Training Volume and Muscle Growth: Part 3

Ok, let’s finish this thing up.  So far I’ve looked at the 7 current studies (as of this article’s writing in October of 2018) in often excessive detail in Part 1 and Part 2 and now it’s time to put them all together to see how training volume and muscle growth relate.  As noted in Part 1, I’m throwing the Radaelli paper into the trash.  I consider the results too random and nonsensicial  to be worth considering.

There is simply no world where growth in triceps in beginners doesn’t start until 45 sets per week but 18 sets for biceps is effective and where LBM gains are higher for calisthenics than low-volume weight training.  So it’s out.  Agree or not, I put my reasoning up front and looked at it in detail to explain why I think it’s garbage so it wasn’t just a hand-wave like most would do.

That leaves 6 studies in trained individuals (minimum 1 year training experience and a usual range of 1-4 years) looking at different volumes of training and the muscle growth response.   Yes, they used varying methodologies, some only used body composition methods via DEXA, some used DEXA and Ultrasound, one used DEXA, Ultrasound and muscle biopsy (of quads only).

As I stated at the outset, I’m going to simply take them at face value for the time being.  It’s the data we have and with the qualifications I was sure to make as I went, it’s what we have to build the model on at this point.  Yes, future data may change the model.   When it arrives, the model will have to be updated with it.

Building the Model

First, let me put the 6 remaining studies together to see if a pattern shows up.

Yes, this is a terrible chart and I probably got at least one of the numbers typed in wrong since I type fast and frequently don’t take the time to check after the fact.  The true guru will dismiss the entire article based on a typo.   But guru gon’ guru and there’s nothing I can do about that.

Instead let’s focus more on the generalities of the data and less on my lack of proofreading.   As no values went down, all numbers represent an increase from the beginning of the study.  Percentage change is a percentage change and any absolute numbers are mm changes in muscle thickness.  I’ve shown the best response in red.

Paper Muscle Low Mod High Volume of best response (sets/wk)
Notes
Ostrowski Quads 6.7% 5% 13.3% 12 Nothing higher tested
Triceps 2.3% 4.7% 4.8% 14 28 sets no better than 14
Amirthalingam Totes LBM 2.7% 1.9% Lower volume better than higher
Trunk LBM 4.1% 1% Lower volume better than higher
Arm LBM 7.8% 3.4% Lower volume better than higher
Leg LBM 0.5% 2.4% Higher volume better than lower.
Tri MT 2.3 4.5 18-19 triceps
12-18 quads
For triceps 26-28 sets upper no better than 18-19
Leg volume of 12-13 and 16-18 so similar that I am considering them as part of a full range.
Bi MT 2.4 0.3
Ant Thigh MT 2.6 1.1
Post Thigh MT 1.2 2.2
Hackett DEXA only 18-19 vs. 26-28 upper

12-13 vs. 16-18 lower

Essentially identical to Amirthalingam with moderate volume superior for overall LBM gains and upper body and the slightly higher volume slightly better for lower body.  Differences were overall small and moderate and high volumes were more or less identical.  Very small number of subjects weakens statistical power.  Lack of direct measure of muscle thickness is not ideal.
Huan LBM changes Gains up to 20 sets/week.

No further from 20-32.

By muscle thickness via Ultrasound, triceps showed growth up to 20 sets and then SHRANK.
By biopsy, quads shrank up to 20 sets and grew after that.  Unclear why Ultrasound did not pick up quad growth but biopsy did.Total summed changes in triceps and quads a whopping 1.8 mm.  Miniscule changes overall.
Over 20 sets per week, water retention measured by TBW and ECW significantly increased, making LBM gains above that point insignificant.Two conclusions: cap to useful volume of ~20 sets per week.  Volume without tension is shit for growth because volume is NOT the primary driver of hypertrophy and never will be.
Schoenfeld Biceps 0.7% 2.1% 2.9% 18 sets By their own statistical methods, the highest volume was weakly/insignificantly better than moderate (described as ‘not worth mentioning’ in stats texts).

Conclusion: 18 sets upper and 27 lower optimal with higher volumes showing no meaningful differences (and certainly not for doubling the volume).

RF and VL changes ADDED together for poor comparison to Ostrowski.

Data in question because of an outright LIE in the discussion regarding Ostrowski triceps data.

Triceps 0.6% 1.4% 2.6% 18 sets
RF 2.0% 3.0% 6.8% 27 sets
VL 2.9% 4.6% 7.2% 27 sets
Heaselgrave Biceps 1 3 2 No statistical difference but a trend for 18 sets better than 9 and 27 no better than 18 (or even worse).

Ok, now let me summarize that horrible chart by showing where each study finds that their optimal results fall in terms of sets per muscle group per week.

Study Optimal Volume
Ostrowski 12 sets lower body (no higher tested), 14 sets upper body
Amirthalingam/Hackett* 12-18 sets lower body, 18-19 sets upper body
Huan 20 sets upper as a cap, possibly 20+ for lower body (needs more study)
Schoenfeld 18 sets upper (compared to 6 with no middle value for comparison)
27 sets lower (compared to 9 with no middle value for comparison)
Heaselgrave Trend for 18 better than 9 but 27 no better than 18.

