Since it is a sum of your TSS for last X days divided by X, couldnt you just have a yet bigger training day (X+1) days ago that dropped from the "X days" window?

TrainingPeaks says that CTL is an " exponentially weighted average of your last 42 days." So you need to compare your most recent workout to the one you did 42 days ago to see why the number changed. Its a rolling number so while you are adding workouts, you are also dropping off old workouts.

And, your fitness probably did not decline. CTL has some uses but in the end it is just a number that can be gamed a bit.

If you had a huge weekend of TSS (ex:Rode a century, ran a HM) that increased your CTL, then the next few days should bring CTL back down a bit.

Well I just completed my ironman build (for IMLP) so yes lots of huge weekends so I guess that explains it.

Kinda depressing to see the number going down during taper (obvious fatigue going down too).

CTL (Chronic Training Load) measures training load, which can represent fitness, but a drop in CTL does literally mean your fitness dropped, nor does a rise in CTL literally mean you are more fit. Look at someone over training, they are gonna see rises in CTL while loosing fitness at the same time.

CTL is just an exponential moving average of the last 42 days of TSS. If your TSS today is higher than CTL, it will go up, if it is lower than CTL, it will go down. If you did a workout and your CTL went down, it just means that the workout had less TSS than your CTL.

During your taper, expect your CTL to go down. That is part of the plan as you increase your form. Your fitness at the end of a taper will also be lower than it was during peak training. Even though you may loose a little fitness, you will shed all the fatigue you were carrying during peak training and perform significantly better on race day than in peak training.

Do you really think your "fitness" has declined?

You are putting to much stock in the Training Peaks terminology.

TSS higher than CTL- CTL rises

TSS lower than CTL- CTL drops

Part of the flaw in TSS/CTL/ATL is that it does not give equal weight for very hard and short workouts.

On these workouts you tend to warm ups and cool down slowly- very slow rest in between efforts (or even descending)- none of this generates a lot of TSS.

It's just a number it's not your fitness

---

Brian Stover USAT LII
Accelerate3 Coaching
Insta

Another thing to consider here is that you are looking at the impact of yesterday's workout on TODAY's CTL. CTL is calculated for a given day using that same day's TSS. If you did not bike today, your bike CTL will always go down that day, regardless of what you did the day before.

It's an exponentially weighted moving average with a time constant of 42 days. It considers more than just the last 42 days, but the further in the past a workout was, the less it counts.

In an interesting (to some) quirk. It's actually easier and less memory intensive to calculate an exponential weighted moving average than a normal moving average with a cutoff.

An exponentially weighted moving average can be restated as (x * today's input [tss in this case]) + (y * yesterday's moving average). Where x and y depend on the time constant chosen.

If you were doing a moving average where the measurement dropped out after some number of days, you have to store in memory what the input was for each of those days.

And yes, your answer to the original post was correct in meaning.

AHHH this is the key. I assumed the current day was "start of day" numbers vs including the current day's TSS (e.g. TSS from Day 1 is reflected in Day 2's ATL, CTL, TSB.

Though something doesnt make sense still if TSB = CTL - ATL

My cycling stats for Monday and Tuesday"

MONDAY
ATL: 51.5
CTL: 50.2
TSB: -8.7
TSS: 0.0

TUESDAY:
ATL: 56.0
CTL: 50.9
TSB: -1.4
TSS: 82.6

So it appears ATL and CTL update with current days TSS applied and TSB is trailing a day (assuming that is to show how "fresh" you were to achieve that day's TSS?).

Bingo... ATL and CTL include today's TSS... TSB does not include today's TSS and is simply the difference between yesterday's CTL and ATL

Thanks for the help. Makes sense now

Totally agree. I've gotten tremendous value from CTL/TSB as being a sanity check for avoiding unsafe overtraining, especially as I am slowly getting back into running and swimming and am wont to ignore the stress created by going for 20-30 minutes at 110%-115% of current threshold (slower than historical easy paces).

That being said, I have found that using CTL as its own end encourages Z3 efforts far and above all else (as easy efforts build minimal TSS and extended efforts at threshold and above are too hard to consistently replicate).

YMMV of course, but TSB especially has been great for me as a dope who always wants to go hard, or at least "medium".

x2.

Plus, you need to keep in mind that in PMC terminology, "fitness" (CTL) doesn't equate to performance ability. Theoretically, performance ability = "fitness" (CTL) + "freshness" (TSB). Performance ability increases during taper because your "freshness" (TSB) increases. The small decrease in CTL is irrelevant if you're tapering correctly.

You can have a high CTL, but can't perform for crap, because you're too fatigued (not "fresh" as indicated by a large negative TSB). It's only after fatigue dissipates that you're able to perform to your potential. This is why the length of optimum taper isn't determined by the length of the race, but instead it's more dependent on the amount of fatigue the athlete is carrying prior to beginning to taper.

---

____________________________________________
Don Larkin
Reach For More
http://www.reachformore.fit/
USAT Lvl1 Coach, NSCA-CPT, NASM-CPT, BS Exercise Science