The pour is running north across the ground floor and everything looks routine.
Then the deck starts to move. Not fall, move, in the way a floor is never supposed to.
Somebody calls it and the site walks out.
Below them, across the full width of the building, props are folding one after another.
By the time it stops, a whole wet floor has come to rest a level lower.
Nobody was hurt, and that was partly luck.
Why a fresh pour loads floors you are not standing on
Concrete that has just left the truck has almost no structural strength. For days it is dead weight and nothing else.
So the wet slab does not carry itself. The formwork carries it, and the formwork stands on the floor below.
That floor is usually young too, so props run down again to the level under that, and sometimes further. Those are the backprops.
The load path is a chain, and every link in it is a slab that was cast weeks earlier and has not reached full strength.
Which means the backpropping design has to know exactly which slabs are in the chain, how old each one is, and what state it is in.
A generic prop schedule cannot answer that, because the answer changes floor by floor.
The sequence, in order
Concrete was being placed on the ground floor, north of a grid line, when the first thing let go.
The backprops between the second basement and the first became overloaded and failed progressively across the entire width of the building, east to west, a distance of about 105 feet.
That left the first basement slab holding a load it was never meant to hold on its own.
It failed next, progressively, along the entire edge of the band beam, and hinged at the grid line rather than dropping cleanly.
With nothing under it, the ground floor formwork slid down the inclined surface of the collapsed slab.
It came to rest on the second basement slab, a full level below where it started.
What the investigation actually blamed
Not the props. Not the concrete. The drawing. Every backpropping failure of this shape traces back to a document rather than to a component.
The formwork design had not been made specific to this project and this area of it, so it did not describe the loading conditions that existed on the day.
In particular it did not account for how a partially cast slab behaves at a construction joint, where the finished portion acts as a cantilever rather than as a supported span.
That is not an exotic condition. It is what every staged pour produces, and it is the normal case rather than the exception.
The design also needed to state the pour sequence and the rate at which concrete would be placed, because both change the load arriving at any moment.
A schedule that says how many props and at what centers, with no reference to any of that, is a materials list.
What actually saved the crew
The evacuation worked because the failure was progressive and visible. Deflections appeared before anything fell.
That is a property of this particular collapse rather than a feature of the system. A punching shear failure at a column head gives no such warning.
So the honest reading is that the sequence bought time, not that the controls did, and the official account says the sequence allowed a full evacuation.
The remedy is also unglamorous. It is drawings, certification and a second pair of eyes, which is why it slips.
Concrete work rewards momentum, and nothing about a formwork review shows up on a program the way a poured deck does.
It is the same reason a mass timber tower can spend 5 years arguing before anyone lifts a beam, and the same lesson in reverse.
The five things a contractor is told to do
Produce formwork and backpropping design drawings that are specific to the project and to the area, not borrowed from another floor of the same job.
Have the design address pour sequence and rate of placement, so the drawing describes the day rather than the building.
Get the drawings certified by the formwork design engineer against the relevant code of practice and the national formwork standard.
Have the project structural engineer review that design, because the person who designed the permanent structure knows what the young slabs can take.
And accompany the inspection engineer on every inspection rather than reading the report afterward, a discipline that also sits behind other formwork alerts.
None of that is new engineering. It is the cheap half of the job, and it was the half missing here, the kind of discipline that usually gets written into a replacement span only after something has already come down.
