all posts post new thread

Kettlebell Periodizing output scaled by movement

Status
Closed Thread. (Continue Discussion of This Topic by Starting a New Thread.)

Brian Johnston

Level 7 Valued Member
Elite Certified Instructor
So I’m scheming a way to maintain “waviness of load” while using different movements and different weights. I plan to do this by estimating the work for a given event, roughly determined by the equation:

reps x mass = work
Yes; I recognize this isn’t the ACTUAL equation for work. No, it doesn’t really matter. Since not all movements are created equal I have to scale my number of reps for each different movement to compare workload between events. So to scale my estimate my equation becomes:

reps x mass x scale = work
Example: If I wanted to do equivalent work for snatches and swings at different weights.

Snatch: 100reps x 28kg x 1 (scale) = 2800 units work
Swing: 125reps x 32kg x 0.7 (scale) = 2800 units work​

I estimated that scale because I think a swing is roughly 70% (0.7) of a snatch. That estimate is based on a comparison of S&S and the Snatch Test. To me, a snatch test at 28kg and a S&S swing session at 40kg have about the same RPE, same time, same reps. 28/40 = 0.7. You may disagree and I’d love to hear about it.

Using the snatch as our base unit of load, what does a clean look like? Is it half the work of a snatch? 60%? 40%? What about a press? A double clean and press? A squat? A TGU???
I’d also be curious if anyone has a way to empirically determine that scale. The PUSH band comes to mind but I have no idea what all that thing can do.
Feel free to let your geek flag fly. If I can’t follow the math or logic I’ll ask for some clarification.
 
To get a rough baseline for different movements I'd probably go with

RPE x reps and work out what the loads are for your lifts at 80% RPE x 8 reps.

and leave it at that. I don't think you can generate units of effort that can be used across exercises, especially comparing ballistics to grinds to dynamic whole-body movements that might not even fit into a traditional rep scheme. The scale multiplier might be a useful means to correlate, but I'd want to whip up the multiplier by actually doing the lifts and noting load and RPE.

It might even be easier to use something closer to
RPE x TUT for a given loading to hit 80% effort for an arbitrary time span, as this accounts better maybe for differences in rapid ballistic vs TGUs etc.
 
Work simplified is really a force X displacement. Obviously the snatch requires a larger displacement (distance traveled) than a swing or clean, and that's why it's more difficult for a given weight. If you were to scale them, it would make sense to scale them by displacement. Cleans & snatches should be more straightforward comparison because the ending spots are pretty consistent. Swing displacement may vary by weight and practitioner. It should be about the same as clean, but typically a little less. It will depend on the standard you hold yourself to for swing height.

I think I get what you're trying to do. But realize this method won't be reliable to compare effort or energy used across drastically different exercises because of different muscle groups being used, and different efficiencies of the movement patterns. (For instance clean vs squat). But simplistically work is weight(force) X displacement.

Another factor you may want to consider waving is density, which time would be a factor. For instance, doing 100 snatches in 5 minutes compared to 10 minutes.

Good Luck
 
@North Coast Miller,
So I like where your head is at and I had some similar thoughts but that assumes my RPE estimations are accurate (they are not). I can comfortably say two events are similar effort, but saying that event represents a 6/10 is beyond me. And I’ve read that a lot of people, even very experienced strength athletes, have a hard time estimating RPE.
I also spent a little time thinking about TUT as a possible metric, but it doesn’t simplify well. Consider 45 seconds of repeated snatch (even at a sedate pace) compared to a TGU at the same weight. Arguably a similar TUT but a totally different experience.

@Papa Georgio,
Similar to RPE and TUT, I considered using the actual work equation to determine a scale. But it also doesn’t encompass the experience (and would also change by athlete). Consider that the work done for a snatch and Clean and press is objectively similar; start from floor end overhead. But the subjective experience is different; difference being rate at which force is applied to get it there.
All that said, I have been using actual work to estimate a scale for movements with a similar pattern for the bell. I’d estimate a combined clean and press is about 1.1 snatches: 0.5 for a clean, 0.6 for a press. Double it for movements with two KBs. A double press would be roughly the same amount of work as a squat with similar weight since it’s the same force over a similar distance in a roughly straight line, so 1.2 for a squat.
I’ve also tried periodizing density and it’s at least part of why I’m trying to work out this scale. Varying only the density gets really old and is hard on the hands if you’re doing something like snatch. If on the other hand I could change density and movement and control for the effort required by movement, then I’m free to substitute at will.

