Scientific Basis for Different FTP Tests
Since FTP is at the heart of our training, I'm on a mission to gain a deep understanding of the physiology of FTP, and why it is that we can have the gold standard 1 hr test, the 20'/20' test, the 5'/20' test, and others that may be out there. I'm sure there's been a ton of discussion about this already, so please point me in the direction of the most fruitful threads. The questions I'm most curious about are:
1. What is the physiological theory behind FTP, why can it be at the center of our training, and what are references where I can read more about it?
2. What kind of sub-hour tests are there (20'/20', 5'/20', etc.), how are they performed, and how do they fit into the theory? Or posed a different way, why can we do sub-hour tests in the first place?
I'm already going through Allen/Coggan - which is excellent as an overview - but I'm looking for more of the hard-core science.
Comments
1. You could also read Philip Skiba's The Triathlete's Guide to Training with Power. It's a bit more sciencey than the Allen/Coggan book. It's still written for mass consumption though so it's not on the level of a pubmed article. There are probably numerous articles that are super nerdy on pubmed. You can try searching there.
2. The reason most do sub-hour tests is because the hour test is just brutal. I did it once several years ago and will never again do it. The shorter tests can get you in the ballpark but I also think... in general... the shorter tests over inflate FTP numbers for most. That's even considering taking 95% or whatever. I bet 90%, and not 95%, of the 20' portion of the test would be a far more accurate number for most. Anyone can ride like a hero for 20'. Few can do it for one hour.
3. The FTP number, however you slice it, is just a basic guide. Just because you tested at XYZ watts doesn't mean you can ride 56 or 112 miles at 85% or 70% of that number without a lot of training and practice doing so. Too many people just take that 20' x 95% x 70% (or 85%) as some sort of gospel right.
Here is a short but very good thread by Joe Hallatschek going over the gold standard 40kTT and how it relates to his 5/10/20 EN testing.
http://members.endurancenation.us/Forums/tabid/57/aft/15320/Default.aspx
Here is a really good blog to read. Highly recommend the posts titled . The Seven Deadly Sins , Turbocharged Training , The Sin of Sins FTP testing #2 , Left Right Out of Balance....
http://alex-cycle.blogspot.com/2009/07/sins-of-sins-testing-ftp-2.html
"FTP is the highest power that a rider can maintain in a quasi-steady state without fatiguing for approximately 1 hour.
When power exceeds FTP, fatigue will occur much sooner, whereas power just below FTP can be maintained considerably longer".
My thoughts based on my experiences and what I have read --- We all know the "Gold Standard is the 60min or 40KTT" This would be your 1 hr power. This is NOT debatable, but very few people do this. Instead we choose to do Various tests such as 5/10/20 or the 2 x 20(2) which may or may not correlate well to an individual's real 1 hr FTP , and they end up being nothing more than an estimate of your true 1hr power. There is nothing wrong with any of these estimates and are perfectly fine for determining your zones for training. I believe the 20 minutes test is your 20 minute power , I believe the 2 x 20 (2) is your 42 minute power , each individual will then have a different percentage of these numbers that would be your "estimated" FTP or 1 hour power. For me I believe to be around 93% of my 20 minute power and around 99% of my 42 minute power... For the 20 minute test my main problem is the 5 min "all out blow out V02" if this is truly done all out then there is nothing left for the 20 min "all out test" I can absolutely kill myself in 5 minutes if I really try... So what we end up with is some that will kill themselves and some that will sandbag the 5 min... The 20 minute portion ends up being completely in your V02 Zone above your FTP ??? Your operating in Z5 to determine Z4 and for me once I go above Z4 there is not much time left.... Again its fine and I do them as written because I suck at them and need to get better but in the end I view it as my 20 minute power and approximately around 93% of that is my FTP... For the 42 minute test I view that as more accurate and I test better, after all its closer to 1hr and operating within Z4 not above , but if truly done "all out" then it is nothing more than your 42min power NOT your 60min power , but its also much closer and therefore I believe my correlation to be around 99% of this number.
