Thursday, February 7, 2013

Some Data in Support of Going Longer on Intervals

One difficulty in training is figuring out why--why we train, why we adapt particular methods and workouts.  This is particularly difficult in January and February, for several reasons.  The season is a ways off, and the effort of a January ride often seems disconnected from the reward of June fitness.  The weather can be awful, and staying cozy can be a more compelling why than a getting fit.

I think something more is needed.

For the past two years I've been fairly motivated to train, but not only for reasons of fitness and racing.  I like riding, even in the cold, but I especially like experimenting with different techniques.  In particular, I'm still trying to figure out the best approach to improving threshold.

Here are two prominent extremes:
-Traditional: ride slow, ride long
-Time crunched cyclist:  do intervals year round

These are extremes, and most plans fall somewhere in between, or involve both.

Since 2011, I've gradually moved away from an almost pure focus on 1-5 minute interval training.  Unsurprisingly, my 1-5 minute power improved a lot, but my FTP did not.  I'd read a great deal of research on intervals suggesting, however, that 1-5 minute intervals improve FTP.  Time and again, studies find evidence of this.

I can't say whether I'm unusual or whether there's something about the typical testing protocol for these studies on intervals; for example, they may test untrained cyclists, there may be omitted variables, and there may be seasonal differences and so on.  In any case, I wasn't seeing the improvements in one of the key variables that measures cycling capabilities:  threshold power.

This year I started doing longer intervals between 15-60 minutes in length.  Every week since November I've done mid-week 2x20 power test.  I've done the test not only to have plentiful data to measure my week-to-week physical changes, but also as a new approach to training.

Here's a chart showing the results, with the data scrambled and jumbled so (1) it all fits together on one chart, so the relationships are clear, and (2) so you don't see my pitiful numbers and you don't know precisely how fat I am.

CPI 20 = power output over 20 minutes (log)
TSS = training stress score for week (log)
Weight=body weight (lbs)
W/kg=watts/kilogram
Now here's another chart showing just my power, as measured by these weekly 2x20 tests, above and below my December 2011 threshold. 

In short, I'm discovering the principle of specificity--the fastest way to get better at a particular skill (and over a particular duration) is to do that skill. Do short to prepare for races where 1-5 power is required (say, Highway to Heaven Hill Climb); do longer intervals to prepare for races where 20-60 minute power is required.  Duh.

A note of caution: my old wired Powertap has been acting strange for several months, so my improvements here might just be measurement errors! Point is, even if you're an old dog like me, switching up your winter training is one helpful way to find purpose. You may find yourself in the opposite situation of having done a lot of longer intervals, or of having focused solely on another type of riding. Try something different, add some why to your how, and see what happens when.

6 comments:

Steven said...

Like you, I used to do a lot of 1-5 min intervals given that the area I live (Newark,DE) doesn't really have hills longer than that. But I think when it comes to increasing FTP, most people tend to respond favorably to the longer "pushing up" intervals like you described.

That being said, I have heard of people getting big gains from VO2max intervals, but it just depends from person to person...

Rollin' Polish said...

I think you are confusing specificity with the principals of aerobic fitness and long term gains.

Sure, 1-5min intervals have an aerobic component, but they also have a lot of things that negatively affect long term aerobic adaptation including:

-increases in cellular acidosis.
-decrease mitochondrial and capillary density and thus o2 transport to muscles.
-incredibly stressful on the cardiac system. Hunter Allen once said doing 1 vo2 max session is almost as stressful as a heart attack.
-incredibly depressive on anabolic sex hormones, such as testosterone. The short of this is that constantly driving down test and increasing cortisol from the stress response prevents recovery and adaptation.
-there is not a single study or body of evidence from any aerobic sport that short intervals are good for long term development. Running, XC skiing, and other sports have all been through the long vs. short debate. It has nothing to do with specificity and all about developing aerobic fitness in terms of a percentage of genetic potential. In that regard nothing beats constant training at or slightly above the aerobic threshold. Other stuff works in the short term, but adaptations cease fairly fast when the negative sides pile up.

Specificity has to do with very finite demands of target events if you want to use the block periodization terminology of Issurin, Bompa, et al. Doing a 2x20 is not specific unless you only do 20 min. tts. Its simply an aerobic workout to create a larger aerobic base. Specific would be something like doing a training race, or doing a long interval at say 90% of FTP or whatever race NP is with bursts thrown in to simulate a criterium.