*I grouped the two studies together since they were an identical methodology and only differed by length
and I’m getting tired of writing and making tables as it’s a pain in the ass in WordPress.

Looked at this way a pattern starts to show up.  Which is that a moderate volume tends to beat out either lower or very high volumes under basically every circumstances in trained individuals (again defined in most studies as 1-4 years of training or a minimum of 1 year regular training).  Or rather, a set count somewhere between 10/12 to 20 sets/week provides about the optimal results in all cases, at least within the limitations of the data available.  Only Schoenfeld’s leg data exceeds this but this is from a low volume of 9 compared to 27 with no middle value for comparison.  We can’t know what would have happened between those numbers.

Let me note that even IF you prefer the conclusion that Brad’s highest volumes groups gave a trend towards higher growth, it STILL contradicts the broader body of literature.  He still can’t explain why he needed to use 2X or 4X the volume to achieve the SAME growth as Ostrowski.  He can’t (read: won’t) explain a damn thing, especially when the data disagrees with him.  Let me note again that James Krieger made the explicit point that you have to look at all of the data and not focus on one study.  Yup.  And what all of the data except Brad’s study says is that 10-12 to 20 sets/week is the right number and Brad’s numbers are wrong.  Gotcha, James.  You played yourself, too.

Spoiler: My conclusion above is EXACTLY what Eric concluded as well in his MASS piece, that 10-20 sets/week was about optimal (and of course he did because it is what the MAJORITY of data supports). This was after he desperately tried to make Brad’s numbers and study not be total bullshit by dismissing the endless problems with it methodologically, by playing the “I do science” card and all other manners of silliness.   Which makes you wonder why he tried so hard to defend it with such pitiful arguments and reasoning.  He doesn’t even think the numbers are right or he’d drawn a different conclusion.  And yet he keeps trying to defend it with weaksauce defenses (leg extension volume load hahahahahaha.  I will never stop laughing at this).  I guess when seminar appearances are on the line you have to tow that line…

But this is kind of interesting because it does actually agree with Brad’s original meta-analysis (I am giving him the benefit of the doubt that it’s worth a shit to begin with and I question that with every passing day) which concluded only that 10+ sets gave the best growth response with no ability at the time to determine an upper cap.  So that passes the first reality check.  In all cases, up to 10 sets there is a clear improvement in growth response.  Above that, TO A POINT, there is a greater growth response but it shows a clear cap where higher volumes do NOT generate a greater response.  In most cases, it’s the same, in the case of Haun and triceps, it was worse.  As this is the only study showing a worse response at the highest volumes so no global conclusions can be drawn here in terms of more volume being detrimental.   It simply isn’t any better.

But despite Brad’s attempts to make huge volumes better, the broader body of work (5 of 6 studies) supports a cap of about 20 sets for upper body, if that, and possibly more for the legs (for which we need far more systematic research).  Not 45 sets/week more.  But possibly more than 20.  This goes along with endless anecdotal beliefs that legs need more training volume but more systematized studies need to be done to show where an optimum might fall and whether or not upper and lower body truly have different optimal volume levels in terms of their growth response.

Now I could cut this article here, going a really long way to reach the above conclusion.  But that would be the easy way out and I’m not done yet.  Because I now want to return to an issue I brought up in Part 1 and said I’d revisit.

The Set Count Issue Redux

I want to return to the issue of how sets should be counted that I mentioned in Part 1.  When the Heaselgrave study came out Brad responded with the following to try to make the study fit his conclusions.  Or he may have been talking about both that paper and the modified GVT paper as they are the only two that used isolation movements and he needed to dismiss the fact that they contradicts his results (nevermind that his own study used leg extensions).

He was jumping around a lot and it was tough to tell what he was trying to dismiss to make his own (incorrect) results happen.  Guru speak can be tough since the goalposts change with every post, sometimes including leg extension load volume data when nothing else will work….sorry can’t resist.

His assertion was that perhaps volume requirements are different for compound and isolation movements and that that changed how sets should be counted (which means that his numbers could still be right).   Or rather, that isolation movements should be counted differently (basically trying to dismiss Heaselgrave’s set count accuracy).

And honestly, this is just total guru bullshit in the sense that is is a change in argument when he needed it.  For years now, in every study he’s done and every meta-analysis Brad and his group have counted sets on a 1:1 basis.  They’ve always done training with compounds movements and measured peripheral muscles and treated the total set count for the compounds as applying to those muscles on a 1:1 basis.

It’s always been 1:1.

