After the commitment, the start of my journey towards my greatest physical challenge was to design a training plan. While the Haute Route challenge is daunting, its great benefit is singular clarity of the goal.
When I started competitive cycling, a coach taught me the basics of structured training. After a year and a half, I went “self-coached” and designed my own training routines. My big problem, however, was the absence of a true goal beyond just “getting better:” I didn’t really have “A” races and nothing truly specific to target or train for. This year is very different: Starting from January 1, 2013, I have 33 weeks to train for a 7 day stage race, focused completely on climbing, during August 18-24.
Knowing the goal, I designed my training plan through a hierarchical 3-step process working backwards from the ultimate goal. Today is about step 1.
(One caveat: I am no training expert so my plan just reflects my best amateur attempts. But I do adhere to one guiding principle: Do only that what you understand and can rationally explain. Below I explain what I believe I understand but do comment/email me with corrections.)
Step 1: Quantify the aggregate training goal: Training Stress Score TSS
Training should be specific to the goal. This means that training should simulate the target goal/event as close as possible. I also believe that if the goal and training can be quantified, it is easier to execute, track and modify if needed.
There are various simple ways of quantifying my HR goal via aggregate demands and workload for the 7day race:
- 21,360 vertical meters, about 3000m per day.
- 25 to 27 hours in the saddle, about 4 to 5 per day.
- 866km, about 125km per day.
While informative, these simple metrics are insufficient to design a training plan. Distance is the least informative training metric as a flat km is very different from climbing one km at 10%. Time is a better gauge of workload but are these easy or hard hours?
I found it useful to break out the 25-27 hours into 17 – 18 hours on individually timed climbs (cols) and the remaining on connecting descents, flats, and rolling hills. (I broke out the true timed climbs using the data on the HR website of each climb by each rider last year. I also estimated, using a simple linear model based on my past climbing performance of power and velocity versus % incline, what my best climbing times would be.)
To make this more specific and measurable, my best estimated times assume a close-to-threshold (more on that in the next blog) power level. This is clearly an upper bound as it is unlikely that I can maintain that performance over 7 days. But this gives me a good estimate of my Training Stress Score (TSS), which is a relative metric of workload customized by person: riding 1 hour at your maximum power (= your Functional Threshold Power FTP) equals a TSS of 100. Training Peaks further explains:
By using TrainingPeaks’ Training Stress Score (TSS) system you can gain points for virtually any sport or activity. This allows any endurance athlete the ability to quantify their workouts based on their relative intensity, duration and frequency of workouts. One single value can now represent how hard, and how long you worked out. 100 points earned by a pro is relatively the same as 100 points earned for a beginner because TSS is relative to each person’s threshold.
Now my training goal becomes very clear and measurable:
Train to achieve a weekly TSS of at least 1700-1800 = about 250 TSS/day.
On the one hand, this is a lower bound because I only count true climbing time. On the other hand, it is an upper bound because I assume the true climbing time is at my max FTP. It will just have to do!
The TSS goal is clear and measurable. Herein lies the great benefit of training with power: my power meter tracks exactly my TSS, adjusting automatically for the mix between intensity and duration. (For finance affectionados: intensity is measured as a mean-variance utility: average power relative to your threshold as well as a boost for power variance. Perhaps more on this in future blogs.) And, as one gets stronger, one should adjust one’s maximal sustainable power level (FTP) so the training adjusts to your improvement.
(Next blog: Step 2: Building the training plan by week: Periodization)
Ex-post Haute Route Alps 2013 Update: It is harder than 250TSS/day: My first day was 470TSS, 2nd day 402 TSS…
Jan Van Mieghem 
Very interesting. I will ask the silly questions: (1) You can simply add the TSS scores for each day of the week?; (2) I mostly look at the data in TrainingPeaks and sometimes instead of a TSS, it gives me an “hrTSS” score. Is that an hourly TSS score?
Patricia: I may explain this later further but here’s the skinny:
(1) yes, as a fine approximation you can just add training stress scores over a small set of contiguous days to arrive at an estimate of workload over that short period.
