What the Science Actually Says About Exercise
In this post, I disputed the so-called “scientific” exercise program of Little and McGuff. So I figured it was probably worth describing the exercise program that I do believe the science supports. The following recommendations hold for the “average” person who simply wants to be in decent shape for every day life.
The key scientific findings are:
– For strength training, one set is enough. The standard regimen at most gyms is to instruct people to do three sets of each exercise. Sets two and three are a waste of time for the average person. Much better to either save your time or do a greater variety of exercises. See here. (For highly trained individuals, you do need to use a multi-set periodized regimen to achieve additional gains.)
– For strength training, once per week is enough. The standard regimen at most gyms is to instruct people to do a complete weightlifting routine three times a week. However, training once per week gets you about 75% of the benefit of training three times per week. 3x the effort for 33% more gain seems like a poor payoff. See here. (For highly trained individuals, your workouts may be so intense that once per week is actually optimal.)
– For cardiovascular training, high intensity intervals produce the best results in the least time. Long distance cardio is beneficial, but you’ll actually improve your performance about as well with intervals in much less time. See here. (For highly trained individuals in long distance disciplines, you must also do distance of course.)
– For cardiovascular training, running has the most cross training benefit. If you want general fitness, running is therefore the best exercise. See here. (For sports participants, sports specific drills are of course the best conditioning.)
– Stretching is not a very beneficial activity. Contrary to popular belief, it does not reduce the incidence of injuries. The evidence here is really strong. Moreover, stretching before strength training or sports actually dramatically decrease your maximum power, potentially increasing the potential for an injury. The only thing it does is increase your range of motion, which does not even indirectly prevent injury though it may improve the performance of activities that require more flexibility. See here.
– To maintain a healthy weight, you need to burn calories.
Thus, here is my recommended program:
– Do 2, 1/2 hour strength training sessions per week, one for upper body and one for lower body. The issue is that most people can’t maintain intensity for more than 1/2 hour. So splitting strength training into upper and lower body enables you to intensely train the entire body.
– Do 2, 1/2 hour interval training sessions per week, one of which is running on a track or treadmill. The pattern here should be 30-60 seconds of maximum output followed by 2-4 minutes of rest. Repeat 4-7 times. The key here is that you should be completely wiped at the end. Increase the interval intensity and time as well as the number of intervals to achieve exhaustion.
– Walk 2 miles every day. If you weigh 150lbs, this will burn the same amount of energy in a year as is contained in 20lbs of fat. At 225lbs, it’s 30lbs fat equivalent. I also find this is a great stress reliever when done at the end of the day. It’s a great time to make calls to friends and family to catch up too.
That’s it. Not a huge commitment. But the gains for the average person would be substantial.
Possible Insight 
“Much better to either save your time or do a greater variety of exercises.”
Probably better to save time (i.e. Rest) Not sure there is any significant benefit to doing more/diverse exercises. Better to Keep it simple, maximize intensity for 30-40 minutes and call it a day.
“For strength training, once per week is enough.”
Is there not some correlation between strength and recovery time? (i.e. The stronger you are, the more time you need to recover?) I think beginners can workout the same muscles twice per week or twice per 10 days. (Everyone is different)
“For strength training, one set is enough.”
One set to failure, no? A light warmup set can be useful. You aren’t going to walk into the gym and immediately try and bench 350 pounds are you?
Daniel Horowitz
March 25, 2010 at 6:22 am
Point 1: Perhaps you didn’t think through the math. Your trainer gave you a 3-set program that took you 60 minutes. A 1 set version will only take you 20 minutes. So to achieve 30-40 minutes, you need to add exercises.
Point 2: There is some debate on this. I tend to agree with you, which is why I made the parenthetical comment. However, the statistics on effectiveness were for beginners.
Point 3: There is some debate on whether going to failure is necessary. I tend to agree with you but the science isn’t actually clear. As for warm up, the average person just needs to warm up the upper or lower body generally before the workout. An exercise specific warm up isn’t necessary. It’s different for me, which is why I made the parenthetical comment.
kevindick
March 25, 2010 at 10:08 am
Point 1: No, I did not even consider the math. But, since you are going there, I will give you a chance to reconsider. You are proposing that workout time is linear with respect to a single variable, # of sets.
I believe that there is additional time involved in switching exercises, thus set 1 (i.e. a new exercise) actually takes more time than sets 2+, so it is possible that a one set program takes 30 minutes and 3 sets on the same program takes 60 minutes.
(e.g. 10 exercises one set each will take you longer than 10 sets of one exercise.)