If someone does 1 set of bench press, that’s 1 set for very muscle involved by bench so one for shoulder and for triceps.  It has to be because even in his own paper he was measuring bicep and tricep size changes in response to compound work.  He wasn’t looking at chest and back thickness (again I question why pec isn’t done more since it is clearly technically possible and wonder if back can be measured).   If you’re going to look at bicep/tricep growth in response to only compound work (and note the odd little bench press study I described that looked at growth in pecs and triceps in response to bench only which means that pec CAN be measured) and count those sets in total, you’re calling it 1:1 for compound and isolation.  Brad and his group have treated it as such from the get go.  It’s also how he reported the ‘findings’ of his recent study too.  He didn’t say it was 30 and 45 sets but this has to be kept in the context of measuring triceps and biceps with only compound training movements.  He said 30 and 45 sets was best for growth with NO qualification whatsoever (well, he didn’t qualify anything until he got backed into a corner).

It’s been 1:1 from the get go for him.

Or it was UNTIL a study/studies came out that he wanted to dismiss. Suddenly, it’s no longer accurate to count it 1:1 which is just terribly convenient. Because you do not after the fact get to decide that the most recent study (or the GVT study) was different due to the isolation work and therefore do not contradict your results.   Even here the argument is totally worthless.  The Heaselgraves study did row, pulldown and curls.  Two compounds and one isolation so at most you count the third exercise differently.  Brad’s leg workout was 2 compounds and one isolation too so what’s the difference except that he doesn’t like data that contradicts him?  The Ostrowski study used isolation work too and Brad was happy to lie about the numbers there without considering set count.  He considered it 1:1 when it was convenient and then lied about the data to change the conclusion on top of it all and then decided it wasn’t 1:1 when it was no longer convenient to do so.

Pure guru shenanigans. The argument changes when it needs to.

I said back in Part 1 that I only used that convention for consistency to what they had been doing and that I didn’t agree with it at face value.  I’ve said the same thing for years.  Brad only said it when he needed it to defend his paper.  But even so, let’s go with this logic and see where it leads us.

Because NOW Brad seems to be arguing that isolation work counts differently towards the training response than compound work.  Presumably it’s worth more since it’s well direct.  That is, nobody is denying that bench works triceps and delts.  The question is to what degree in terms of tension and volume overload it works them and how it compares to direct work in that regard.

Let me add for honesty: in the discussion section of his most recent study, Brad does acknowledge in the limitations that the use of compounds and measurements of isolation might alter the set count conclusion and he used this as an argument for why his change of attitude wasn’t just a convenient excuse.  Which might fly except that this was NEVER brought up until he was backed into a corner and needed to bring it up to dismiss a study that contradicted his.  So he can say all he wants that he considered it but he knows that the majority don’t read the discussion (maybe why he thought he’d get away with his lie about the Ostrowski data).  But when every post he makes crowing about his results IGNORES it, he’s just bullshitting after the fact so far as I’m concerned.  An honest scientist mentions the limitations of their work UP FRONT when they present it whether in research OR PUBLICLY.  Like I have done in this article series for each paper, addressing the potential limitations (small subject number, DEXA only) as I went.  I’m not waiting to get backed into a corner to change my argument or magically find new data (cough cough leg extension load volume…..fucking seriously?)  He did just like James Krieger did.  You can ask me any question about this article series, and nothing I report will change from what I’ve already written unless I made an explicit mistake (which I will then fix).

But let’s go from the assumption (which I have felt from the get-go) that you need to count volume differently for compound and isolation work in terms of determining the growth response to training (note that Eric said this was also true in MASS even if he hemmed and hawed about the ratios involved).  I always have in practice and I’d say anyone with real-world training/coaching experience does as well.  We don’t consider a set of compound chest work to be 100% a triceps exercise (and nobody counts pec deck as a triceps movement although it does involve a little biceps although absolutely nobody counts that).  Most people aren’t even aware that triceps long head is involved in rowing due to it’s function as a shoulder extensor but nobody on the planet would count that towards triceps volume.

It might be conditionally true for trainees with very specific levers who can just bench and build big tris (and even perfectly built benchers do extra triceps work) but we don’t generally count it that way.  Well I don’t and neither does anybody else I know with actual training or coaching experience.  If you look at every workout routine I’ve ever written, total sets of chest (which always includes a compound movement which might be followed by a second compound or an isolation movement depending) are higher than for direct arm work because I’m counting some of the compound chest (or back) work towards arms in terms of daily and weekly totals.   Delts is always funky but I do the same thing kind of.  Shoulders are complicated since it’s three heads with different functions and one is pushing, one is pulling and one is either neither or both depending on you look at it (it’s really humeral abduction but let’s not get too entrenched in this).

How Do We Count Sets?

The question is then how should you count the sets of a compound exercise towards smaller muscles.  I don’t know but I’m going to start with an assumption of a 0.5:1 relationship.  That is, I will count one set of bench or row as one half of a set for triceps or biceps.  Is this right?  Doesn’t really matter, ignore the specifics and follow the logic.  You can math it out for a different value if you’d like.  Call it 1/3rd or 2/3rds.  Call it 3/4ths.   Whatever fits your personal bias. This is my assumption but it’s only that.  Make your own and do the math based on it.  It will only change the specifics but won’t change the general conclusions that come out of the exercise.