Moreover, specificity like this is saved for the transmutation, competition, and peaking phase, which are only a few weeks long. Being specific outside of this band would either blunt a peak or slightly retard longer term development because, at least for races that go above the aerobic threshold, some aerobic base is lost during competition in order to prevent undue fatigue accumulation.

So what you've seen is years of riding combined with sticking to what works- working at or slightly above your aerobic threshold on a consistent basis. This is kinda why lots of people dial back their top end work after the season, rest, and chill out on rides and BAM, suddenly find a quick burst of fitness (supercompensation plays a role here). For some reason since the popularity of Coggan and Allen's book has increased cyclists have made up reasons to get away from what has always been proven to work long term. For further testament to this there was a great article recently in running magazine that examined trends in aerobic vs. interval training over time. The verdict was that running speeds either decreased or stayed static during years when a high amount of top end work was en vogue. It then increased when they went to mixed methods that were aerobically dominant.

Rollin' Polish said...

Part 2 of 2:

A personal example for me is of a friend that races on an NRC elite team with some pretty good victories to his name. After 2 years on the NRC scene he had stagnated and done everything under the sun- motorpacing at altitude, oxygen tents, went through 3 'coaches' worked his fatigue profile, etc. Last year he dialed it back and returned to doing a lot of general aerobic work and did a single mid-week 3x20 workout at 90% of FTP. Won more races than ever and packed on ~10% to his FTP. I can give you more info in person but don't want to publish his numbers even though Velonews has in the past.

Some quotes to leave you with:
"In an endurance workout lactate content must not rise too high ... if it does then lactate tolerance is trained instead of endurance capacity ... Intensive workouts going together with high lactate values may be damaging to endurance capacity ... Endurance capacity may deteriorate by this kind of training."- Peter Jansenn

"Back in the early and mid 60s the German's training approach ... (placed, ed.) ... a greater emphasis on high intensity intervals. What they found was that, to a great extent they did reach high performance levels with this training program. But, they were not seeing progressive improvement from year to year among their elite athletes. Every year they came up to the same level, fell back down in the off-season, and repeated the process the next season. Then they changed the composition of the training to higher volume, lower intensity (fewer killer intervals at max speed) and the long term progress began to occur."
Stephen Seiler

"Overloading this training intensity ... prevents the body from developing the aerobic base. Rowers can even fall so far behind that they have to start developing the aerobic base from the beginning. It can take weeks or even months to correct such overloading."
Wolfgang Fritsch

"During the 1950s and 1960s, the top runners' training heavily emphasized intervals. But the interval-trained champions were soundly trounced when Arthur Lydiard's runners came on the scene. Peter Snell, Ron Clarke, and Murray Halberg did just 6-8 weeks of speedwork, after laying in a 12-week base of pure aerobic endurance running. Runners who've done tremendous volumes of speed work - like Emil Zatopek and Bill "Mad Dog" Scobey - couldn't match the times of the endurance-trained Lydiard athletes."
George Beinhorn

Kevin Cross said...

Thanks, as always, for your insights, Rollin'. The quotes and arguments are awesome.

I think we're actually in agreement, although I do think there's a massive payoff for doing shorter intervals, but that tends to be limited to short-term improvement, and it may not benefit FTP to the extent that some studies suggest.

You're right that specificity differs for different people. For me, I specifically want to improve my FTP, which I define as anything between 20-60 minutes (I know, it's supposed to be 60-min best, but in practice, then, why do we then perform 20-min tests (*.95) to determine it?)

But say my target is a 40k TT. Suppose I could arrange a 40k TT training race every two days? Would that be the most specific, and hence, best way to prepare?

Well, much research is ambiguous on this.

Here's a study that compares long and short intervals, and finds that they both lead to improvements in 40k TT performance: http://www.lunekildecoaching.com/web16/images/stories/stepto.pdf

Here's evidence from a study on "highly trained cyclists" who did just 4 weeks of 5-min intervals and saw big improvement in 40k TT improvement. http://link.springer.com/article/10.1007%2Fs004210050119?LI=true

Here's another study that found FTP improved whether subjects did short intervals (30") or longer ones (4').

There's even evidence that 1-rep max squat is correlated with FTP!

Now, another issue concerns release of hormones.

HIT increases production of HGH: http://jap.physiology.org/content/87/2/498.short

I'll say more about this in my next post, but I'm glad you brought this up! Interesting stuff to ponder.

Rollin' Polish said...