Point 2: I think we agree
Point 3: I think we agree. I don’t believe going to failure is necessary, though as sub-maximal efforts are less stressful, part of the advantage is the ability to workout more often and increase muscle efficiency with respect to specific exercises.
danielhorowitz
March 26, 2010 at 9:05 am
Point 1: Sorry, but no. If you do three sets of each exercise, you must either (a) do three circuits or (b) rest 60-90 seconds x 2 for every exercise (between sets 1 & 2 then again between 2 & 3).
In case (a), the time for 1 versus 3 sets is _exactly_ 1/3.
In case (b), it’s typically _less_ than 1/3. A properly constructed single set workout requires rest only after every 4th exercise.
I assure you that I have spent hours figuring this stuff out analytically and years verifying the results empirically. 1/3 is the _upper_ bound on the relative time factor.
kevindick
March 26, 2010 at 10:25 am
So, If you were doing 1 set Bench, 1 set Rows, 1 set Deadlift, it would take you the same time as 3 sets of Bench?
Daniel Horowitz
March 26, 2010 at 2:45 pm
Yep. If you do the algebra comparing 1 and 3 set workouts, you find everything cancels out and the 1/3 equivalence point is when the transition time between exercises is the same as the rest between sets. If it’s less, the 1 set workout is more than 3 times as fast.
I typically see people resting 60-120 sec. The only transitions that take anywhere near this long are for olympic bar exercises which I would never make part of a 1-set per exercise regimen.
kevindick
March 26, 2010 at 2:58 pm
I believe the conclusion is better summarized as: “Warm-up is important and don’t stretch too hard”:
“Stronger evidence demonstrates that various approaches to conditioning that include warm-up and stretching along with other techniques such as strength training, plyometrics, and proprioception training both enhance performance and prevent certain kinds of injury (52,64,83,94). This suggests that strength training, conditioning, and warm-up play an
important role in injury prevention. In addition, stretching of specific muscles and joints for specific activities might enhance the effectiveness of these other preexercise activities, an approach consistent with a multifactorial model for prevention (97). At the same time, there might also be a risk of injury and impaired performance associated with stretching without adequate conditioning and/or warm-up.”
Alex Golubev
March 25, 2010 at 7:09 am
Yes and no. Warm up is important. Stretching, whether hard or not, is not important. This is quite clear if you read the CDC meta study and its component studies.
kevindick
March 25, 2010 at 10:12 am
What’s your opinion on the kinds of exercise that combine strength and stretching dynamically (e.g. many forms of yoga)?
BTW, this is my favorite post of yours so far 🙂
Rafe Furst
March 25, 2010 at 11:52 pm
I would not have guessed this would be your favorite. But thanks.
My opinion of yoga is that the science definitely shows it will increase your flexibility. It also appears to provide some improvements to cardiovascular fitness and muscle strength. However, these gains are relatively modest and occur in programs that require a lot of time (typically 3-4 hours per week). So yoga is not a very efficient form of cardio and strength training.
Actually, the biggest benefits seem to be in reducing stress. However, my guess is that you could achieve the same stress reduction more efficiently through a pure meditation program.
Now, there seems to be a large social component to yoga. This is of clear value to many people. And if that gets people to exercise, I’m all for it.
kevindick
March 26, 2010 at 1:19 am
I was more referring to the idea that dynamic and weight-bearing stretching that happens in yoga feels different to me. I certainly have bigger flexibility gains than I’ve been able to achieve through static stretching, and it also feels like I’m less at risk for injury while stretching (contrary to what I’d always heard through American fitness folklore). Is there data to support this?
For someone who is naturally very tight (but not at all prone to bulking up) the athletic performance gains I can achieve with marginal flexibility improvements are far greater (empirically) than I can with marginal strength improvements.
And yes, as an aside, I get more out of yoga than just the strength, cardio and flex. All in all it’s super-efficient for me.
Rafe Furst
March 26, 2010 at 7:17 am
Swimming probably offers the best exercise value. Easy on you joints + cardio and strength. You MAY pull a calf finding a local Olympic size pool however. 🙂
Alex Golubev
March 26, 2010 at 8:58 am
Actually, that’s probably wrong. Follow the link on cross training. Swimming provides the least cross over cardio benefit. It appears to be a very specific activity.
kevindick
March 26, 2010 at 9:59 am
My guess is Rafe likes the uncertainty, the variability, the complexity.
“There is some debate”
“The science is not actually clear”
“It’s different for me”
danielhorowitz
March 26, 2010 at 8:40 am
I have a few comments in no particular order:
1. I like the general gist. Training follows the power law concept, which I like. Very time efficient way to go.
2. Max effort intervals are probably not appropriate out of the gate for the unconditioned. I’d work up to them. (I’m sure you’d agree). Aerobic intervals (easy/harder) are probably a better place to start working up to max effort/rest type intervals.