Let me note that the ratio must be lower than 1:1 since nobody would EVER count a compound movement as MORE than 1 set for the smaller muscles (ok, I know how people read my articles and someone will make a strawman about poorly done rows being more biceps and that’s fine. Let’s define this as the movement being done properly in a technical sense).  The question is simply how much lower.  Pick your ratio and break out the calculator.

Just don’t call it 1:1 until it’s convenient not to do so like Brad did.

Re-Analyzing the Volume and Hypertrophy Data

Because if Brad is now going to say that compound and isolation movements count differently in terms of sets then EVERY OTHER ONE OF BRAD’S STUDIES AND ALL THE REST have to be recalculated in terms of their effective set count including his original meta-analysis AND his most recent paper and all of the others I’ve examined.   His study used only compound movements (with leg extensions for quads) but looked only at single joint muscles and many studies seem to do this.  If he NOW thinks isolation exercises count differently than his effective bodypart volume changes.  And so do the set counts on every other study (the set counts on his meta-analysis will also change but I don’t know what body of literature it used and what proportion used compounds versus compounds and isolations).

So let me recalculate them based on my assumption that a compound set counts as 1/2 a set for smaller muscles and a  direct exercise counts as a full set (i.e. bench press is 0.5 sets for triceps, triceps extension is 1 set for triceps).  Again, this is my starting assumption and nothing more.  Use whatever ratio makes you happy so long as it’s less than 1:1.  And don’t play silly buggers at the extremes, we all know it’s not 0.9:1 to one or 0.1:1.  It’s probably not 1/4 to one or 4/5ths to 1 either but somewhere clustered around the middle depending on the movement.  Maybe 1/3, 1/2, 2/3, 3/4…again I don’t know for sure and nobody else does either. So I’m using 1/2 and literally splitting the middle.

So all I did was go back and count the sets.  If it was compound exercises, I cut the number of sets in half (10 sets per week becomes 5).  If isolation, I didn’t (5 sets per week = 5 sets per week).  If there was a mix I counted compounds as half and isolations as one and added them (10 sets compound chest = 5 sets + 5 sets isolation triceps = 10 total sets for triceps, down from 15 originally).  Only two used a mixture so I probably didn’t screw the math up too badly.  Even if I got a number slightly wrong it doesn’t change the overall conclusions.

Now let me be clear again, I am NOT saying that this is a perfect analysis or that a 0.5:1 estimate is right so spare me the strawman arguments that I’m trying to force a set of data.  I’m simply saying that if Brad is going to dismiss a study result he doesn’t like based on isolation vs. compound needing to be counted differently, that opens the door for this type of analysis.  Beating a (very) dead horse, you can redo it assuming compound is worth 2/3rd of set of 3/4.  The numbers will change slightly and that’s fine and they’ll be marginally higher than my 1/2 assumption. If you go with 1/3rd they will be marginally lower.  But for rational set counts, the differences aren’t even that much.  Focus on the principles, not the specifics, folks.

But they will ALL go DOWN from what Brad was claiming them to be originally.  EVERY SINGLE STUDY.  That means in his meta, in his own study and in his examination of previous studies the numbers will all decrease.  Even in Ostrowski which he lied about in his discussion.  All of the set counts decrease.  None of the studies I examined used only isolation movements so there is NO situation where the numbers don’t go down.  And since nobody would ever count a compound movement as MORE that one set, they can’t EVER go up.

Yes I am beating a dead horse but I know how people read my articles. At least one person will claim “Lyle said all compound movements are worth one half a set for the other muscles involved” which I am not saying in the least.  I am saying this is my working assumption for lack of a better one and that’s all it is.  Again, use your own number that is lower than 1:1.  Just follow the logic here.

Yes, we need data on how to compare the exercises to see what the best counting approach would be.  I am aware of one that compared pulldown to biceps curls for recovery and the biceps curls took longer to recover from than the pulldowns so clearly it’s NOT 1:1.  The pulldowns didn’t hit the biceps as much as direct arm work did.  Dadoi.  Until more data exists, we make assumptions.

It might even and probably will turn out that different movements should be counted differently.  An undergrip pulldown is more biceps involvement than overgrip and a parallel grip is halfway in between (with more brachiails).  A high bar squat is more quads than low bar and a close grip bench is close to an compound triceps exercise but a flared elbow bench is more pec specific and how you’d count those towards triceps would likely differ (I’d call close grip almost 1:1 for triceps but flared elbow as 0.5:1), etc.  Back gets super complicated as we’re dealing with the traps (with multiple sections), rhomboids, teres, lats (which have two segments with slightly different orientations) and back movements work them to varying degrees based on movement, grip, bar, etc.

Coaches make adjustments for this based on experience.  If I were using overgrip pulldowns with a trainee, I’d give them slightly more direct biceps work to compensate compared to if they were doing undergrip pulldowns which would have worked the biceps more. If they did V-bar rows, I’d make adjustments to biceps compared to doing an undergrip row.  I’d give a low bar squatter who sits back more direct quad work than one squatting high bar for example.  Everybody with any real-world experience does this in practice to some degree.  We do this based on 1/2 guesswork,  1/2 experience, 1/2 science, and 1/2 intuition (and sometimes1/2 luck).