Bottomline...Its not really that complicated.... Data is your friend... Test .... Set Zones... Train... ReTest ... Collect Data .... Correlate collected data over various time-frames and Races.... For me I like to track various time frames of my power , such as , 5min , 20 min , 40 min , 60 min , 2.5 hrs , and 5hrs... I also consider the w/kg ratio relative to my races (this number is not debatable being how much you weighed that day and what your power was for that race or ride) more relative to fitness/performance rather than the w/kg ratio of "your weight on that day to your estimated FTP ".... Your weight and FTP are dynamic changing numbers day to day but when you know your weight and power of a specific ride you then you have an exact w/kg ratio of data that is true!
If your interested later I can post up my test data and race data from last year that is important to me and how I interpreted it and use it going forward? Hope some of this helps!
If you haven't looked through these two Wiki posts, check them out for a comprehensive look at how we use FTP and power in general at Endurance Nation:
http://members.endurancenation.us/Resources/Wiki/tabid/91/Default.aspx?topic=Alternate+Means+of+Determining+FTP
http://members.endurancenation.us/Resources/Wiki/tabid/91/Default.aspx?topic=EN+Power+Training+Resources
Building on what Bob and Tim said, here is the way I look at using FTP testing in the context of training for and racing long course triathlon. First, a twist on the original question: Is there a scientific basis for using FTP as the basis for training and racing for LT triathlon? I doubt there is a formal study of the relationship of the physiology of one hour power to the physiology of 3 hour - 6 hour races with a run to follow. There is a LOT of knowledge about the various energy systems used in muscular effort in general, and in cycling in particular, and which of those systems play the biggest roles in races of various types and distances. From that, we have learned that we need to train ALL of those systems to be the most effective on race day. The general pattern used is to train the shorter systems first, the progress to the longer ones closer to race day.
In that context, FTP is simply a tool to allow systematic training off of a known baseline. We use the term "FTP", but of course, we are actually using 20 power as the basis. In addition, we have the 5 minute power as another standard. Since most of our intervals are in the range of 2-30 minutes, this makes a lot of sense for training purposes. Periodic testing allows us to both monitor progress, and up the effort level when indicated by new test results.
Racing is a different story. For a sprint or Olympic tri, one hour power makes a lot of sense. But as the distances get longer, the FTP becomes less useful. Back in 2007/8, the EN coaches and team pooled their racing and testing data, and developed a rough and ready chart to correlate FTP, race time, and IF. It was purely empirical, and works fairly well within a range of, say, 5-6.5 hours for the bike leg, for someone who has both trained well, and is willing to ride at a steady power. But there may be better ways to guide race day execution, or at least additional metrics to follow. The one which seems to help the most people is (for an IM) "five hour power", meaning the normalized power one is able to sustain for a five hour stretch during training, as calculated on the Performance Management Chart in Training Peaks. There may be others, waiting to be found.
For now, in the OutSeason, I think of FTP as simply a number to guide me in how hard I need to work in various intervals I am doing on my bike. Come race season, I don;t really start to focus on race execution until after my first race rehearsal coming six weeks before race day. At that point, I start gathering metrics such as HR, RPE, power during 5+ hour rides, and RR rides as well as FTP to help me decide what effort level to target during the bike leg.
Final note on the value of testing... the way to improve cycling performance is by working HARD in intervals which stress the systems we will be using on race day. I learn how hard I have to work during those intervals by working to exhaustion in a test, then work at the same effort level for shorter times with some rest in between.
....give me the science behind the best (and, thus, shortest time) way to get stronger and faster....
Remember when we were 100% self coached?!?! How many hours did we spend sorting thru studies/articles and patching plans together?!?! In my case, it was referred to as ''Paralysis by Analysis''.
And, yes, it is important. Clearly, you've done plenty of homework on this. BUT, now it's time to start your own 'study', your own 'experiment' on you. RnP have done the homework for you.....your job is to do the work, just sweat! That's the reminder to what you already know. That should free up some of your time and energy right there
If you want try a different test, make sure you do the same test throughout the entire season. Check the results.
Hell, try a different test next season and see if the results vary.
