Training to Failure Might Cause More Growth Than We Thought!?

if you kind of go back and look at all the the categorical meta analyses that we talked about those all kind of painted the picture that there's nothing to gain from training closer to failure if you look at those a little bit closer pretty much every single one of those effect sizes does lean in favor of the failure condition you were heavily involved with I guess it's a re a review metaanalysis which is a a review of the literature but it goes about things in a mathematical way so you're not just kind of eyeballing all

the studies and trying to put them together but statistics helps you do that in a way that's more objective I guess uh that's that's the idea yes met analyses have suffered a few critiques but they you know if done right they're excellent and so the metaanalysis was um for hypertrophy and strength and basically integrated the concepts of proximity to failure and dose response that's a whole lot of fucking brainy words I don't understand so what um what did you do in this meta analysis what was kind of the goal and uh what basic methods did

you use in super plain terms ABS um and then um what did you find yeah so throughout my masters I started getting interested in this topic and so the the topic of proximity failure is what I did my Master's thesis on um but throughout that process and reviewing the literature and also reviewing the available met analysis that had been done to that point pretty much everything was analyzing things in what's just called a categorical fashion or a versus B failure versus non-failure and as we just described with the concept of RA being so prevalent in

training for hypertrophy and training for strength at this point obviously power has been doing that for a really long time I knew there had to be a way that we could analyze that a little bit more granularly and understand the the dose response relationship like you kind of mentioned so just a basic way to kind of understand that is if we have some sort of uh pain relief drug or Neds or Tylenol or something like that um typically the greater the dose of that medication that you're ingesting the greater effect you're going to see but

those but those relationships can look a lot of different ways right you could have strongly diminishing returns after you reach a moderate dose of the medication maybe the effects really taper off maybe it's a perfectly linear effect there's a bunch of different ways accelerating curve yeah exactly exponential those kind of things those could all look very very differently and so the goal of the project was ultimately to say hey let's go through all these studies do our best to estimate the number of pro r that they were training with and then let's see what we

can find in terms of these do response relationships for both hypert and strength so basically what you're saying is it's totally fine from a scientific perspective to look at the problem and go people who didn't train to failure versus people who did or whatever conditions where various people did or didn't train to failure is cool for some kind of insight yeah but if you kind of lens into the problem you go actually okay so some of these groups did one rep to failure some did four reps to failure and like we're grouping that in the

same group of didn't go to failure if we actually look and see okay what's the difference between roughly four and three roughly three and two roughly two and one roughly one and zero and then we can determine like oh okay the closer you get to failure XYZ happens that's basically what you're doing first study of its kind to do that fucking awesome yeah is it it's I I'm not going to oversell it too much in the sense that n fuck it oversell it these people will buy anything they they buy our app it's a fucking

dog shit app yeah it there there's definitely some one things that say on the front end I guess is that because the reporting of studies is pretty challenging to um you know very few studies actually are going to report the average number of re that are participant trained with just as a basic example we had to go through each kind of bucket of different studies we can talk about that if you'd like and ultimately estimate the re that they train with now I wouldn't have done it if I don't think it's like pretty decent estimations

but of course those come with some some limitations in terms of we didn't actually know what they were training with but we did the best to go through a pretty systematic series of estimations to get those actual numbers but once you get those numbers then we can kind of play with them in ways that people haven't really been able to before yes I guess um there's some imprecision in estimating R from just reading a manuscript but at the same time I'll say it uh nice and nice and sarcastically or cynically if you think science is

anything other than incrementally more educated guesses uh you must be doing religion or some shit like that cuz it's all guesses all the way down but they um the more reasonable the guesses get and the more times you examine some phenomenon and it basically yields the same result you're like oh okay I'm more confident like you know the philosophers will be like like are you really there you're like I don't know 100% but like I I um I hit my uh hand once with a hammer and fuck that hurt man that t really me so