Again, let’s not get too mired in the specifics here (as I do that very thing).  Follow the logic.

Rebuilding the Model: Part 1

And with my assumption of a 0.5:1 relationship, here are the re-mathed set counts for each of the 6 studies I’ve included. I’ve shown the original set counts in parentheses next to the re-mathed value and I probably messed at least one of these up because math is hard, my brain is tired, and I don’t bother to run it twice.  And this is total sets per week.  I’ve indicated in red which group did best (based on the analysis above) and might have even gotten it mostly right.  I am quite sure that anybody wishing to dismiss my conclusions based on a single typo will make me aware of that typo and I will change it because that’s the intellectually honest thing to do.

Study Muscle Low Moderate High
Ostrowski Lower 2 (3) 4 (6) 8 (12)
Upper 5 (7) 8 (14) 20 (28)
Amirlingtham/Hackett* Triceps 10-11 (16-18) 16 (26-28)
  Quads 7-8.5 (11-13)  8.5-9 (16-18)
Huan Increasing from 5-16 sets with a cap on growth at 10 sets for upper and possibly more (up to 16) for lower.
Schoenfeld Upper 3 (6) 9 (18) 15(30)
Lower 6 (9) 18 (27) 30 (45)
Haeselgrave 6 (9) 12 (18) 18.5 (27)

*I grouped the two studies together since they were an identical methodology and only differed by length
and I’m getting tired of writing and making tables as it’s a pain in the ass in WordPress.

And this brings the results into even starker view.  Ostrowski fits with all other data which shows a clear dose response relationship up to 10 sets for legs although the lack of data above 8 sets limits this finding and we can’t know if more would generate more growth. For upper, 20 sets (down from 28) wasn’t better than 8 down from 14.   Since 8 and 20 got the same growth, it seems unlikely that a middle value would get different results although it’s possible that 14 would but 20 was too much for some reason.  Without data, this is a guess and we need studies examining different intermediate values to know for sure.  Test 8, 14 and 20 next time.

For the GVT studies 10-11 sets per week as a mix of compounds and isolation was as good as 16 sets/week for upper body.  Inasmuch as the differences were miniscule, 8.5-9 sets per week for legs was better than 7.5-8.5 but at this point, we’re looking at a single set difference when it’s re-mathed and that alone would explain why the results were essentially identical (the stimulus was essentially identical).  You wouldn’t expect 1 set to matter but maybe if it were 7-8.5 vs. 15-16 it would. That group has already done the same study twice, now do it a third time with real differences in lower body volumes (give them a second leg day).

Schoenfeld’s data becomes a lot less idiotic now and at least starts to pass the reality check, in line with the other studies.  9 sets was as good as 15 in terms of triceps growth (because his stats did NOT show that the highest volume was more than insignificantly superior to the moderate).  Even if you believe that his highest volume was superior, it’s cut to a realistic 15 sets per week from an absolutely moronic 30 sets per week.  This starts to fit the reality check and is still well within the realm of 10-20 sets.  Like I said, big picture whether you accept my contention that moderate was as good as high or potential trend for highest to be superior, when you count the sets rationally, it stops mattering, at least for upper body where both moderate and high fall within 10-20 sets/week.

We still lack data on chest growth per se and it might require more volume or it might not so whether or not the original values matter is unknown (i.e. does the chest somehow need 30 sets of direct work…I doubt it). Until it’s measured, we don’t know.  No study can address that yet  and the bench press only study I referenced in Part 1 didn’t compare different volumes although it would be hard to see how the 9 sets that worked for 6 months suddenly needed to be tripled after one year.   But he doesn’t get to count chest volume and then measure triceps to draw conclusions about optimal sets for all muscle groups (which he essentially did) and then decide that you have to count volume differently for isolation exercises after the fact (which he actually did).

For lower body the 18 sets was as good as 30, again passing the reality check.   Here, if you take his higher volume claims as better that’s a pretty high set count (30 sets/week) although there might very well be a plateau value between 18 and 30 sets (we don’t know) which would be consistent with Haun (maybe) and anecdote.  Maybe.   If more than 20 sets IS optimal for legs (and this is still in the IF stage), a third group at 24 sets might have done better.  Testing 18 vs. 24 vs. 30 sets would be very informative but it has to be a lab that isn’t Brads.  The stats and strength gains still don’t support it and the fact that he lied about data should make his study inadmissible on fundamental grounds.

There’s still that pesky ECW issue to worry about above 20 sets per week which now ONLY the highest volume leg work in Brad’s study crosses (maybe that explains the almost significantly higher leg extension load volume.  Hahahahaha.  I’m never gonna stop laughing at that shit).  Then again, Haun was using pure compounds so that probably doesn’t make any sense as I think about it since I’m now comparing a compound only study to remathed sets.  So yeah, forget that bit, it’s wrong.  Based on initial volume, several groups in Schoenfeld cross 20 sets/week.  And that means ECW might be playing a role or artificially increasing the results.  I’d only note that with a spread of 18 to 30 sets, we don’t know if a middle value (i.e. 24 sets) would be superior until it’s directly tested.  Finally is Haeselgrave which found that 12 sets was better than 6 but no better than 18.5.