But, then there's all these other variables....another year of fitness under your belt, maybe you ate better/worse this year, came in lighter/heavier, different race schedule, didn't crack a collarbone in the group ride, didn't have to move for a job, etc. So, it's really hard to see which is best. Basically....
Pick one, stick with it.
The science is done. Smarter folks than us have already put in the time and are allowing us to apply it to our own circumstances.
I've noticed that I can hold a strong power level for 5' or 20'. It's killer, but I can do it. But when I do the 3x 15' @ Z4 (like my today's workout), I can barely hang on. I strongly doubt that I can hold that for an hour. This may have to do with me being more of a short course specialist than an IM or even an HIM athlete. Very interesting to hear that even some of you IM folks are around .93 instead of .95.
I've been looking into Critical Power (check out this calculator: http://www.cyclingpowerlab.com/MonodCriticalPower.aspx) which uses various data points such as the 5' value as well as the 20' as factors. What's interesting here is that if there is a substantial difference between, say, 5' avg power and 20' avg. power it will REDUCE your FTP estimate. To illustrate, if 5' = 310w and 20' = 276w, your estimated FTP will be LESS than if you did 5'=286 and 20' = 276. This seems counter-intuitive at first blush, but when you think about it, it makes sense: If you're more of a sprinter like me you're going to have a greater fall off the longer the duration. Your hour should fall off substantially. And this may be the power of critical power (pardon the pun), it seems to take in account the kind of cyclist you are.
The problem with this kind of assessment, tho, is that if you tank the 5', you get a better FTP than if you pushed it to the limit. Bottom line: Your test MUST be max achievable. Anything else and you're cheating yourself.
Since I am doing sprints and olys, the 1hr FTP is extremely useful: I absolutely do not deny the effectiveness of the metric for training purposes. There's enough silverbacks like you all (a title of honor, not of age) that have been kicking butt for years who swear by it, that it's withstood the test of time. I absolutely am sucking it up and hitting the training. I just want to understand the science of it for my own curiosity.
On that note, one of the most important things I've gained out of joining EN, and my first foray into NOS is that I can now hold heart rates on the bike north of 170. This used to be only the domain of my running game. My bike at max push used to hover in the 160s and I knew I had untapped potential. Now I'm tapping into that potential! This is because I'm pushing into those Z4s!
As for consistent testing, I know that's the crux of the biscuit, but I think I've botched both my NOS FTP tests: The first one I don't think I fully achieved my 5' power (my max HR was not in the upper 170s then), and this last test I launched the 20' only after resting 4 minutes (confused the FTP ride RI with the test RI), so I was horribly depleted when I turned my 20' in. I'd love to be consistent on the next test with respect to my previous, but I'd really just like to do a correct test for once!
I'm probably going to do a range of testing and test types as the year rolls on: I habitually do a "burn the rubber for an hour" session around the Rose Bowl when things warm up. That should give me good real FTP data, plus more 5/10/20s, and I have a mountain behind my house that I can hammer up for a good 40' straight with almost 6% grade the entire way - zero stop signs. Knowing the physiology will let me see into my relative strengths and weaknesses.
@TC: I'd love it if you could post your nums!
Anymore info that comes to mind, please post! Thanks all!
2014 Last years tests listed below with my thoughts on my FTP as I went along .... My peak FTP tests in 2013 were in the 230-235 range via the 2 x 20 (2) tests , assumption was minimal loss prior to OS .... I adjusted my FTP above some of my tests and below others...
1st test inside 12/13 5'@248 20'@232 232x.95=220.4 CP-226.67 I adjusted my FTP from 230 down to 225
2nd test inside 1/11 5'@249 20'@233 233x.95=221.4 CP-226.33 I kept my FTP @ 225
I did not believe my FTP had dropped all the way to 220 , setting the FTP to 225 and hitting all intervals based on this number proved my thoughts.