I'm going to bet like 99.999% so the confidence we have that your methods for assessing how close to failure people were it's not like dude this is 100% the revealed truth it's like probably no almost certainly better than not having that analysis exactly a bit more information yeah I like to think about it it's almost yeah I almost use almost certainly too it's probably like directionally accurate as the way that I like to think about it um and we'll talk about that here in a second in terms of like the exact RS I'm less interested

in saying exact claims that way but the directional aspect of things that's what that's what the analysis gives us whereas otherwise if we are making that failure non- failure comparison if you just think about it at a as a really basic level and probably just something you've said before is like okay failure versus all the conditions we training with two R that'd be a very different comparison than failure versus all the condition training at five reps in reserve and so those those analyses tell very different stories but without able to actually examine that dose response

relationship along the r Continuum yes our confidence and the conclusions that we can draw are just very very different right yes so when you say directional uh what I think you mean correct me if I'm wrong and by analogy is like if you you're you got like 10 guys in a room that you're looking at you know my Tuesday night I would say more than looking at first looking of course just dive in I don't know about respect at first respectfully and then not respectfully um and you don't have a scale uh you can't weigh

anyone but you can arrange them by rough body size so you don't know exactly what anyone weighs but like the guy who looks like Brian Shaw is probably bigger than the guy who looks like Harry Potter and the average dude in between now you have like three factors you're like someone's like well how do you the Brian Shaw guy beats up the little Harry Potter guy and you're like I think combat performance is related to size like well how much did he weigh you don't know like yeah we don't know exactly what he weighed yep

so with with train of failure we don't know exactly if it's one r or two but we know the conditions you met out the one R subcondition they they probably trained closer to failure than the 2 R almost certainly than the three R and so we can get a rough idea of what's going on beautiful beautiful analogy oh stop that's great fantastic we pay him off camera to say Scott pulls everyone aside that comes to the studio like Mike loves compliments let me put that another way if he doesn't get them he's like broken an

iPad before and stormed out so like if you want this to go well just every every 5 minutes and it can be totally random you could just interrup your own train of thought and be like Mike I love your sweatpants I'm like isn't that how Donald Trump works all right we're getting cancelled no Donald Trump doesn't cancel people fuck it it's all jokes all right so um meat and potatoes um not application just yet because we'll get that in a bit but what did you what did you find in your analysis yeah so ultimately like

we said we kind of went through things in a lot of different ways and there's kind of a lot of different offshoots that kind of color the main findings But ultimately the the biggest like headline findings were for hypertrophy we basically saw a linear best fit relationship that ultimately Trends as you are getting closer to failure the point of essentially the stimulus per set is like the way that I like to think about it ultimately improves incrementally in a linear fashion whereas for strength there was essentially no relationship and and important thing here is that

both of our kind of primary uh analyses you can think about them at a given percentage of one RM or a given load on the bar or a given um rep max whatever kind of term you want to use there and that's important for strength because um if I have a set of five let's say and I'm getting closer to failure I'm also increasing the percentage of one RM that I'm lifting and we wanted to kind of take that out of the equation so we wanted to see the actual influence of proximity to failure directly

and so hypertrophy tends to improve per set for strength essentially no relationship so basically what that means is your analysis found roughly speaking that if I'm interested in getting at least one session's worth of any kind of workout and maybe even a few weeks worth or a few months worth of gains the closer I get to failure in my sets at least in that analysis the more likely I am to get more hypertrophy in a way that seems pretty linear so like more proximity to failure is generally better for growth and then for strength the

relationship you guys found and I'm sure you didn't analyze like a bunch of studies at rp2 yeah it was a lot a lot across the Spectrum for strength just because strength is obviously a little bit different in terms of what the programs are designed for but yeah it's essentially the the the model was essentially entirely flat which just means that load is probably much more indicative of strength gains than proximity failure directly which makes total sense principal specificity and whatever we get out of failure in strength um even if it comes with an extra hypertrophy