To be precise, however, you should consider the TSS per day as a time series with physiological adaptation and decay effects. Doing so (and I’m throwing a lot of detail under the rug here) gives rise to exponentially-smoothed moving-averaged values like Acute Training Load (ATL), which is a better estimate of your training load over the “recent past” days (about a week) and thus “fatigue,” and Chronic Training Load (CTL), which takes a longer window of past days (about 6 weeks), estimates your fitness. Here’s the math:
Chronic Training Load = [Todays TSS * (1-e^(-1/42)] + {Yesterdays CTL * (e^(-1/42)]
Acute Training Load = [Todays TSS * (1-e^(-1/7)] + {Yesterdays ATL * (e^(-1/7)]
If your TSS were constant over time, both your ATL and CTL would converge towards that TSS value. If TSS exceeds CTL (or ATL), then CTL (or ATL) will rise as must happen during build-up; otherwise they fall.
Training stress balance (TSB) = CTL – ATL is an estimate of your "freshness" or "form". During build-up, TSB < 0…
(2) I use WKO+ but I don't use Training Peaks so haven't seen this hrTSS…
hrTSS is TSS based on heart rate, intensity factory based normalized heart rate and threshold heart rate. A far cry from power, but better than nothing I guess. It is reasonably accurate for steady state efforts. It fails miserably for high intensity work.
In retrospect, do you think you should have been training to a higher TSS, based on the reality of the event (i.e. it was that much more difficult)? I mean, what would you do differently if you were doing it again (like in 2015 perhaps…?) 😉
Excellent question Gerry. In retrospect we know that a typical HR day is > 400TSS, with less for the TT stage and last stage.
Clearly, the race demands go beyond what you train for, so the question is what CTL you want to achieve/be at when starting the HR?
As you notice in a later post, I think I reached 180 CTL and that is MASSIVE for me. I could not aim higher… nor should I.
The question is: would it have been better with LOWER CTL? i.e. LESS training volume?
In hindsight, I think so: I didn’t rest enough in July — actually I rode every day for some 40+ days in a row just to “make the plan and the numbers” + I rode in Geneve on Wed, Th and Fr before the start on Sunday. If I were to do it again, I would always take a rest day per week + I would actually be off the bike for at least 5 days before the start.
I did well my first day, but quickly started deteriorating. The stolen bike and not riding your own bike gives me some excuse, but I think I just was too tired at the start. It is better to be fresh and rested.
I also would change the nature of my training to be more climbing specific: (as I wrote in some “ex-post” blog) building towards 3 x 1hr climbs = 3 x 1hr LT (a typical HR day). For that I would want to live near mountains–just too painful to do in the flatlands of Chicago. But you can and you did train more like that!
Last, what DID work, where the build-up blocks. Definitely keep them.
Rob said the same think about ‘volume’, but he was doing really extreme amounts of time on the bike. I’m interested to hear that you would also reduce it. I’m going to do a 3-wk on, 1-wk off (or recovery anyway) program this coming year, as opposed to the 4+1 we did last year. I think the extra recover will be helpful. I also rode too much before HR and paid for it right from the 1st day. Just never felt fresh the whole time.
And, like you, I deteriorated throughout the week, while everyone around me was getting better (or so it seemed). This was is difficult to quantify for me. Is it just not enough miles in the legs? Not enough rest (look how Horner did in the Vuelta this year with 4 or so months off)?
Anyway, it’s all academic right now because neither of us is doing HR next year.
Interesting to know that Rob said the same. We all probably overtrained–the typical error of newcomers. I know believe that, whenever in doubt, better to undertrain and be fresh, than overtrain. Definitely blocks of 3wkon, 1wk off; 4wk on is too long.
I also felt everyone was getting better–especially on last day I was terrible. But, in hindsight, it may be that I just deteriorate more than the others.
I’m not thinking about it either; I just know that I won’t train again, EVER, as much as I did in 2013. I recall that there were days in July that I told Shannon: “Darn, I’d rather go to work at the office, than do my training now.” (I know pro’s talk about “another day in the office”; for me I mean my real office, not the bike!) You know it’s not good when the hobby becomes worse than work… That will never happen again.
It never got that way for me, but my office is home, where I did much of my riding! It was a chore to be on the bike 6 days a week, though, but I’ll be doing that again this season, I think. I don’t think I overtrained, by the way. I think I probably should have taken it easier before HR, but I doubt I trained too much. I’m a slacker at heart!
Your articles are really interesting to read, now that I’ve got a power meter. Suddenly, they make some sense 😉
Excellent Blogs this one, I have been recently attempting to create my own training plan. this blog will be a big help on getting started on making a personalized plan for my goal!