Rebuilding the Model: Part 2

Recreating the chart from above with the new numbers we get the following optimal volumes per week.

Study Optimal Volume
Ostrowski 8 sets lower body (no higher tested), 8 sets upper body
Amirthalingam/Hackett* 10-11 sets upper, 8.5-9 sets lower (the huge drop in set count is due to the single leg day)
Huan Increasing from 5-16 sets with a cap on growth at 10 sets for upper and up to 16 for lower.
Schoenfeld 9 sets for upper body, 18 sets for lower body (15 and 30 if you accept the highest volumes)
Heaselgrave 12 better than 6 but 18.5 no better than 12

So we get systematically lower numbers here, as expected.    And again, if you disagree with my 0.5:1 and use a different value, the numbers change slightly but they still all go down (i.e. if you use 3/4:1 Ostrowski’s leg data might be 10 sets instead of the original 12 or my 0.5:1 assumption 8) Basically, for any moderate set count, the differences in remathed sets just isn’t that significant.  I mean, consider a group that did 10 sets/week of compound.  If I assume 0.5:1 that goes to 5 sets.  Assume 1/3rd and it goes to 3.  Assume 2/3rds and it goes to 6.  Assume 3/4 and it goes to 7.5 or whatever and we are looking at a 4 set spread.  Use the lower ratio and it’s a little lower, use a higher ratio and it’s a little higher. And it all more or less stays in the same overall range we’re looking at here.

But it’s ALWAYS less than the original value which is my point here.

Looking at the new numbers, Ostrowski’s upper body optimal volume is 8 sets/week.  Lower body matches 8 sets/week with no higher values tested so we can’t know what would happen above that.   The two GVT studies are 10-11 sets for upper and 8.5-9 sets for lower with no higher volumes of lower body tested.  Haun finds a cap on upper body of 10 sets (down from 20) and possibly up to 16 for lower (down from 32).  Brad’s number stop being totally moronic when you don’t count them in an ass-backwards way with 9 sets for upper and 18 sets for lower body, generally matching the results of Haun.  Heaselgrave is at 12 sets for triceps but 18.5 was no better.  And all of this basically agrees with the original 10+ set meta-analysis even remathed (though it’s conclusions should probably change if it is remathed) except that we now have a much better idea of the upper caps on weekly set volume.  There’s a dearth of leg data at higher volumes and more study is needed here.

So my next to final comment: whether you look at the original unadjusted data or the semi-adjusted data for set count, you still see a general optimal range of 10-20 (original count)/8-16 (remathed count) sets per week per muscle group which is all close enough for government work.   Let’s just call it 8-20 sets/week and move on with our lives.  And again, this is consistent with Eric Helm’s own conclusions in MASS of 10-20 sets/week after his pitiful defense of Brad’s paper.

There is still the slight indication that *maybe* more for sets/week for legs would be better but it’s understudied and any conclusions would be tentative as hell.  But there is no way on god’s green earth to justify the 30 and 45 sets Schoenfeld et. al. is so desperate to prove as optimal.  His own data doesn’t support it, his stats don’t support it, the bullshit apologism by everyone involved doesn’t support it, a rational re-analysis of the set count doesn’t support and neither does the broad body of literature, his lie about the Ostrowski data notwithstanding, support it.  Nothing supports it except his burning desire to support it with guru games and because he’s believed all along in these types of high volumes.

Now We Refine the Model

Because like I said, this is how science works: you take all the available data and you make a model.  You don’t fixate on individual studies and it’s the overall body of literature that is relevant (again, I thank James Krieger for making my point for me on this).   And ignoring Radaelli, I have presented 5 studies showing that, in the aggregate, moderates volumes somewhere between 8-20 sets per week provide the maximal growth response and 1 that fails the reality check so hard it hurts, where data was lied about in the discussion (a fact that NOBODY has yet to address directly for me) and which should be dismissed on that fact alone.  Legs maybe need a bit higher but we need more data.

Now, it’s possible that more work will change that and I’ll change my model and opinion when and if they do.  But unless we do find out that Ultrasound doesn’t measure muscle growth or something and all of these studies go into the junk pile I won’t hold my breath.  They match one another despite varying methodologies and they match (for what it’s worth) with real-world training practices.  They pass the reality check is what I’m saying.  At best we’ll refine the above numbers with more targeted research.

That is, future research might start from the idea that 10 to 20 sets/week is optimal and determine what specific volume is optimal within that range.  Or more systematically compare lower body and upper body.  Perhaps look at 10,15,20 for upper body exercises and 15,20,25 for lower. But stop doing 9,18,30 where the variance is just too huge to know what happens in the middle ranges.  Is 12 the same as 18, is 24 the same as 30?  Stop focusing on sets per exercise.  If you want to test sets per week, set it up to do that in a rational way.