stimulus which we are to assume it probably does somehow gets canceled out somewhere by something so that uh we do not see that linear because people would say like okay so hold on strength is largely based in size you're okay that's true in the longer term especially was say in the context of training studies maybe that's a little bit different for sure but also like you expect the guys going close to failure to probably grow somewhat more muscle in a strength training study too and that's probably true so how come they didn't get stronger and

then we're talk about Med factors and all that stuff but um okay so that makes sense linearly now um as far as nuances yeah uh we had an excellent conversation on the revive stronger podcast about that sort of thing I might have been one of the only people on the internet quote unquote to read the study it's kind of strange and people like what do you think about the new study I'm like I haven't read it yet then I read it I'm like oh I have a lot of thoughts about the study but one of

the things that really really impressed me was you guys did a whole sub analysis of all the kind of mediating factors and what I got out of it and please tell me if this is mistaken is as you ramp up the number of sets per week as you ramp up the number of sets per session as you ramp up the tendency of the study participants this is an odd way to take a a two-factor variable more male um and as you ramp up I think their training experience the benefits of going close to failure start

to substantially dilute yeah so this was so we actually had two versions of the the uh met analysis in preprints which is basically we put an early version of the manuscript out getting a ton of feedback and make some adjustments prior to publication it's now been published um this was one of the things that we Sly adjusted the way that we did it all of what the statements you just said are directionally accurate but in terms of you know the the traditional threshold of significant or meaningful things like that they didn't quite cross that for

pretty much any of the the moderator analyses where we look for okay what what factors in the different studies are ultimately maybe panning this picture a little bit differently so like you said all of the the training volume the kind of all the things you would expect that make the training program in general a little bit more on the advanced side of things harder more stimulative all directionally do do kind of play that out and I think uh just calling on some other research that kind of helps that's come out since uh Martin Ruffalo had

an awesome dissertation project that was uh performing you know relatively High training volumes with relatively Advanced people and there wasn't a clear distinction between um you know failure and about two reps in reserve and relatively Advanced folks so I think that relatively makes sense we can kind of talk into that a little bit more detail if we want but I think the way that I view it even alongside those factors is it's still kind of a principle is the way that I really like to think about it rather than an actual um an applied outcome

that applies in every single scenario so if we take the premise that on a per set basis a set is more stimulative as you get closer to failure and probably at the point of failure and maybe even Beyond with some of the you know the lengthen uh training stuff and we can go into that as well to me I just use that in kind of my my my mindset when I'm planning a training plan but that doesn't mean that two R wasn't extremely effective and maybe in certain scenarios is is is valuable for other reasons

and obviously you're kind of hinting at those um in the case of a higher volume training program okay conceivably there's a trade-off between doing every single set to failure and not being able to perform as many sets in the week so if those kind of things you have to start playing with these things but that still can mean that training to failure maybe on a per set basis is still more stimulative but if I'm playing like a game like the price is right let's say and I have to meet a a doll3 and I don't

want to go over maybe a set to failures like playing with quarters and I might eventually go to a125 when I have to play with some nickels and dimes to really max out the amount of stimulus that I could handle within a week that's kind of how I think about it if that makes sense that makes a ton of sense I was actually a contestant on the prices right one time really yeah and unfortunately it never aired the episode because Scott uh remembers this happening um they asked me to estimate a price for something and

um I didn't see this clause in the contract so my eyes rolled back in my head I reached out my hands and I accessed the the the Judaism force and the camera started flick the light shut down a couple of the studio audience they screamed and ran off uh I like I ended up floating in the middle of the room electric arcs everywhere my eyes were like Raiden white you know and uh I I predicted and guessed every price exactly uh to the fractional scent um and they they said that was cheating or something at

uh just being me you know I actually don't remember the incident I'll just remember streams of green dollar that's all I it's it's a religious experience um so that Ridiculousness aside what I always try to kind of shift back away from direct results of studies and go up a hierarchy of understanding and try to integrate the study results into Concepts and so to me when your guys's study came out it fit real nicely with a a concept um I described to you and you just tell me like that I don't know if that supports that

