VBT: Deadlifts are not your friend

If you want a how-to-deadlift guide, I suggest my favorite website. That's not what this is.

Edit 8/8/2017 - Since I posted this two days ago, my google scholar news alert pulled up a new article examining the relationship between velocity and prediction of number of reps at a given intensity (XRM)/reps in reserve (RIR). They created generalized equations to predict XRM and RIR, but found that like all things VBT, it truly needs to be individualized. Furthermore, they showed that XRM can be determined from velocity, but RIR does not appear to be identifiable. This seems to say that it is possibly less ideal to use the exertion-velocity relationship identified by practitioners like Mladen Jovanovic than it is to use velocity to determine XRM. On a practitioner level, this seems inconsequential since you can just use the velocity-XRM relationship of that individual to drive exertion. If you want 2 RIR, just perform the velocity XRM plus 2 reps.

In effect, even though I have weakly advocated the exertion velocity-relationship, it is likely more appropriate to use the velocity-XRM relationship using your individualized data. Even though I resolved this solution specifically for the deadlift, it's actually more appropriate for more movements. I'm not going to edit this to reflect this new finding.

Introduction

The deadlift is a bratty kid in velocity based training. It thinks it's the center of the universe and doesn't play by anyone's rules. If you sort through many of the velocity based training research articles, you will find an overwhelming amount of information on the bench press, bench throws, squat, jumps, squat jumps, pull ups, and prone rows. Likewise, many of the principles established in the one exercise you can often apply to the other. But you don't see much for deadlifts. One possible explanation for this is it's hard to publish negative results.

Key Points:

• Most exercises in PL have you start at a velocity and slowly bleed that velocity off with each rep. Deadlifts do not do that though.
• Since velocity loss is a factor in auto-regulation in VBT, there has to be a reason why.
• Nope. Not a reason I see. Maybe deadlifts just suck?
• I have an idea of how to solve this by re-working something else Mike T does. I steal from Mike all the time

Velocity Loss Across a Set

If you take a set for the bench press and go until failure, you will see velocity start off fast and slowly drop off to grindtown. You see the same thing with the squat. This has been termed velocity loss and is talked about to at length here.

Mean concentric velocity across multiple intensities of the bench press and half squat from Izquierdo et al.

In a previous article about the future directions in VBT, I alluded to the potential for research in different variations of movements. The research has ventured here in terms of the load-velocity relationship and appears to be continuing that way. For sake of simplicity though, here's how that has looked in the past year of my experience:

A comparison of velocity loss across multiple movement types

You see the general decrease in velocity with successive reps. It appears the deadlift doesn't do that though. You have high coefficients of determination (R²) for everything except the deadlift. Moreover, if you look at the trend line for the deadlift, it's relatively flat. It's like a table that has that one leg that needs a shim. And much like that table, I also get the impression that fixing it isn't on anyone's to-do list and we'll just continue to live with it.

Velocity Across Intensities

Deadlifts aren't entirely useless in VBT. It appears velocity is useful across different intensities, but confusing across a set. Here's an example:

Velocity across different sets in the conventional deadlift. After each set, load is increased. This plotted using only the initial velocity rather than the average of multiple reps  (3 for light, 2 for moderate, and 1 for hard sets) as I have advocated in the past.

As you can see, with each successive increase in load, velocity appears to decline. The coefficient of determination seems high enough (above 0.90 would be optimal). If you were to isolate this into the %1RM that you're most likely to work with in powerlifting (for strength or hypertrophy), most sets will fall between sets four to seven.

So while we can't regulate volume within a set (and therefore exertion), we can obvious regulate load. Likewise, you could be specific about how you regulate fatigue. An example of this would be hitting 80% 1RM for as many sets as possible. Once initial velocity drops below a predetermined velocity threshold, the volume (in sets) has been achieved.

Determining Confounding Factors

There are maybe a few things that could contribute to this phenomenon, and I'm not sure what it is. I tried to identify as many potential candidates as possible and figure out a way to tease out the results.

Sumo vs. Conventional

If you pull your deadlifts sumo: 1) you eat ass 2) you're a goddamn cheater 3) you make that singlet look good. 

For sake of argument, there's nothing that makes this phenomenon unique to deadlift type. The first graph of velocity loss across sets is sumo. The second graph was sets of 3, and in each set you get a rather noisy picture in nearly every set. So that's not it. 

From a motor learning perspective, it makes more sense for me to limit this to the sumo deadlift. It's my competition deadlift. Over the past year, I've done half a dozen different variations of the sumo deadlift. In that same time period, I've only done one type of conventional deadlift. If I was regularly cycling in conventional deadlift variants, it would make sense to prefer this. In the end, the two lifts are more similar than dissimilar. Even if you don't agree with that, take your first world problems and GTFO.

Concentric Only Lift

The common thread that squat and bench share that make them differ from the deadlift is that the lifts are preceded by an eccentric movement. The running idea here is that this eccentric loading transfers some force through the stretch reflex. Even with a pause, the stretch reflex lingers for some time. This was what prompted Louis Simmons to use the box squat and unload the muscle.

There's also the argument that velocity loss across a set could be different with touch and go vs pull and reset. I wasn't seeing such a pronounced difference on a cursory survey. Also, the rationale doesn't make sense to me since I generally see concentric-only lifts (pin presses and pin squats) to be more reliable. To tease this out, I attempted to emphasize the concentric part and start the set with the eccentric portion of the movement. I basically did what all the pull-and-reset Nazis complain about and made it into something they would hate even more. I started the bar an inch below lock on an adjustable height monolift. Then I pulled the bar that final inch as the monolift swung away. At this point, I'm able to start the deadlift by lowering it. 