Feel free to contact me with help with the study design.  I can also probably figure out how to pre-register the study, describe the randomization in the methods (and randomize the subjects in a blinded fashion to do the Ultrasound) and efficiently write the discussion with accurate data representation for anybody who just doesn’t have time…..

But expectationally based on the broad body of literature, optimal results are likely to be found between 8-10 to 20 sets per muscle group per week.  Once again, Eric drew this same conclusion after his desperate efforts to make Brad’s paper not be shit.   And I just read something by Bret Contreras of all people saying to stop doing insane volumes and focus on intensity.  Good lord, when Bret is the rational one and Mike Isratel is the intellectually honest one in all of this, Mercury must be in retrograde.  But Bret was on Brad’s paper and he better be careful or Brad will kick him out of the paper publishing circle jerk or prevent him from getting seminar appearances for not towing the party line.

Ok, two more comments and I’m done.

The Generic Bulking Routine

With the above in mind, a rough volume of perhaps 10-20 (or 8-16 depending on the analysis) sets per week as an optimal growth number, I want to look at what I have presented for years as my Generic Bulking Routine.  This was an intermediate program I drew up absolute ages ago that has proven to work for intermediates for over a decade.  I report this only anecdotally and nothing more, I’m not James Krieger who thinks anecdote counts as ‘science’.  But if we’re going to pretend to integrate science and practice, then it is always nice when practice actually matches up with the science.

It was an Upper/Lower routine done 4 times per week with each day having the general structure shown below and was meant to be done as 2 weeks of a submaximal run up and then 6 weeks of trying to make progressive weight increase (progressive tension overload being the PRIMARY driver on growth with sufficient volume within being optimal) prior to backcycling the weights and starting over with the goal of ending up stronger over time.   It was mean to be an intermediate program used from about the 1-1.5 year mark of consistent training to maybe 3 year mark before my specialization routines were implemented.

Weights didn’t HAVE to be increased every week or workout, that was simply the goal (as Dante Trudell put it in his Doggcrapp system, you should be trying to beat the log book at each workout).  In my experience, so long as folks were eating and recovering well and started submaximally, they could do so over relatively short time periods like this (over a longer training cycle, I’d do different things).  Women perhaps less so than men for unrelated reasons but no matter, Volume 2 is coming eventually…

I’m going to provide specific exercises in the template but just think of them as either compound or isolation for the muscles involved since exercise selection is highly individually dependent. RI is rest interval and note that I use fairly long ones so ensure quality of training with real weights and ideally all sets are at the same heavy weight (oh yeah, in ‘Merkun a single apostrophe is minutes and a double is seconds and I am told this is the opposite of the rest of the world).  Big compound movements get 3 minutes, smaller muscles get 2 minutes and high rep work gets 90 seconds since it’s meant to be more of a fatigue stimulus to begin with.

But for big movements, a 90 second rest interval is bullshit and means that you’re probably squatting with 95 lbs on the bar by your fifth set ‘to failure’.  Better to do less sets and give yourself long enough to do quality work.  In that vein, after the submax run up, the goal RIR was maybe 2-3 for the initial set which would likely drop to 1 or even near failure by the last set.  The goal is progressive TENSION overload over time (meaning multiple training cycles).  When your workouts don’t use stupid volumes, you’re in the gym the same amount of time but can actually do quality work than when you’re trying to fit in 45 fucking sets and get done before tomorrow.

Upper SetsXReps(RI) Lower SetsXReps(RI)
Flat Bench 3-4X6-8(3′) Squat 3-4X6-8 (3′)
Row 3-4X6-8 (3′) RDL or Leg Curl 3-4X6-8 (3′)
Incline DB Bench 2-3X10-12 (2′) Leg Press 2-3X10-12 (2′)
Pulldown 2-3X10-12 (2′) Another leg curl 2-3X10-12 (2′)
Lateral Raise 3-4X6-8 (3′) Calf Raise 3-4X6-8 (3′)
Rear delt 3-4X6-8 (3′) Seated Calf Raise 2-3X10-12 (2′)
Direct Triceps 1-2X12-15 (90″) Abs Couple of heavy sets
Direct Biceps 1-2X12-15 (90″) Low Back Couple of heavy sets

The exercises are for example only and the other two workouts per week could be a repeat of the same movement or different within that general structure (exercises can be also be changed with each succeeding training cycle).  So start with incline bench and pulldown for the sets of 6-8 and do flat bench and row for the sets of 10-12 or whatever.

Now let’s add up the set count:

Compound Chest: 5-7 sets twice/week for 10-14 sets/week (counted as 5-7 sets for tris at 0.5:1)
Compound Back: 5-7 sets twice/week for 10-14 sets/week (counted as 5-7 sets for bis at 0.5:1)
Side delts: 6-8 sets per week.
Rear delts: 6-8 sets per week (gets hit somewhat by pulling but hard to math out)

Note: Effective delt volume is likely a bit higher than this but it’s a pain in the ass to estimate how much side delts do or do not get hit by compound pushing.  Or rear delts via compound pulling.  It might math out to 8-10 sets/week or less or maybe more.  Again, hard to say but most report just fine delt growth from the above (and no shoulder problems which is why it’s an upper/lower to begin with).