3. I’d probably make the cardio prescription 3x/wk with at least 1 interval session and at least 1 steady state session, but that’s just me. The third session would be a matter of preference. I think there’s a place for a bit of moderate intensity running/jogging in any program. If time is an issue, I’d sacrifice one of the walks.
4. The resistance exercise Rx is probably best for the 8-12 rep range. Lower than that and you’re really training for strength and should start periodizing (as you mention).
5. I’d love to see a 1 set per exercise lower body workout that takes a whole half hour. Seriously though, assuming one can get some form coaching, squats and deads are easily the biggest bang for the buck and probably require a bit more warmup (although maybe not so much in the 8-12 rep range).
6. While I agree that stretching as means of acute injury prevention before/after workout is unproven (and is likely detrimental to performance before a workout), I’d have to argue that most people could benefit from stretching. This would include pre-workout if one does not have an adequate resting range of motion to perform the exercises. This could be accomplished as part of the warmup set(s) potentially (with very light weights). For a long time this was critical for me before squatting deep. This also applies to weekend warriors as they age. Playing basketball in one’s 40s without a dynamic range-of-motion warmup (not static stretching) is probably a bad idea.
I would also argue that most people (who don’t actively try to address the problem) have muscle flexibility/strength imbalances which can lead to postural distortions and chronic problems. Typical are tight hip flexors, hamstrings, calves, and upper traps. These muscles should probably be lengthened and their antagonists strengthened in most people (especially desk jockeys). Runners are also prone to tight IT bands and glutes. With the stuff that I do, if I don’t keep my glute medius stretched regularly, I’m prone to SI joint problems.
Just my $0.02.
Todd White
March 26, 2010 at 9:23 pm
5. Right. I would actually Squat one week and Deadlift the next. And my workout would go:
(Deadlift)
Warmup 10 reps one plate on each side
Warmup 3-5 reps 2 plates on each side
Working set of whatever I’m doing.
/End workout
And, if I’m squatting 20 reps deep, there is no way I’ll be able to do anything else afterwards, but that’s just me.
Daniel Horowitz
March 27, 2010 at 8:07 am
“And, if I’m squatting 20 reps deep, there is no way I’ll be able to do anything else afterwards, but that’s just me.”
Wuss 🙂
kevindick
March 27, 2010 at 12:32 pm
And in the everybody’s different category, I found out that if I’m going to squat heavy, I need to do it at least twice per week. Otherwise I get the worst DOMS in my adductors and can barely walk the next two days. Once per week is probably better for strength gains, but I learned to take slower gains in exchange for comfort.
Todd White
March 27, 2010 at 2:43 pm
(2) Actually, I do disagree. The evidence for HIIT with untrained subjects is actually very good. 4 x 30 sec intervals twice a week really isn’t that much. Of course, what an untrained subject thinks is maximum effort is actually not. So they are working up to it. I certainly don’t advocate driving an untrained person to exhaustion with drill sergeant tactics. So it’s self-limiting.
(3) That’s the conventional wisdom. But the evidence seems to say otherwise. That’s why I have the 2mi/day walking. To make sure caloric output goes up in the absence of classic cardio.
(4) Agreed.
(5) Oh, that 30min includes 5 minutes of warmup. So that’s only ~12 exercises. In my recommended protocol, I use 3 groups of 4 exercises with a 2 min rest in between each group. Easy to do that.
(6) The evidence here is strongly against a pre workout stretch. Maybe if your activity is something like gymnastics or wrestling that requires an unusual ROM. Or if you have a particular injury that restricts your ROM. But as a general recommendation, there’s no evidence for it. Even for 40 year old basketball players.
I’d like to see some evidence for generic stretching preventing exercise-created muscle imbalances in the general population. It seems to make sense and I do recommend 5 min of post workout stretching as a hedge. But I haven’t seen any convincing studies.
kevindick
March 27, 2010 at 12:30 pm
2. If I tell an untrained 50 y.o. woman that she needs to do 4x100m as fast as she can, I risk either (in order) scaring her away from exercise, pulling a hammy, having an MI. I’d rather have her work up to it over a month or so. That’s just me.
3. I have a hard time breaking completely away from the conventional wisdom, because it’s so … conventional.
5. You have a routine of 12 lower body exercises? Do tell.
6. I didn’t say pre-workout stretch. Well, actually I kinda did, but listen to what I mean, not what I say. I totally agree, no pre-workout static stretching. I do really believe in a pre-workout dynamic set of movements to take the muscles through their range of motion for when the workout is likely to challenge their limits, especially in a fast or unpredictable way. Thus the squatting and basketball examples. I know for one that if I play basketball without doing this, I’m more likely to get a minor muscle tweak. This is probably unnecessary if the workout will not challenge the range of motion (or will do so in a very controlled fashion).
