The knee gives way faster than a reflex can arrive

A ligament ruptures inside a window shorter than any muscle response, which makes the position taken beforehand the only thing that mattered.

Video analysis of non-contact knee ruptures keeps landing on the same figure. From the moment the foot hits the ground to the moment the ligament has failed, something in the region of forty milliseconds have passed. A stretch reflex takes roughly thirty to forty. A voluntary correction takes two hundred or more.

That arithmetic settles an argument. Whatever the muscles do during the injury, they are not preventing it, because they cannot generate meaningful force inside that window even if they start immediately. The protective work had to have been done in advance, in the tenths of a second while the athlete was still in the air, positioning the limb and setting the stiffness of the joint before contact.

The mechanism at the moment of failure is a combination rather than a single direction. The knee is close to extended, taking a large vertical load on a foot that is planted, with the shin rotating and the joint moving into valgus while the trunk is off to one side. The ligament resists forward translation of the tibia and a share of the rotation, and when the load exceeds what it can carry, it does not stretch gracefully. Ligament is stiff by design. Stiff materials fail suddenly.

What decides the position, then, is preparation. Where the trunk was when the athlete committed. Whether they landed with the hip and knee flexed enough to have travel available. Whether they were looking at the ball, at an opponent, or at the ground. Whether they had planned the movement or were reacting to somebody else's, since unplanned cutting produces substantially higher joint loads than the same cut done on command.

So the useful training targets are the ones that operate before contact, and they are unglamorous. Landing with the hips back and the trunk over the base. Deceleration capacity, which is what buys the option to plant somewhere sensible. And enough exposure to reactive, chaotic situations that the athlete's preparation is running automatically rather than being computed.

I am unconvinced by any programme that spends its time on the last forty milliseconds.

That window is already lost when it opens. The injury was decided while the athlete was still in the air.