Long and short to both those blurbs is I think RPE, TUT, and the actual work are components of the scale I’m looking for, but lack gestalt that I’m hoping an experienced crowd can quantify.
 
@ Brian Johnston,
I have been known to use a sort of waviness on my park bench programming but do it by using different variations of similar exercises.
Eg. Good Mornings one day for higher reps, single leg DLs for low reps the next, PUs one day, OAPU the next. It is too difficult to get a lot waviness within a bunch of exercises (to me at least) and really doesn't require a lot of variety to make a program like this work - two variations of each movement that hit similar patterns.

The method of swapping out exercises avoids a lot of the guesswork, some movements are just a better fit at specific loading ranges.
 
I've been working on this for a while now for my programming. I have a bunch of different solutions which I'll probably be posting about in the upcoming months as I finish testing them.

My background is in building models in quantitative physical sciences (fancy way to say theoretical physics) so I tend to not like to make up numbers and then use them in further calculations. The errors multiply and the meaning decays pretty quickly. So rather than trying to make up a number (e.g., 2800 units of work) for disparate movements, I try to find the fundamental comparable quantity.

So if I were trying to both wave the load and change the movements, I would probably look at a different variable. One that works for me, is to wave the time. You could do something like this:

You could keep reps constant and vary the total time

10 swings OTM for M1 minutes on day 1
10 snatches OTM for M2 minutes on day 2

with the waviness being M1 vs M2. As you change exercise your work time will vary for the same number of reps

Or you could keep work time constant

:10 on / :50 off swings for M1 minutes on day 1
:10 on / :50 off snatches for M2 minutes on day 2

now the reps will vary with the same amount of work time.

And then you can get really fancy and vary the work time, rest time, and total time. But at some point the complexity become more effort than the gains you get.
 
@ Brian Johnston,
I have been known to use a sort of waviness on my park bench programming but do it by using different variations of similar exercises.
Eg. Good Mornings one day for higher reps, single leg DLs for low reps the next, PUs one day, OAPU the next. It is too difficult to get a lot waviness within a bunch of exercises (to me at least) and really doesn't require a lot of variety to make a program like this work - two variations of each movement that hit similar patterns.

The method of swapping out exercises avoids a lot of the guesswork, some movements are just a better fit at specific loading ranges.
So I don’t disagree with anything you’re saying. But I think it’s slightly tangent to my aim. Your aim in varying exercises is development of a particular movement pattern (good mornings and DLs for hip hinge, PU and OAPU for press). My aim in this would be more like an estimated metabolic load per repitition, so I could measure cumulative conditioning effects without doing the same movement every practice.
 
I've been working on this for a while now for my programming. I have a bunch of different solutions which I'll probably be posting about in the upcoming months as I finish testing them.

My background is in building models in quantitative physical sciences (fancy way to say theoretical physics) so I tend to not like to make up numbers and then use them in further calculations. The errors multiply and the meaning decays pretty quickly. So rather than trying to make up a number (e.g., 2800 units of work) for disparate movements, I try to find the fundamental comparable quantity.

So if I were trying to both wave the load and change the movements, I would probably look at a different variable. One that works for me, is to wave the time. You could do something like this:

You could keep reps constant and vary the total time

10 swings OTM for M1 minutes on day 1
10 snatches OTM for M2 minutes on day 2

with the waviness being M1 vs M2. As you change exercise your work time will vary for the same number of reps

Or you could keep work time constant

:10 on / :50 off swings for M1 minutes on day 1
:10 on / :50 off snatches for M2 minutes on day 2

now the reps will vary with the same amount of work time.