3rd test outside 2/8 5'@258 20'@246 246x.95=233.7 CP-242 Not believing either of these numbers I adjusted my FTP back up to 230... training on this number proved very difficult but accurate
Next was something totally new... I participated in 2 TT's , 1st was 4 miles and the second was 12 miles with 1hr rest in between them... Note: these were 100% aero
4th test outside 3/7 10'@252 35'@226 --- great training but no usable data to change FTP.... FWIW CP for those 2 efforts was 215.... So I have CP #'s between 215-242??? If I had to guess at this point I would say FTP pretty close to 100% of the 35' effort.
Race season was upon us .... NO more testing until after a Sprint , IM , OLY races...
5th test outside 6/3-- 2 x 20'(2') outside , 100% aero-bars , absolutely picture perfect test , first 20-235w , 2nd 20 230w , 42min FTP = NP230 --- good confirmation of my FTP setting
6th test outside 6/14--- 20 min all out hill climb effort after 1hr of riding with NO 5' blow off prior 20'@245 245x.95=232.75 I kept FTP @ 230
7th test outside 7/19--- 20 min all out hill climb effort after 1hr of riding with NO 5' blow off prior 20'@252 252x.95=239.4 I kept FTP @ 230
Tests- 6 and 7 were exactly the same ride....This was an all time high 20min power on #7 but with no 5' prior I just kept 230...
So as you can see I had results from 220-242 depending on the tests and how you interpret it... I never had any results indicating to me I would be able to hold greater than 230 for one hour so never adjusted above that... To be honest , if you apply my theorized .93 x 20' efforts or .99 x 42' efforts I would have an FTP in the high 220's and that would be a number I would go for on an hour effort! So even though I had tests indicating above 230 my thinking of keeping it at 230 was appropriate maybe even high , but while training I was able to confirm this 230 number by hitting 40 minutes of FTP work @ 100% most weeks. Sure I would have liked to "SAY" my FTP is 240 and my w/kg is 4.5 but reality comes into play! Its as simple as, If you cant hit your numbers in training its too high ,If you exceed them every week too low, if you can barely hit them its probably just about right.
My Races and Quick Correlations... These numbers are more important than tests to me.... Assumptions FTP = 230 and Weight 120lbs... With the exception of my KONA race you can clearly see a pattern of watts/IF/w/kg at each distance... These also correlate very well to past years performances...
1 SPRINT 37min @ 218w IF.95 w/kg 4.0
2 OLY 1:16/1:20 @207/210w IF.90/.91 w/kg 3.8/3.86
3HIM's 2:31/43/26 @194/192/194 IF.84/.83/.84 w/kg 3.56/3.53/3.56
2 IM's 5:06/10 @176/174 IF.76/.75 w/kg 3.23/3.22
outlier KONA 5:43 @ 155w IF .67 w/kg 2.85
I think the key is to not overthink FTP... Pick a TEST , TEST , set zones, train , retest , race , collect data ,look for patterns, and be realistic... they are just numbers and they change everyday just like our weight!
1) Bike power training is a mathematical model being superimposed on a complicated physiological system. It's an approximation. It's worth getting reasonably close to right, and it's worth having the specialists really pound on it to figure out if they've got it right, but for most of us in the real training world, the approximated view is fairly reasonable...but you have to live with the fact that it's still currently an approximation. That's why you hear things like "Not all TSS points are created equal." Well, the whole point is that they are SUPPOSED to be all equal, so when you have to say they aren't, that is pointing out a flaw in the model.
2) Understanding this, the important thing to understand for our baseline single-benchmark method is that "FTP" (or whatever you want to call it "CP60", "CP" etc) is a good single benchmark and that all three are strongly correlated, even though the "theory" behind each of them is completely different. There are other benchmarks that could matter, but if you're going to pick one to work with, this thing is the best available.
3) The 2 x 20 vs 5/20 vs whatever isn't that important. What is important is that you can do it reproducibly and that you are willing to do it. If you really want to know your 1-hour max power, obviously the best way of figuring that out is to do a 1-hour max power TT. But hardly anyone is happy with the thought of that or can even do it without the mental stimulation of a real race. So we have these proxy tests (2 x 20 or 5/20) that are approximations that are a lot easier on the mind and get very similar results. As long as you are doing the same test under similar conditions, your reproducibility from test to test is more important than the variability between your test method and The Grand Truth of your 60 minute power....in part because even knowing your 60 minute power exactly is just an approximation of the physiology.
