the handling
Eighteen feet each way at two and a half feet a second is 14.4 seconds a trip, on top of whatever the piece takes to lift and set. For a chair that takes nine seconds to handle, the walk is longer than the job. For a rostrum at forty seconds the walk is a quarter of the trip.
On the default list that travel term is 374 of the 1,020 person-seconds — well over a third of the shift spent walking rather than handling. Six chairs alone are 54 seconds of handling and 86 seconds of walking. Whether shortening the run beats adding a person depends on how long the run is and how big the crew already is, so try both in the fields rather than trusting a rule. What is reliable is that the travel term multiplies by every piece and every pair of hands, which is why a list of many small items is a walking problem and a list of two heavy ones is not.
The wings are a floor problem before they are a time problem
Four flats, a door unit, a table, six chairs and two rostra is not a lot of scenery. Laid out singly it covers 113 square feet; stacked two deep it still wants 69. A ten-foot wing with a four-foot running lane leaves six feet of storage depth, so 69 square feet is eleven and a half feet of wing, or under six feet a side if you can use both — before the second scene change adds to it, because the pieces from change one are still there during change two.
Stacking deeper buys floor and costs access. A stack six deep is fine for pieces that go on at the end of the show and useless for pieces that come back in the next scene. The plan that works is the one that orders the stack by when each piece is next needed, and that ordering is a thing to decide at the production meeting rather than at the half.
What this refuses to answer
How many hands a piece needs is an input here, not an output, because it depends on the piece, the floor, the people and the route, and the honest answer comes from the technical staff of the venue trying it. Nothing on this page addresses flying, hanging, rigging, or any piece that goes overhead: that is the work of a qualified rigger working to the rules of the house, and a page of arithmetic has no business anywhere near it.
Questions people ask
How long should a scene change take?
As long as the cue that covers it, and not a second more, which is why the cue length is an input rather than a target. What the calculator gives you is a bracket: the time if everybody is busy every second, and the time if the shift becomes a queue. If the cue sits between those two, the shift is a choreography problem and a dry run will settle it. If the cue is shorter than the lower bound, no amount of rehearsal fixes it and something has to come off the list.
How do I work out how many crew a scene change needs?
Add up the person-seconds — count times hands times seconds each, plus the walk both ways — and divide by the cue length. That gives the crew you would need with perfect parallelism, and you should treat it as a floor. Then check the largest single piece: if it wants four hands, four is your minimum whatever the division says, because a piece that needs four people cannot be moved by three working harder.
How much wing space does a set need?
Add up the footprint of everything that is offstage at the busiest moment, divide by how deep you can stack each type, and lay that against the wing depth you actually have once the running lane is taken out. The mistake is measuring at one scene change: the pieces from the first change are usually still there during the second, so the peak is later than people check and larger than they expect.
Why is the calculated shift time faster than the real one?
Because the lower bound assumes nobody ever waits, and in a real shift people wait constantly — for a piece to clear a doorway, for the hands a heavy piece needs to become free, for the piece that has to go on before another can come off. There is also a dependency the arithmetic cannot see: some moves have to happen in an order. Time a dry run with the actual pieces and use that, and treat these numbers as the sanity check that says whether the dry run is worth booking.
Does this cover flying scenery?
No, and that is deliberate rather than an omission. Anything that goes overhead — flown pieces, hung goods, anything on a line — is the work of a qualified rigger operating under the rules of the venue, and the failure mode is a piece coming down on a stage full of people. You will not find a capacity figure, a procedure or a rule of thumb about it here. The floor moves on this page are the ones the arithmetic can honestly describe.