Saturday, April 29, 2017

Reality Check: Accelerometer Systems Aren't Useless

I feel obliged to talk about why I don't discount the accelerometer systems in all cases. Firstly, I will remind you that the intent of this blog is to provide a powerlifting perspective on velocity based training. For perspectives other than that sport, I suggest seeking out other experts that will flesh out their opinions beyond 3 main lifts. In very simple terms, I'm interested in velocity based training as it pertains to fairly slow velocities, often emphasizing velocities above 80% 1RM. Secondly, I should also remind you that my first sensor comparison article is not a definitive study. The sample size is 1 (me). And lastly, there are practical reasons for using accelerometer systems. Axes measured is an important part of that, as well as different training modalities. There's always room for growth in these devices.

Anyone that thinks I'm trying to throw a manufacturer under the bus doesn't understand that I've kind of spread it around on all parties. The last thing I want is someone to look at the sensor comparison article or any article and use it as an argument to disparage a company. All of the companies out there are bringing something unique to the table and should be recognized for that. I'm not throwing shade. I'm not trying to hamper their efforts. Each of these devices has advantages and disadvantages.

Of the three companies I've included in my original comparison, I've spoken to a representative from each (PUSH, Beast, and OpenBarbell). At one point in time, I was even a blog contributor for PUSH. If I was indeed bashing accelerometer systems, why would I ever work for PUSH?


This is the most important part of VBT for strength training. This is not the most important part for all training. Even in strength training, VBT is not the most essential way to monitor, augment, or otherwise manage your training. I've said it before:
If VBT is the most important part of your training plan, your training plan probably isn't good.
That said, if you want to use VBT for strength training, then you will certainly need a device that measures reliably at slow velocities. It does not need to measure accurately, but it must be done reliably.  Accelerometer systems seem to work best above 0.40 m/s or around 80-85% 1RM on many lifts for average folks. Tethered systems (or linear position traducers - LPT's, optical rotary encoders, etc) tend to work reliably around this range for barbell movements. That does not mean they are without fault.

I'm not an engineer, so I can't really talk with great confidence about the full range of capability of accelerometer systems. That said, I have seen systems that use hardware that plausibly should be capable of measuring velocity below 0.40 m/s. If the hardware could measure it but isn't actually returning reliable measures below that, I would surmise this could be an issue of how the signal is processed or broadcasted. All of these devices work through Bluetooth and all seem to perform better on iOS (Apple) devices because of how Bluetooth low energy works on Android (protip: at least buy an iPod if you want less Bluetooth frustration). Additionally, it seems plausible that data can be bottlenecked by the constraints of Bluetooth, limiting their full capability. If Bluetooth was the limiting factor, that's truly a sad state of affairs because 1) the manufacturers are limited by developments on the smartphone end and/or 2) future iterations of the devices must be developed to keep pace with signal broadcast if Bluetooth is the limiting factor. The upside of this is manufacturers could alter the way the signal is broadcast, providing real-time information in short bursts and transmitting extraneous detail after completion of the recording period (after your set).

The other limiting factor could be related to how frequently data is sampled and processed. If this is the limiting factor, the great hope is that manufacturers can update the software end to mitigate this. Anyone that's an early adopter knows this experience. Most devices within the first year of the commercial release act buggy. It isn't until after some time that these issues fall by the wayside. You should keep this in mind when you're looking at accelerometer device reviews on youtube from over a year ago, or even the comparison data I provided in future time.

When I fault devices for not performing well with deadlifts, the common response is some people CAN get deadlifts to work with the device. It usually involves performing the movement in a special way that is arguably too far away from the mode of training to make it feasible.  And just to be clear, I haven't talked to a single manufacturer that was satisfied with how their device was assessed for deadlifts. Maybe there isn't a good solution, maybe deadlifts behave erratically, maybe my consistency in deadlifts suck, or maybe my attempts to measure deadlifts objectively are flawed. Could be any single one, or it could be all of them.


Powerlifting barbell movements are predominantly in 1 axis, but you could argue that all movements happen in three axes enough for it to warrant attention. Even the bench press, which moves in two axes more than the others, doesn't seem to be hampered by measurement in one axis alone. Different devices measure in a different number of axes. Most devices allow for a change in configuration so you can measure in an axis/axes of your choice (floor/ceiling/horizontal mounting). S

That said, depending on what you're doing for assistance/accessories and how much VBT integration you intend to have, it might be advantageous to choose a purpose driven implementation of a VBT device. In particular, lunging in general (side, traveling, and possibly forward/reverse lunging), some types of landmine work, step ups, glute-ham raises, pull ups/chin ups, could be limited by your device selection. 


Consider how much a GymAware cost. At a price tag of $2200, that's not including a subscription. Next, pricecheck a Fitrodyne Tendo or a T-Force unit - don't forget to consider how convenient these set ups are. Last check on the tethered-unit supremacists: tell me the next estimated ship date for OpenBarbell V2. Compare this to the availability of accelerometer systems and they obviously have a leg up. Accelerometer systems have a more favorable equilibrium on both supply and demand.

Just as much as I recommend using some form of autoregulation training before jumping into VBT, I would also recommend incremental implementation of VBT devices. I would recommend trying a product out that has a solid return policy or using one from a friend/facility that already has one. Alternatively, you could minimize your losses by buying at the lower end in terms of price if you weren't sure if you wanted to take the $2200 dollar dive.


Don't assume I chose the best design method. I didn't. I went with what was practical and tolerable in order to develop a rationale for my case to implement VBT. The target was the consumer market. The only other comparison of multiple devices I've seen talk about them in generality, occurred long before the devices were well-established (IE: before they worked out the major bugs), or simply stood as opinion pieces. I don't think any of these really helped people determine which unit to buy.

If you think I'm going to toss out my PUSH now that I have an OpenBarbell, you're wrong. Depending on what I mean to do, the PUSH still comes out.

This isn't a validation study. At best, the comparisons are akin to what DC Rainmaker for cycling power meters or NotebookCheck does for laptops. At worst, it's little better than Men's Health reviews or Gizmodo.


Accelerometer systems aren't garbage in all things lifting. Accelerometer systems may not even be the worst option in all things. In the end, it's almost like asking what's a better motor vehicle: a Prius, an F150 truck, or a Formula 1 racing car? Depends. Are you focus on commuting, hauling cargo, or getting from point A to B in the least amount of time?

No comments:

Post a Comment