I could go on about a bunch of other approximations, but this pretty much sums up where I've landed. I did all the reading, and I will do it again after WKO 4.0 comes out and Coggan sees fit to tell us his new, better theory...but for now, I'm happy enough to do my 5/20 test and my workouts. :-)
FYI, this is purely an intellectual pursuit fairly separate from my faith in the training program in which I'm engaged. I have no question about the value of consistent metrics, as long as they correlate well with performance - which I'm confident our FTP tests do because VERY experienced athletes such as your selves are devoted to them. You all would drop them like a hot potato if they didn't work. The truth is in the pudding.
I'm simply intellectually interested in all this and want to understand it better. There's gotta be something to it if it works so well!
@WJ: In your research what references did you discover in your research that were the most helpful in giving you some insight? I'm also curious about the impact of rest intervals on performance as well.
Again, thank you all for a very rich and rewarding discussion.
In terms of references, I like Skiba's "The Triathlete's Guide to Training with Power", and "Scientific Training for Triathletes" — he gives plenty primary references.
And of course (as WJ states) Allen and Coggan "Training and Racing with a Power Meter".
1. I want a reliable model for planning and executing races, and
2. I want a reliable model for planning and executing training in preparation for races.
These goals have, to a large degree, taken me away from the science behind the model and focused me on whether or not the model accurately represents the stress accumulating in my body. Interestingly, while I use FTP as the starting point for all this planning, I have noticed (Similar to Tim's comments) that measured FTP is not always the best predictor of performance. FTP alone (my opinion) reflects a "potential" performance better than an achievable performance in much the same way that Jack Daniel's VDOT score reflects a potential race performance. In both cases, it is necessary to have the appropriate "race/volume" training to achieve the predicted performances. Specifically to our approach to measuring FTP, our training may cause FTP to be over or under stated in terms of it's usefulness predicting stress accumulation during a race.
For example, early in the year, after pounding short intervals in the off season and crushing hill intervals to stay warm in cold Michigan rides outdoors in early April, my FTP is higher than it is in August after I have started adding more volume. I see this represented in Golden Cheetah on the Power Duration curves as gains/improvements in the left half of the curve and not in the right. I see it on the EN test numbers as having a 5-minute test (4.46) that is higher on a watt's/kg basis that my 20-minute number (3.52) when I look at a chart. If I were to use this FTP to calculate my 140.6 bike pacing, my legs would be fried on the run.
Later in the season, after I have adapted my training to include the volume, my FTP (measured using the 5/20 test) falls to a lower number, ˜5% lower, (although a number that is larger than the previous year) than it was early season. This number seems to be more reflective of what I can sustain for long course races.
Together with my goals (above), I will try to adapt my power-based training this year. Like last year, I will focus on driving my short duration power higher. However, as the season matures, I will compare the w/kg on the 5- and the 2-minute tests to determining interval targets and durations as I phase in to long-course training. I foresee more workouts with sustained efforts near FTP and durations of 20-30 minutes that evolve into even High Z3 to low Z4 intervals at 40-50 minutes.
I believe that a fresh test ˜3-4 weeks before race day, provided that all the volume work has been done and the efforts as seen on the power duration chart are within range, is an excellent tool for race planning and execution... if the training efforts as shown on the power duration curve are not consistent, then a more conservative approach should be considered.
Interestingly - apart from a level of intellectual curiosity - I have little interest in the accuracy of any of these models in physiological terms. Perhaps this is the engineer in me, but I am only interested in how well the model helps me adjust my training and plan/execute my training. For me - it is similar to the tachometer, speedometer and fuel gauge on a car... they do not help me go faster, they are tools measure the effort and manage the fuel needs of the trip.