The natural question is "how do you finish the set?" There's two ways. You can get to lock out and try to duck walk the weight back to the monolift like an idiot or finish by putting the bar back down on the ground. I'd like to tell you I didn't learn which method the hard way. But I did. Yes. I did. I don't think this movement helps develop anything particularly better than normal deadlifts, but if you're going to do them then just end with the bar on the floor.

This emphasizes the eccentric part across the whole set - not just from the second rep and onward. Let's just call it the "sumo droplift." You heard it here first folks.

SUMO DROPLIFT^TM

Proudly presented by Zero Sum Gains

Range of Motion and Bar Deceleration

Powerlifting often comes down to the weak range of motion. In the deadlift, that either happens off the floor or at the knees. Typically conventional deadlifts fail due to deceleration near the knees and sumo deadlifts fail to generate enough acceleration off the floor to power the move into lockout. It could be possible that acceleration from the floor to the knees or from the knees to the lockout is the confounding factor. When that irregular behavior gets mixed in with the rest of the movement, that might be what's causing the rather noisy picture of velocity across the set.

The easy way to tease this out is by separating the parts of the movement. Block pulls are a common supporting movement. I'm just going to do them as sumo block pulls, which is less common. To get the range of motion from the floor to the knees, I've decided to use spotter arms upside down. Once both sides contact the safeties, I lower the weight again. 

If the range of motion is the issue, then obviously one of these two should come out significantly cleaner than the other.

Jefferson Deadlifts

Obviously, this will show us everything we need to know about deadlifts. This is really the only deadlift you ever need. Not even in powerlifting. This is the only deadlift you need in life.

SMDH - What it looks like

Normal sumo deadlifts are flat. This particular set has a lower coefficient of determination than the one I showed in a previous graph (0.003 vs 0.3). The sumo droplift is not only flat but has the worst coefficient of determination (R²) out of all of the lifts (translation: no. Dumpster fire). Block pulls have a downward trend and half sumo to the knees have a counterposed upward trend. Just by the look of the trend lines, it would seem like half sumo to the knees are what contributes to the noise with velocity gain across a set. In reality, though, coefficients of determination (R²) of 0.03 and 0.11 are hot garbage.

I'm not sorry you read through all that bullshit just to arrive at the fact that I don't know why deadlifts for reps doesn't work. Maybe God just doesn't want you doing high rep deadlifts. High reps=cardio, right? All I can tell you is that the things I identified as potential confounders don't appear to be it. 

So why bother? Why waste your time? Everyone seems to have this idea that because VBT doesn't work because of a lack of understanding of how to use it. It's still very useful to regulate load/intensity with velocity. Don't let frustration with VBT deadlifts drive you away from VBT.

The Work-Arounds

I've harped on the idea that VBT is only one way to autoregulate training for powerlifting. If you want to blend your methods, that's something you can easily do. If you're experienced with subjective measures like RPE, you can easily integrate that. Alternatively, you could just program your deadlifts via percent of training max. You can even augment that training max at whatever frequency you want with AMRAP sets using the usual calculation Jim Wendler uses or this. Wendler's favorite:

wt. x reps x .0333 + wt.= 1RM

There is an alternative though. Let's take for example the RPE chart from Reactive Training Systems:

Not original content. See source link

We know that the load-velocity relationship works. In that same instance, the %RM-velocity relationship works. The only issue we have is obtaining the MVT. MVT is usually the last rep in the tank whether it is a 1RM or a multiple rep max. In this case, though, MVT is best obtained through testing 1RM and can be updated on occasion with reps slower than MVT that occur in training (because that happens). With that said, we can use the %RM-Velocity slope and intercept (nerd shit, go back to the "How the Sausage Gets Made" article and download the excel files) to do a reverse look up on the velocities. 

So rather than getting %RM value to use, you can use a mock up of this chart to determine what your initial velocity should be. Here's a visual:

Assuming this works, you can better control exertion/proximity to failure. Unlike bench and squat, you can't auto-regulate within a set, using a stop velocity to tell you when to terminate a set. Instead, you're working on delayed feedback. Intensity/load is still real time, but you're basically flying blind and assuming work capacity is fairly stable. If you give up the gas mid set and hit a wall, obviously the only thing determining the end of the set is subjective feedback (IE: feels heavy, bro. I be done). 

This is the streamlined way I've done this. After playing around with how to auto-regulate deadlifts, I've taken to using MyStrengthBook's personal record table screenshot below:

The menu gives you a table of your rep-maxes, which I initially used for programming. Upon realizing I could convert the raw weights into percent 1RM, it became obvious that there is a non-real time alternative to terminate sets to a predetermined exertion level using velocity. I'm not going to say it's perfect since I'm still evaluating it, but it's food for thought.

Continuing Forward

It's been a good three months since my last update. I said from the beginning that this wasn't going to be an indefinite series. Potentials on the horizon: Gym Aware included in the comparisons, Open Barbell V3 included in the comparisons, PUSH evaluated on update sumo deadlift algorithm, an RPE-VBT excel template, and VBT/RIR in typical bro exercises (depending on lit review).

Some of the things I said in the past year have been elucidated by newer research. Not anything refuting things I said (at least none I'll admit to), but I think about half of my future directions have already been knocked off the list if not currently in-process. I'm not going to re-write that. Likewise, some issues on the manufacturer side have been changed. I try to add footnotes at the beginning where appropriate, but for the most part it's hard to keep track since I primarily train with one device.