Fantastic episode! I’ve been citing your work for a while when peer reviewing papers in sport/exercise science journals. I’ll be sharing this podcast as well to other students and researchers to help with knowledge translation. Appreciate you both putting this info out there in a consumable way!

If you’re interested in learning more, I’d be happy to share the paper!

5. We assume that the devices used to measure velocities over time are valid and reliable across a spectrum of movement velocities. However, depending on the device you use, this is not the case and you may in fact see that at very fast (or slow) velocities that your device errors change.

In other words, most LVPs are capturing behaviour and not necessarily capacity. This is fine, but we need understand the difference between assessing behaviour vs. capacity.

4. These prediction models assume that different constraints on movement impact load-velocity relationships similarly over time. However, we do know that different personal, task, and environmental constraints on movement can yield different movement strategies and thus impact movement velocities.

3. It is assumed that someone’s “strength” and “speed” capabilities recover at the same rate. While not tested specifically in the context of velocity-based training, there is other data to suggest that maximal strength abilities will recover at different rates from RFD capabilities.

2. There is an assumption that 1RM strength fluctuates more than movement velocities at submaximal loads. Current data doesn’t support this assertion, but again we need more ecological valid studies to assess this assumption.

1. It is assumed that the accuracy of the models we build during an assessment are consistent over time. In other words, that those models generalize well to new data. This assumption has not rarely tested, and the few studies that have been conducted do not appear to be promising.

My personal goal is, like with most of my work, that I look back on this review in 3-5 years and chuckle at how outdated the ideas were, both from having new data and studies and from just standard personal growth. With that made clear up front, let’s dive into some of the key ideas:

This does not mean we believe that this is a “settled” area of research or that we believe velocity data cannot be useful to inform training practices. In fact, our hope is that this review prompts more research in this space to refute or confirm the points raised in our paper.

To summarize the paper, there are five key points. Before diving into them, there’s really one key idea underpinning the entire paper - we believe there is not enough data published yet to substantiate the claims being made with respect to using velocity data to regulate training intensity.

I’m guilty of this too btw and I’m actively being better with this moving forward. I have a ton of python code I need to make more open, but that just wasn’t really the culture in academia. Do we fix this by creating more open source tools for researchers and students to learn? #openscience #science

Why in biomechanics are there so few open source tools, or github repos accompanying studies, that allow people to at least reproduce your claims if you sent them the raw files? Given the amount of data in a typical biomechanics collection (and thus more researcher DoF), this is especially worrying.