Having said that - knowing where the research is going helps to interpret the data we see. Towards that end, I also follow Coggan on Facebook (https://www.facebook.com/TrainingAndRacingWithAPowerMeter?ref=br_tf). His posts are a little bit defensive and not always practical... but he is also posting some snippets of the research he is doing and foreshadowing how (I believe) FRC (functional reserve capacity) will be a major component of his new approach to power duration modeling in WKO 4 rather than driving everything from FTP. Here is a link to a PowerPoint on the subject: http://www.slideshare.net/TrainingPeaks/the-new-power-duration-model-in-wko4.
** please take all this with a grain of salt - I have not achieved a fraction of the success executing as the other folks posting here. I have spent a lot of time thinking and reading and talking on this subject and am just sharing some ideas/thoughts hoping to further the conversation. **
So as Rich suggested, I checked out Coggan's FB page, and found this graph, which demonstrates our belief that raising one's FTP will have a significant impact on the ability to race over periods longer than an hour. The title of the graph says it all:
https://www.facebook.com/TrainingAndRacingWithAPowerMeter/photos/a.390014401020864.93440.378267952195509/912206232135009/?type=1&theater
Hunter Allen's book, the documentation for Appolo RaceDay, and (to a lesser extent) some of Joe Friel's background material are pretty good. There's actually a set of four Hunter Allen youtube videos that were made and published through Training Peaks when WKO 4.0 was supposedly going to come out (maybe 12-18 months ago?) that points out some of this stuff and talks about the failures of many of the older models (favoring his new, unpublished one).
I'm not an engineer, but we chemists are trained in the basics of data analysis and modeling.
https://www.facebook.com/TrainingAndRacingWithAPowerMeter/photos/pb.378267952195509.-2207520000.1420142483./827916983897268/?type=3&theater
(This graph is amazing in its own right, because the data is from an 80+ yr old WR time trial rider!)
Another great thread, thanks for the discussion! A few notes:
Functional Power Threshold
. So the key is to do some flavor of testing to determine about where this threshold is. Then, once identified, do training at or near this threshold effort because we know that at this threshold we are recruiting a lot of muscle fibers, forcing them to be get better at what they do.
As you ride harder / produce more watts, your body recruits more and more muscle fibers to do the work you're asking of it. At some level of effort, all of your slowtwitch fibers are being put to work and your body recruits fasttwitch fibers. These fibers produce more lactic acid, fatigue more quickly, etc, and after a few minutes above this threshold (ie, the point above which we are recruiting FT fibers) things begin to go downhill quickly. We've all been there
Smart folks have determined that lactate threshold is a good enough indicator for where this max muscle recruitment is happening. But true lactate threshold testing is impractical for most of us. More importantly, HR is a reflection / symptom of the work performed, not a measurement. Your HR can elevated or depressed for any number of reasons for any workouts. What we are interested in improving is our ability to perform work, ie put out more watts, so it makes more sense to focus on the work performed (watts) rather than the symptom of the work performed (HR).
So these same smart folks have determined that a 1hr TT test is good enough measurement of this "work threshold." However, the rather significant practical problem is that a true 1hr TT effort is probably more impractical for all of us than a lab tested LTHR -- admin, safety, route considerations, etc.
And so there are several "good enough" ways of determining FTP through testing. If I can get a good enough number with a 5/20 test, or a 2 x 20' test, or a TT up my local hill and subtract 7%, or a TT up another longer hill and subtract ~2%, or TT up that 7% hill in the aerobars on my tri bike and subtract 10%, or don't do formal testing at all and instead use software tools to help me approximate my FTP...it's all good.
But as for the science, for me, it's all about:
For some reason, this morning I came across these two links, the first of which shows Coggan's initial attempt at outlining both his proposed zones, and the physiologic impact of training in each of them:
http://freewebs.com/velodynamics2/traininglevels.pdf (this is a document from 15 years ago, which is an extraordinarily succinct statement of the reasoning behind using a 60 min TT as an "anchor point" for power zones, training based on them, and the impact of that training on muscle physiology and cardiovascular system)
http://home.trainingpeaks.com/blog/article/power-training-levels (this is just a prettied up version of his earlier post, published 7 years later.)
References aren't given for the assertions implicit in the impact table, but given Coggan;s history as exercise physiology Ph.D, I assume he has good justification for it all.