And then you can get really fancy and vary the work time, rest time, and total time. But at some point the complexity become more effort than the gains you get.
Time as an independent variable leaves a lot to be desired if you’re looking to also vary movement. This is especially true with KBs since a your pace is limited by pendulum motion more than your will to go faster. With regard to your example, 10sec of swings is far less taxing than 10sec snatches with equivalent mass. It’s certainly a more concrete variable but doesn’t facilitate variety in programming.
And the units of work need not be arbitrary. Call it coefficient of effort or metabolic-load and cook up some cocktail of measures to generate it. So long as it generates consistent results ¯\_(ツ)_/¯ call it science!
In the end this may be just an academic exercise because it’s very possible this is more effort than the gains are worth and I should just stick to something tried and true... but that’s not much fun :p
 
So I don’t disagree with anything you’re saying. But I think it’s slightly tangent to my aim. Your aim in varying exercises is development of a particular movement pattern (good mornings and DLs for hip hinge, PU and OAPU for press). My aim in this would be more like an estimated metabolic load per repitition, so I could measure cumulative conditioning effects without doing the same movement every practice.

Distilled down, what adaptations are you looking for? This is why I don't think too much about the particular aspect, the variables compound so rapidly my notes become a big mess.

You can't even use HR because the differences from ballistic to grinds and especially lifts that change levels.

End of day, even though your initial plan with multiplier is subjective in many respects it will probably work well enough to be useful. You can always tweak the multipliers as you go should you notice trends emerging that are inconsistent.

Hopefully some more responses coming, this sounds like the sort of thing Kenny might have an opinion about...
 
Distilled down, what adaptations are you looking for? This is why I don't think too much about the particular aspect, the variables compound so rapidly my notes become a big mess.

You can't even use HR because the differences from ballistic to grinds and especially lifts that change levels.

End of day, even though your initial plan with multiplier is subjective in many respects it will probably work well enough to be useful. You can always tweak the multipliers as you go should you notice trends emerging that are inconsistent.

Hopefully some more responses coming, this sounds like the sort of thing Kenny might have an opinion about...
So this may be taboo on StrongFirst forums but my goal is increased work capacity across broad time and modal domains :p
Doctrinal CrossFit programming is trash, but there is a lot of wisdom in widely varied movements. So if I can devise a way to standardize my measure for a variety of movements I can vary the plan but generate a comparable metric for each practice.

I’ve spent a little more time thinking and while HR probably wouldn’t be right, O2 consumption might work. Hook up a VO2 max meter and compare metabolic impact of each exercise. Might work?

Definitely hoping for more responses.
 
Last edited:
Interesting line of thought.

Kettlebell work is hard to quantify because you can't easily separate the strength and conditioning benefits. So much depends on weight/reps/sets/rest as to which end of that spectrum you are targeting.

I think that's what I would try to do, though... For any session or portion of a session, pick one or the other that most closely represents the objective. For things that get you breathing harder and maintain the HR above, say, 110 bpm bpm for the duration, use HR. For things that require strength where the HR recovers to a lower baseline frequently between efforts, use something like your "reps x mass x scale = work" scale.
 
So this may be taboo on StrongFirst forums but my goal is increased work capacity across broad time and modal domains :p
Doctrinal CrossFit programming is trash, but there is a lot of wisdom in widely varied movements. So if I can devise a way to standardize my measure for a variety of movements I can vary the plan but generate a comparable metric for each practice.

I’ve spent a little more time thinking and while HR probably wouldn’t be right, O2 consumption might work. Hook up a VO2 max meter and compare metabolic impact of each exercise. Might work?

Definitely hoping for more responses.

I've tried to do a rough correlate across lifts using HR but was all over the map. O2 meter might work but then you're probably overthinking things.

Not unique to CF, interval training and circuits come in many flavors and are proven to improve work capacity. Mix well with your usual strength training. In that case you could go by HR, timed intervals (work and rest variations), composition of the circuit a la traditional metcon.
 
This is how I would approach it:

- Separate grinds from ballistics.

- For ballistics, if you stay well away from failure, use HR as your measure of equivalence. For instance, start the set at 100 bpm and aim to finish at 150. You´ll have to play with reps and weights for each exercise, but eventually you´ll know that 10 swings with 32 get you there, as well as 7 snatches with 28, and XX double cleans. If you go close to failure I dont think HR is a reliable measure of effort because your strength would be the limit before the HR rises.

- For grinds, just use RM as your measure of equivalence. For instance, you could do 3x5 with 70% of 1RM for presses, and consider that equivalent to 3x5 with 70% RM of weighted pull ups.

- TGU are too weird to be compared with anything else, IMO. You know your own TGU capacities, so just program around that.
 
If you go close to failure I dont think HR is a reliable measure of effort because your strength would be the limit before the HR rises.