Bis/Tris: 1-2 direct sets added to compound work = 2-4 sets/week + 5-7 indirect sets/week = 7-11 sets/week.  Add a third or even fourth set if you like to get 8-12 sets/week or 10-14 sets/week of combined indirect and direct arm work.  I certainly agree that if you do heavy pushing and pulling you don’t need a lot of warm work.  I simply do NOT agree that the sets count 1:1.  But my workout designs usually have proportionally less direct arm work since I partially count the compound pushing/pulling and always have and always will.

Let me comment before moving forwards that while this might seem like a low per workout volume to some (and high to others), it matches the set count data based on my analysis above.  As well, I have contended for years that if you can’t get a proper stimulus to your muscles with that number of sets, volume is not the problem.  Rather, you are.  Whether it’s due to suboptimal intensity, focus, technique sucking, etc. you are the problem with your workout.  Doing more crap sets will never top doing a moderate amounts of GOOD sets.

Regardless, looking at it now, with 15 years of experience with it and the data analysis I just did, I might bump up the side delt volume a bit.   As noted above, the contribution from chest work is tough to really establish here and the delt has three heads with differing functions.  But no matter.  Let’s focus on generalities.  Which are that my general set count for this workout is and has always been right in the range of what the analysis of the majority of the training studies found to be optimal.

This template could be adjusted in various ways.  The second chest and back movement could be isolation which would reduce the indirect set count on arms, necessitating an increase in direct work.   So if someone did 4 sets of flat bench and 3 sets of incline flye that’s still 14 sets/week for chest but reduces indirect arm work to only 4 sets/week (8 sets of compound pushing divided by 2) so you bump direct arms to 3-4 sets per workout to get to 6-8 direct sets per week and 4 indirect sets for 10-14 sets per week.  I think that makes sense.  The point being that I am looking at total set counts per week (actually I was counting reps but it all evens out) and adjusting volumes for smaller bodyparts based on exercise selection.  If you use more isolation movements for chest or back that decreases the indirect set count for bis and tris so I’d add more direct work there.

The same holds for legs where quads are worked for 5-7 sets twice weekly or 10-14 sets, same for hams and calves.  I might bump this up slightly although high volumes of truly HEAVY leg work is pretty brutal, add a third movement like leg extension and another leg curl to for a couple of higher rep (12-15 rep) sets apiece.  Now it’s 7-9 sets twice a week or 14-18 sets.  Towards the higher end of volume but until we know for sure that it’s 20+, I’m not changing much here.  And, again, a workout with 20+ heavy sets of legs (including quads, hams and calves) is gruelling.

But overall upper body comes in at somewhere between 7-14 sets for upper body muscles and 10-14 for legs.  Again, intermediate program from like 1.5-3 years or so.

Those numbers look so very familiar.

The Wernbom Meta-Analysis

Let me finish by revisiting the original Wernbom analysis that looked at intensity, volume and frequency in terms of optimal growth.  It’s become pretty fashionable these days to dump on it for various reasons.  It’s fairly old, there is more data now and there was simply very little work done on intermediate much less advanced trainees at the time.

Irrespective of that, within moderate intensities (the typical ‘hypertrophy zone of perhaps 70-85% 1RM), it concluded that a volume of 40-70 repetitions twice/weekly was optimal for growth with triceps and quads being the muscles of interest.   I honestly think using reps per week is a better approach than sets since obviously 10 sets of 1 and 10 sets of 10 are not the same stimulus.  That said, since almost all of work on this topic stays in the 8-12 range or so, set counts are at least conditionally appropriate.  Within any rationally accepted repetition range, it just all sort of balances out.

If you add up the reps on my GBR you can see where my numbers come from. I use a combination of heavy 6-8’s for tension and 10-12 or 12-15 for more fatigue which is why I mix them but you end up with roughly that number of reps for every muscle group (you can count reps on compound chest/back/legs as half the reps for arms but it should all math out more or less correctly because that’s how I set it up).

No Wernbom wasn’t on well trained subjects but none of the above studies used elite guys either because a 1.1 bodyweight bench is not elite in men, it’s advanced noob.  Wernbom was basing on a limited data set in, at best, limited work on even intermediate trainees (again, just like the above studies) and still concluded 40-70 contractions twice a week gave optimal growth compared to lower and higher values.

So we double 40-70 and that’s 80-140 repetitions per week per muscle group.   Some quick maths.

At 10 reps per set 80-140 reps per week yields 8-14 sets per week.
At 8 reps per set 80-140 reps per week yields 10-16 sets per week.
A mix of 4X8 (32 reps) and 3×12-15 (36-45) for 68-77 reps per week is 14 sets/week.
A mix of say 5X5 (25 reps) and 3-4X1012 (30-36 reps) for 55-71 reps twice a week is 16-18 sets/week.

So for any rational workout design an optimal repetition count of 40-70 reps/workout done twice per week for 80-140 total reps per week put us somewhere in the realm of 8-18 sets/week for the optimal growth response.

Well whaddya know about that?