I’m not really talking about exercise created muscle imbalances (unless you’re a bench and biceps only doof). I’m talking about sitting-at-a-desk-all-day created muscle imbalances. I’ll see if I can find some references.
Todd White
March 27, 2010 at 2:40 pm
(2) I’d probably start your hypothetical woman with HIIT on a stationary bike. Obviously, the form of exercise must be appropriate to ability. But the HIIT pattern is what’s optimal. In fact, it’s exactly what I started my 67 year old father with a replaced mitral valve on (after getting it cleared by his cardiologist).
(5) Sure. Group 1: dumbell deadlifts or squats, dumbell Romanians, calf raise, toe lifts. Group 2: reverse lunges, reverse crunches, forward lunges, cable torso twist. Group 3: leg press, leg curls, abductor machine, adductor machine.
(6) Oh, I agree that a set of warmup exercises that take the muscle through the expected ROM of the activity is a very good idea. It’s what I do for wrestling and boxing. But I don’t stretch.
I was confused by your runner’s imbalance statement. For sitting-induced imbalances, I’m not sure. I’d want to see a copmparison of just stretching, strength plus cardio, and strength plus cardio and stretching. My guess is that getting the proper amount of activity is what’s important, not the stretching. For prevention. Once you’re messed up, you may need stretching.
kevindick
March 27, 2010 at 4:31 pm
I’ll bet you can get 75% of the results from 1/3 of your 12 lower body exercises.
Daniel Horowitz
April 14, 2010 at 11:15 pm
wait what study showed:
– To maintain a healthy weight, you need to burn calories.
We hear this a lot and it makes intuitive sense but how do we know this is true?
Hendrick
March 28, 2010 at 11:36 pm
Here are the most highly cited studies on this topic:
http://content.nejm.org/cgi/content/abstract/319/18/1173
http://grande.nal.usda.gov/ibids/index.php?mode2=detail&origin=ibids_references&therow=83440
Click to access 92.1.full.pdf
kevindick
March 31, 2010 at 10:32 am
That would be one factor but it is also advisable that you should also monitor your diet if you want to stay healthy and fit.
Recumbent Exercise Bikes
June 2, 2010 at 7:25 am
more specifically, I have read that exercises such as interval training work because of the effect they have on your metabolism and insolen reaction not because they burn more calories. Hence interval training > long distance running for conditioning and weight loss even thought long distance running burns more calories. So how and why do people think burning calories is a necessarily good thing for your exercise routine?
Hendrick
March 28, 2010 at 11:42 pm
Thanks. Just read “Body by Science” and a number of on-line papers on infrequent high-intensity slow interval training. I’ve started a once-a-week routine at my gym and will try to remember to report back here after 12 weeks. Keeping track of weekly strength improvements, etc., pushing to the limit on each. It kills, and I’m sore for days (53), but the theory is fascinating. I augment during the week with fast walking, about 1mi/day avg.
My old gym routine was 2-3 times/wk of low to medium circuit. Didn’t do much for me.
John L
April 17, 2010 at 4:28 pm
The program works, at least in the early power law curve (first 12 weeks). I've done McGuff's “big five” routine once a week for roughly 3 months, plus fly and leg curl, and have seen an average strength increase of roughly 25% (starting/ending), with some muscles showing over 30%, and some under 20% improvement. This is from 15 minutes, once a week, of a high intensity circuit – 5 to 10 ultra-slow reps to failure. I never understood why people yell at the gym. Now I know. This kind of workout is incredibly intense. Also starting an interesting high-intensity, short duration cardio program that Mercola recommended. http://bit.ly/ccuq7R
John L
July 8, 2010 at 3:56 pm
“Stretching is not a very beneficial activity.”I feel you missed the boat on the importance of flexibility, and my need to revisit the issue.Strength training does not increase ROM, activities like yoga, Tai Chi, and even resistance training will increase ones range of motion. It has also been shown that Tai Chi can have a positive impact on cardiovascular performance. Why is flexibility so important, because in an aging population, lack of flexibility is what leads to inactivity and injury, not lack of strength.
Jasonmckinley
September 20, 2010 at 8:33 pm
In fact, I did acknowledge stretching increases ROM. However, I cited a very large body of evidence showing it doesn't decrease injury. If you have evidence to the contrary, please reference it and we can compare.I'd love to see your evidence on Tai Chi increasing cardiovascular performance. My guess is that you have to do many hours of Tai Chai and then get a marginal benefit. See my comments on Yoga in response to Rafe's comment.
kevinsdick
October 13, 2010 at 4:31 am