This kind of depends on the person and the exercise. In a typical 5-minute snatch test for example, some people are limited by breathing and HR, and others are limited by strength or strength endurance.

But yes, I like this line of thinking:

Separate grinds from ballistics.

And this:

interval training and circuits come in many flavors and are proven to improve work capacity

So if you used HR for ballistics and circuit training, and "reps x mass x scale = work" scale for grinds and strength, that might narrow in on it.

I think I would include TGUs in the grinds/strength, although they are a little different, they could be assigned an appropriate "scale" that also takes into account the fact that you're moving your bodyweight in addition to the kettlebell. So "scale" might need to be two factors -- one that represents the difficulty of the exericse relative to other exercises, and one that represents the degree to which the kettlebell weight matters. If that makes sense.
 
So the things you guys are saying is what I’m wrestling with and why I think this is difficult and worth consideration.

Kettlebell work is hard to quantify because you can't easily separate the strength and conditioning benefits.
I agree! That’s why I’m trying not to separate them! Just because a Snatch is a ballistic movement doesn’t limit its benefit to conditioning. Conversely a military press can develop conditioning and endurance.

- Separate grinds from ballistics
That’s the opposite of my purpose in this. Using HR and %RM are well documented methods for determining progressive overload.

Thinking of strength and conditioning or grinds and ballistics as dichotomous is what leads to events a, b, and c during a practice. You work a for skill development, b for strength, c for conditioning. You set your goals and periodize them individually. If you’re good (as many StrongFirst instructors are) the separate goals and periodization compliment each other. Success is obviously possible, people train like that all the time with varying levels of success. But if you’re not good, you could be wasting your (or your clients’) time, overworking some days, under-performing others.
If, on the other hand, you could determine a consolidated figure by practice (my suggestion being a work calculation, but you could just as easily call it “strength-skill-score” or similar) then you have a tool to compare a planned day in its entirety to the next day, and the next, and so on, until you’ve got a cycle mapped out. A/b/c plan one day, complexes the next, timed circuit the next, and they all feed a consolidated goal.
 
Thinking of strength and conditioning or grinds and ballistics as dichotomous is what leads to events a, b, and c during a practice. You work a for skill development, b for strength, c for conditioning. You set your goals and periodize them individually. If you’re good (as many StrongFirst instructors are) the separate goals and periodization compliment each other. Success is obviously possible, people train like that all the time with varying levels of success. But if you’re not good, you could be wasting your (or your clients’) time, overworking some days, under-performing others.
If, on the other hand, you could determine a consolidated figure by practice (my suggestion being a work calculation, but you could just as easily call it “strength-skill-score” or similar) then you have a tool to compare a planned day in its entirety to the next day, and the next, and so on, until you’ve got a cycle mapped out. A/b/c plan one day, complexes the next, timed circuit the next, and they all feed a consolidated goal.
Ok, I think I understand where you are going, but not completely. I still dont fully understand how would you put together a program from these "coefficients". Lets supose you investigate further and obtain the following coefficients:

- Snatch= 1.00
- Clean and press = 1.00
- goblet squat = 0.8
- Swing: 0.7

How would you go ahead and design the program?
 
If, on the other hand, you could determine a consolidated figure by practice (my suggestion being a work calculation, but you could just as easily call it “strength-skill-score” or similar) then you have a tool to compare a planned day in its entirety to the next day, and the next, and so on, until you’ve got a cycle mapped out. A/b/c plan one day, complexes the next, timed circuit the next, and they all feed a consolidated goal.

Hmm...here's a link to a spreadsheet I made a loooong time ago. I took a bunch of measurements and tried to make an average power calculator. This was in the ancient times when I was part of CrossFit, don't judge me :) There might be something here that you can use to come up with a metric.

It's read-only, so you'll need to make a copy if you want to edit it.

Read Only Share - Crossfit Estimated Power Output

Mike
 
I agree! That’s why I’m trying not to separate them! Just because a Snatch is a ballistic movement doesn’t limit its benefit to conditioning. Conversely a military press can develop conditioning and endurance.

