Yago Ferrieros: Back Stability

[Mark Rippetoe]

Squatting and Back Stability | Yago Ferreiros

[Shiva Kaul]

To summarize: driving your hips briefly increases the back angle θ by Δθ, which increases the moment τ by Δτ. If Δτ is large, even if τ is also large, your back is still strong enough to support τ + Δτ. But you can't react quickly enough to prevent instability/flexion. The solution is to start with large θ and τ (bend over), which makes Δτ small (per Figure 4).

[Mark Rippetoe]

That's an awfully short article.

[Shiva Kaul]

It's a good observation. His control-theoretic perspective, even without difficult calculations, helps focus on the right issue: dynamic instability, rather than statically “crushing the spine.”

[zft]

But you can't react quickly enough to prevent instability/flexion.

The article doesn't explicitly model time and makes the implicit assumption that Δt (i.e., the time it takes to change the back angle by Δθ) is constant in all scenarios and dθ/dt is constant for a given starting θ. The assumption on Δt is definitely an unrealistic assumption---timescales are crucial in regards to responding to pertubations; more realistically (and very consequentially to the analysis), both Δt and dθ/dt are functions of θ (and probably also L and F), with dθ/dt also being a function of t. This means that Δτ is an insufficient descriptor of stability and that dτ/dt is the actual quantity of interest; it may be the case that in this more representative model, max (dτ/dt)_horiz >= max (dτ/dt)_vert.

[Shiva Kaul]

Involving time seems useless. The whole point is: the Δt of hip drive, for any θ, is too quick for the lifter to react and prevent potential instability. To the lifter's nervous system, it is indistinguishable from a constant. The assumption is the lifter does not have superior kinesthetic sense. Highly-skilled squatters (you may be one of them) manage to high-bar squat with hip drive and small Δt while staying balanced and rigid. Yago's not talking about them.

The entire mathematical analysis here is "the derivative of sine is cosine." Which is nice. If you did want to open the can of worms, with time and differential equations, why would dθ/dt be smaller at small θ?

[zft]

The whole point is: the Δt of hip drive, for any θ, is too quick for the lifter to react and prevent potential instability. To the lifter's nervous system, it is indistinguishable from a constant.

My contention with this is that no evidence to support this is actually the case has been presented and it's not intuitively obvious to me at all. I'm not sure Δt is even that small itself and I don't know how it compares with the timescales of a lifter's ability to react. You can certainly appreciate that as Δt becomes larger, Δτ becomes increasingly less informative in regards to stability.

why would dθ/dt be smaller at small θ?

I didn't say it would be and I imagine actually answering that question is hard.

I do not mean to unnecessarily complicate the model here; I just think that the conclusion of the article may not actually correspond with reality because the model is too simplistic.

[Shiva Kaul]

My contention with this is that no evidence to support this is actually the case has been presented and it's not intuitively obvious to me at all. I'm not sure Δt is even that small itself and I don't know how it compares with the timescales of a lifter's ability to react. You can certainly appreciate that as Δt becomes larger, Δτ becomes increasingly less informative in regards to stability.

Hip drive is normatively very quick. At heavy weight, you're not supposed to purposely increase θ; that is an unfortunate side effect of driving hard. That is why, in Yago's model, Δθ is a random perturbation. Δt is a therefore a period of time during which error accumulates in θ. To say it is longer than the reaction time seems contradictory.

Is there empirical evidence that lifters are destabilized by Δτ from hip drive under a vertical back? I'll let coaches comment on that.

I do not mean to unnecessarily complicate the model here; I just think that the conclusion of the article may not actually correspond with reality because the model is too simplistic.

A more complex model is fine, but there should be some intuition about the different solution it will obtain. How do you think a more vertical back keeps the lifter more stable out of the hole, assuming they want to use their posterior chain?