or maybe that's wrong maybe it's right but this isn't our shit so it seems like training stimulus is analogous to filling a glass and if you um have training very close to failure every time you pour a a lot goes in if you have training very far from failure every pour is less and so if you have not too many sets a week like let's say you have someone who's like I can train for 30 minutes twice a week you're going to be like okay so we're not worried about fatigue a ton because the hell

are you going to do to yourself to make yourself super chronically fatigued from that and we need as much stimulus as possible just don't have a lot of time to do it anything other than damn near failure is like you just misunderstood the problem it's like I'm I have two seconds for you to fill my cup and I'm really thirsty going go pour a ton of shit in whereas if you have a program that's six days tons of volume if you uh go everything to failure you can overfill the cup potentially uh or get to

a point where you're like filled it halfway through the week and the rest is like well where does the orange juice go at like like fucks up your laptop that's sitting under the cup all this other stuff so is that cup filling analogy of like the more intense the stimulus the more powerful it is but potentially the less of it you need to get to where you're going something like what's happening yeah I think I think so I think that's obviously we don't have uh concrete answers to literally every permutation of this problem but I

think from you know combining practitioner experience with the data that we have to me that's a very very reasonable kind of mental model to operate from every set is a little bit more stimulative we might talk a little bit why that may have a cost you want to consider but even just from a practical sense like anyone knows if you go in and try to do 20 to 30 sets of a muscle group where every single thing is the failure that's a very different subjective experience than going in doing the same thing with two or

three reps in reserve now is there does it have to be one versus the other is there a combination there all those things are up for um you know practical discussion that probably you can find a combination of things but ultimately to me understanding these principles from a directional perspective and we can kind of paint all these basic training variables from that lens then you just have the optimization problem of like okay from these dose respons relationships that we do have from proximity to failure from volume from load or rep range um let's create this

optimization problem to get the greatest net kind of effect for the desired outcome in this case hypertrophy so I think that's exactly kind of how I think about it is uh you can apply that to pretty much all training variables and you just have to pick out the ones that maybe are the most uh have the strongest effect and and proximity failure seems to have a pretty decent effect but compared to some other variables like volume it may be one that is worth trading off some of the time volume is King so basically what you're

saying is if I think about okay I am interested in the Max Muscle results and I'm read The Meta metanalysis and I'm like uh didn't get to the discussion at all I didn't read at all I just read the title and um I just crank I want the most growth it I should know that for any given set of training the closer it is to failure the more that set grows me and it could come with other things we'll talk about later like fatigue but at the very least I should know that yeah in the

same way that I know that the velocity of an object that's headed to somewhere is like a big determining factor of how soon the object will head somewhere but if I'm assuming that's the only variable maybe that could get a little muddled cuz like it's technically true that if you have a Lamborghini and as everyone knows I have hundreds at this point um if I take my Lambo to the store and I floor it the entire time on a first principal's analysis considering only one principal yes I'm more likely to get to the store sooner

but if I crash into a tree because I forgot the variable of velocity trades off with a variable of vehicle control then I'm like okay so given that I want to go as fast as I can but in the context of I still have to slow down for turns it's still super important to know that like look like getting there faster is like you you got to step on the gas some people like you got to be careful and make the turn slow yeah but the store closes in five minutes so like go as fast

as you can given the circumstances in much the same way when you're designing training programs if like someone writes you a program that's like everything is 5 R you could at least ask the question of like okay so what concerns do we have that make us um take the very slow road because it could be like look like the freeway's fucked up uh the slow Road has no cops on it or some shit like that or actually this is a shortcut something but it's got to be some kind of reason so when you see someone

at the gym who isn't cranking real close to failure that might not be doing the wrong thing and in context might be doing the right thing but you have to at least ask and on the strength side um if someone's cranking to failure all the time you should at least be asking like what advantage does that have for you because it seems like if you went shy of failure there's not that predictive force of getting you better results for first and foremost I love that analogy the way that I usually kind of think about it