Good point. What if you factored in both strength and conditioning to each? Both would depend on how it was done (i.e., plenty of rest, vs. metcon/glycolytic effort) more than the sets/reps/weight. So, these variables:

A = total volume of work (sets * reps)
B = weight
C = degree of conditioning
D = degree of strength
E = scale, as defined in original post

With the forumla being: Units of work = Units of conditioning work (A x B x C x E) + Units of strength work (A x B x D x E)

So, using your example:

Snatch, done as a 5 min snatch test: 100reps x 28kg x 1 (scale) = (100 x 28 x 1 x .9) + (100 x 28 x 1 x .5) = 2520 + 1400 = 3920

Snatch, done as a A+A repeats of 5 snatches plus rest: 100reps x 28kg x 1 (scale) = (100 x 28 x 1 x .5) + (100 x 28 x 1 x .7) = 2520 + 1400 = 3360

So, how could 100 snatches with the 28kg end up as a different amount of work? Well, I think it could, because to some degree work - stress, and a 5-min test is more stressful (and theoretically more effective in driving gains in the short term, though certainly not an advisable way to train in the long term).

Swings, done as sets of 5-10 EMOM: 125reps x 32kg x 0.7 (scale) = (125 x 32 x .7 x .6) + (125 x 32 x .7 x .3) = 2520

So the snatch workouts are 3920 for a snatch test, or 3360 for A+A, and the swing workout is 2520.

Sounds about right to me.... :) Anyone follow?
 
Ok, I think I understand where you are going, but not completely. I still dont fully understand how would you put together a program from these "coefficients". Lets supose you investigate further and obtain the following coefficients:

- Snatch= 1.00
- Clean and press = 1.00
- goblet squat = 0.8
- Swing: 0.7

How would you go ahead and design the program?
So this will be WAY overly simplified with weird weights and wonky progression, but it makes the math easy and the entry short. 4-day training week, using only those 4 exercises

Week 1, 10000 total work units
Day 1, heavy: 10x10 snatches at 30kg = 3000 work units
Day 2, medium: 5x5 barbell squats at 125kg = 2500
Day 3, light: 2x10 C&P at 40kg, 3x10 Snatch at 40kg = 2000
Day 4, medium: 3x25 swings at 48kg = 2500

Week 2, 11000 total work units
Day 1 = 3300 work units
Day 2 = 2750
Day 3 = 2200
Day 4 = 2750

Week 3, 12000
Week 4, 11000
Week 5, 12000
Week 6, 13000
Week 7, 13000
Week 8, 14000
Week 9, 10000
Week 10, test

How you meet the work goal for the day and week is immaterial. So long as it passes a common sense check, is safe, feeds your program goals, and meets the desired work capacity, you can schedule whatever you like.
 
Good point. What if you factored in both strength and conditioning to each? Both would depend on how it was done (i.e., plenty of rest, vs. metcon/glycolytic effort) more than the sets/reps/weight. So, these variables:

A = total volume of work (sets * reps)
B = weight
C = degree of conditioning
D = degree of strength
E = scale, as defined in original post

With the forumla being: Units of work = Units of conditioning work (A x B x C x E) + Units of strength work (A x B x D x E)

So, using your example:

Snatch, done as a 5 min snatch test: 100reps x 28kg x 1 (scale) = (100 x 28 x 1 x .9) + (100 x 28 x 1 x .5) = 2520 + 1400 = 3920

Snatch, done as a A+A repeats of 5 snatches plus rest: 100reps x 28kg x 1 (scale) = (100 x 28 x 1 x .5) + (100 x 28 x 1 x .7) = 2520 + 1400 = 3360

So, how could 100 snatches with the 28kg end up as a different amount of work? Well, I think it could, because to some degree work - stress, and a 5-min test is more stressful (and theoretically more effective in driving gains in the short term, though certainly not an advisable way to train in the long term).

Swings, done as sets of 5-10 EMOM: 125reps x 32kg x 0.7 (scale) = (125 x 32 x .7 x .6) + (125 x 32 x .7 x .3) = 2520

So the snatch workouts are 3920 for a snatch test, or 3360 for A+A, and the swing workout is 2520.

Sounds about right to me.... :) Anyone follow?
... I like it!
I’d have to sit and think about that for a while though. It makes a lot of sense but doubles the measures I’d be looking to estimate.

A shortfall to all of this is its math heavy so you almost have to use excel or similar spreadsheet software to do calculations.
 
Status
Closed Thread. (Continue Discussion of This Topic by Starting a New Thread.)
Back
Top Bottom