Re your 40km question- it would be a bit overall, but maybe ideal during your specific prep block, which would be about 5-6w out from competition (3 weeks specific, 2 week taper/peak). The effects of extreme specificity do not last long. Weightlifting is an easy example- rarely do lifters do workouts arranged in the competition lift order with full lifts at high intensities despite the fact that this would be ultra specific. They only do this during the specific prep period and work the components of the lift during general prep. Same goes with swimmers and seems to be the case with Wiggins last year. For a road racer this might mean working FTP through many different methods slowly moving towards more time on the TT bike and at the target event power/duration riding the target course if possible.

For testing I prefer the Monod critical power spreadsheet (google it its like the first result). It uses anaerobic and aerobic tests to create a power curve that takes the 3 different energy systems into account. I think the longer the test the better, but you gotta be practical. Testing a 20 at Hains is a lot easier than testing a 60 logistics wise and Monod can give you a more accurate multiplier than .95. Most people vary from .88 to .98 of their 20, so Monod helps hone that a bit.

You don't need to test weekly either. I've never understood this. These gains take months to accumulate and PM accuracy is only +/- 2.5%. So say you test one week and you're at 350 and next week you're at 355. You realistically are just inside the margin of error so nothing can be derived from this. Moreover, you can train FTP as low as 85% of threshold up to 108% of threshold so the zone is HUGE. Personally I feel that if I go out and bust out 95% for a 2x20 one day, but then next week I hit 98% its really not necessarily a gain since there are so many long and short term factors that go into stuff you know? In my mind a rider can adjust targets by PE as they go and test as infrequently as at the start and end of a block, which could be 6-8weeks.

Rollin' Polish said...


I've seen those studies and they have flawed designs and the time frame is too short. Of course people will gain threshold power by doing 3-8weeks of vo2 max work. In the first there is no control for lower intensity and they didn't take valuable blood bio markers to establish CK, hemoglobin, hematocrit, etc. values. This is really important as vo2 work can actually LOWER hematocrit and hemoglobin when used too long and decrease aerobic efficiency even though watts go up. This is one of the main reasons why EPO, artificial hemoglobin, and saline infusions have such huge benefits- imagine doing vo2 max workouts and being able to erase all of the negative effects thereof. I digress, but if you think about the average training week a rider can easily blunt their aerobic workouts by driving down these values by doing too much top end. Judging by blood work from last year to this year, my hematocrit was 39 in January 2012 and was 47 this week. I'm also faster this year and the key difference is that I've done 0 L5/top end work so far. I think that training in the mid 40's is probably better overall than burning out and training in the high 30s. My hemaglobin is also dramatically up, so overall my body can carry more oxygen at the same threshold so I've accomplished a higher stroke volume (prime benefit of vo2 intervals) without the undue sides. Pretty interesting, but N=1

I agree that in the short term these types of intervals make sense mainly to quickly tune up for a peak race. 3 weeks isn't that much, but some coaches have 3 of these workouts a week for months on end and I just don't get that. Numbers might go up, but aerobic efficiency might not its a huge trade off.

4 minutes is simply not long. Blood lactate might surpass 4mmol, but in lots of cases it might not either. Also, time at 4mmol is extremely low- you might get 1 minute. This is why the longer intervals make sense when you're essentially trying to increase your wattage and duration spent at 4mmol of blood lactate. A 20min interval will be at the very least 75% of time at this level so a 2x20 gives you 30min of time at threshold vs. a 6x4 giving you 6min. So yah, they have an effect, but its pretty miniscule compared to longer efforts. Combined with the high risk factor and the fact that they have a limited usage window, its just not as sustainable in the long term for overall development. I think at some point though a rider does need this work and they can be used for FTP development i.e. 8-10x5min @100%-105% of FTP with 1 min rest or less being one protocol that comes to mind especially for someone that might race in MABRA.

Lastly on this subject, most studies that actually compare shorter to longer intervals rely on a very flawed study by Baumgarter et. al, which compared these intervals to people riding 4hrs of L1 a week and then determined that the aerobic utility of the workouts were high. The study also only lasted 8 weeks.

The HIT study re HGH really isn't HIT since only 25% on the bike was done at lactate threshold. HIT would be around 115%+ of Lactate threshold or more, over 200% for Tabata style workouts. Also, the problem with that study is that HGH does spike with exercise in the short term, but over the long term the spikes in cortisol and other stress hormones can impact its release. Diet also plays into this as well.

Hope to roll with you soon