is I as you know the the kind of the fatigue and failure thing sometimes I think is kind of overblown in a way so I I've I've tried to sit on the fence as much as I can with this kind of thing um and acknowledge that yes they may be um some additional stimulus on the table per set but also not try to be like a like a put me in like the failure Advocate bucket and then also acknowledge that yes there might be a little bit you know other factors like fatigue that we'll talk

about um that may come along with that but also that's not always an issue either I I kind of view it like an if then principle so if you have some sort of uh roadblock from training by training everything very close to failure then you may respond in some of these other ways to manipulate your training program pull back on that a little bit but if if you never run into that first roadblock in in in the first place that's where understanding that the stimulus maybe there is something to gain would allow you to optimize

that program a little bit further so for from the hypertrophy perspective that that driving analogy is exactly the same thing um and then from a strength perspective yeah I think the the simplest way to think about it is that in the short to moderate term so the length of the training studies that we have I think that's an important um a few months basic exactly that's an important distinction because ultimately for the very long term for a strength athlete you're basically training for hypertrophy to some degree like you kind of said um so that's where

I would kind of shift the conversation to say like that would make sense if you're training closer to failure you're your some of your reps are grinding things like that but in the short to moderate term the the the length of the studies that we typically have so more like your meat prep or or uh you know your peaking phase things like that in that kind of that length of time I would say basically the easiest way to think about it is your sets should be slowing down because they're heavy and not necessarily because you're

fatiguing within a set it's really the simplest way to think about it um because if I can make the set slow and specific to a strength test because I'm adding load to the bar that's a very different you know neuromuscular kind of condition than hey I just did a 20 rep max and that's like an extreme example but those kind of the principle specificity there and how you're getting to those slow grindy reps one of them very much reflects the sport or the goal that you're trying to train for the other one is just very

very different um so that's basically for strength I think that's the easiest way to think about it is you want to be heavy not necessarily fatiguing yeah yeah yeah um or like if you can add a rep to your strength sets but or you can add some weight adding weight's probably a good idea CU that's what you're being tested on okay super so as far as the most basic take-home stuff from this meta Y what are kind of a few things you think people should absorb from this and carry with them in their little wisdom

pouch sure yeah I think the most basic thing is that if you kind of go back and look at all the the categorical meta analyses that we talked about so the a versus B failure versus non-failure I think those all kind of painted a picture that there's nothing to gain from training closer to failure and if you look at those a little bit closer pretty much every single one of those effect sizes does lean in favor of the failure condition but for statistical reasons we don't have to bog down too much doesn't reach kind of

that threshold of what people would take notice of um when you do kind of these different analyses where we kind of open up the the the the range of the rears that we're looking at and look at things in more kind of a continuous fashion it just makes that a little bit more precise and I think we were able to pick up on the fact of what most lifters kind of do know is that you know Arnold has been saying it for a long time right like you know those last couple reps that's where the

magic happens and I don't think it's a black and white cut off by any means but ultimately the fact that as a set gets closer to failure the stimulus increases I think that's just a generally accepted principle by LIF um everywhere and I think that makes perfect sense and I think this this gives people like a positive affirmation of that intuition so I think that's the that's the first thing um the second thing would be that I do think it's a it's a helpful case to kind of see where strength and hypertophy do diverge a

little bit in terms of the short-term aspect of of uh training studies and just understanding that when you look at other other data and just understand that in the short to moderate term the length ofy studies there results may not be the same so informing one to inform the other probably isn't um always always the best idea I think that's probably something that came out of it as well um and then third is just ultimately small effects like this are hard to detect and it takes a lot of studies to kind of combine to really

be able to detect some of this stuff so I think um that's where you know individual studies are useful but it it's tough to put a ton of stock into painting the entire picture until you get a lot of data to be able to detect these smaller effects that we're looking for yes and on that smaller effects thing it's probably also true to say that like as long as you're not fucking eight reps from failure you're getting some decent stimulus either way yeah for sure H and then the question is like per set the stimulus

is higher for closer to failure then you have to ask how many sets and so on and so forth exactly awesome [Music] [Applause]