To train for the Haute Route 2013 challenge, I designed a training plan through a hierarchical 3-step process working backwards from the ultimate goal. This blog is about tracking and forecasting fitness, fatigue and form, given the training plan.
Recall that Step 1 defined the aggregate training goal as simulating a training stress score of 250/day (1700-1800 for the 7 days). We know that 100 TSS corresponds to one hour riding at your max. Averaging over hard and easy rides, a good estimate is that you gain about 1 TSS point per minute or 60 per hour of riding (again, hard training rides earn higher points but a reasonable estimate is that 250 TSS is about a 4:10 ride).
Step 2 used periodization to come up with a weekly TSS training plan. We saw how Chronic Training Load (CTL) measures your fitness or long-term response to TSS training load. The cool thing about these metrics is that they can be calculated and thus forecasted. Recall:
CTL(t) = CTL(t-1) + (TSS(t) – CTL(t-1))*(1-e^(-1/42))
This equation says that the one-day change in CTL equals the incremental load (over your CTL) * 0.024, so you only gain about 2.4% of your incremental load in fitness per day. (That’s why it is so important to build a good base during the winter/early season as you cannot build CTL quickly.)
Now you can enter your planned TSS in a spreadsheet and project out your CTL. But before doing so, let’s introduce a third metric (trust me, it took me a while to really “get” these metrics but in the end it’s simple and useful):
Acute Training Load (ATL) measures your fatigue or the short term response to TSS training load. It is similar to CTL but focuses on the last 7 days (1 week) instead of 6 weeks (42 days). The computation is similarly done using exponential smoothing but only over 7 days:
ATL(t) = ATL(t-1) + (TSS(t) – ATL(t-1))*(1-e^(-1/7))
Clearly ATL responds much quicker to a TSS increase than CTL: the one-day change in ATL = the incremental load (over your ATL) * 0.133, so you experience about 13.3% of your incremental load in fatigue per day. As the figure below shows, in about 2 weeks (16 days) your ATL is at 90% of a constant TSS/day; but it takes 3 months (96 days!) for your fitness to reach that same 90%! The good thing is that if your TSS is below your ATL, then you quickly recover!
That leads us to the last metric: Training Stress Balances (TSB) which relates fatigue to fitness and measures “Form”. TSB is simply the difference between fitness and fatigue:
TSB(t) = CTL(t-1) – ATL(t-1)
Positive TSB means you are “on form” and have higher fitness than fatigue. The goal of a training plan is not only getting the fitness (building CTL) but also timing it such that you have positive TSB before the start of the race. Peaking exactly accomplishes that: after a peak in workload, fatigue and fitness, the workload is reduced below the CTL one to two weeks before the race. Given that CTL moves slowly, fitness falls little but ATL responds quickly (falling below CTL) leading to a big positive spike in TSB.
Using the weekly training plan, all three metrics can be projected for the Haute Route. (For simplicity, I have kept the TSS/day constant within each week.) You clearly see my remaining three training cycles: Build 2, Build 3, and Peak, when form (TSB) is finally turning positive: READY TO RACE! Beautiful, no?
One sad realization: at best, my max CTL will be just under 180 TSS/day at the beginning of August 2013. This is far short from my original goal of 250, which would require riding more than 4 hours each day for three months. Oh well, it is what it is: one can only do so much as an amateur and this will just have to do!
Awesome! This exactly what I was looking for. I wonder if Training Peaks or the like have forecasting functions built in. Regardless, formulas w/some scripting are perfect for DIY!