Fast Talk Episode 178: The Good, the Bad, and the Ugly of Chronic Training Load (CTL)

I really feel this episode needs a rebuttal. A lot of the ‘bad and ugly’ @trevor presented have nothing to do with CTL, and even his description of CTL was very lacking. Reminds me of a quote attributed to albert Einstein, “if you can’t explain it simply, you don’t understand it well enough.”

CTL is an aggregation of duration and intensity meant to be more useful than just miles(km), kilojoules, or hours over longer time frames. @chris said in the pod that knowing how it’s constructed is very helpful. That was just after Trevor’s explanation of CTL and the concepts it was based on. I’d argue that the concept is way more important and useful than the calculations, but if the calculations and concepts CTL is built on are useful Trevor should have presented them accurately.

I’ve often suspected Trevor has a bias against Coggan’s concepts and tools; his description of NP (normalized power) and its formula, while dismissive, wasn’t wrong (that I heard). His description of the calculations for TSS, absolutely was wrong. I don’t understand how he got it so wrong. He didn’t even mention IF and described the different training zones as being separately weighted ‘buckets’… that sounds like Bannister’s TRIMPS. the equation for TSS is actually very simple:

TSS = ((s x W x IF)/(FTP x 3600))x100
s: seconds. W:NP in watts. IF(intensity factor): NP/FTP.

A criticism the podcast made is that a TSS value doesn’t describe the nature of the training that was done with the implication that this fault continues downstream to CTL. There was the example of a “really slow” 4 hour ride accumulating 150-200TSS which is actually riding at ~62-72% of FTP which is a pretty big range smack in the meaningful part of zone 1 in the 3 zone polarized model compared to an interval session that accumulates 150 TSS in 1-1.5 hours. This is absurd and basically impossible unless your FTP is wildly underestimated. This was pointed out later in the podcast, as one of the drawbacks of CTL, that those values rely on an accurate FTP estimate but that has been the expectation since its inception. Coggan often says GIGO garbage in garbage out. And, it is a worthless argument, using miles per week means you need an accurate distance measurement so I guess that’s a major drawback to distance…

Is CTL being misused by athletes? Probably. Coaches? Yeah, probably. Trevor? Yes, unequivocally. any discussion of CTL and its benefits needs to include discussions about ramp rates and their impact and meaning as well as the rest of the PMC. An understanding of average and individually sustainable CTL ramp rates is very protective against non-functional overreaching or overtraining. Understanding how ATL(acute training load; basically the training load over the last week or how much fatigue the athlete is experiencing)drives changes in the CTL overtime and how it impacts performance short term. Or, TSB which is the relationship between ATL and CTL and is very predictive of athletes relative performance.

Really the PMC should be a single topic if the guys were trying to actually help listeners. The legitimate drawbacks of CTL presented are actually already addressed in the PMC. Teaching listeners how to use the tool as intended rather than portraying it as broken then not providing any hint of an improvement is pretty lazy. I’m pretty sure Trevor knows that too because he used it correctly when he was talking about his athlete who he trained to beat up on his buddies annual ride, he built the CTL up pretty high (80’s) then did a big dose of training to peak CTL (~90) (and ATL) then allowed both to drop as TSB came up. That is the classic peaking cycle described with a very useful concept.


1 Like

This continued argument that TSS (and NP, CTL, ATL, TSB)is arbitrary and not rerflective of physiology has got to be can calculate with pretty good accuracy how many calories a rider used and from what substrates to generate that power. with way better accuracy than from heart rate. You can use power data to quantify fast vs slow twitch muscle fiber composition almost as good as a muscle punch biopsy. This isn’t arbitrary it is very useful estimation based on sound principles.

CTL can be informative for sure, but I think the point being made is it’s secondary. It’s better to get the training distribution right first. If your principle goal is to maximize CTL, there’s only one distribution available (loads of sweetspot). Is that the best type of training? Certainly not in all cases (some would argue never).

Once you have identified the proper distribution of training for your fiber-type and goals, you can build CTL over time. But now we’re back to basically increasing volume (assuming we don’t want to change the distribution). So, the metric’s utility is lessened.

On a related point, I find Xert’s metrics more constructive than TrainingPeaks’s. XSS tends to value endurance work more which helps moderate the desire to do high intensity work. For example, I did two hours of zone 2 on Saturday that resulted in 102 TSS but 139 XSS.

1 Like

@fazel1010, Thanks for chiming in. I think hear what you’re saying; chasing CTL at the cost of executing a well designed training plan is a bad idea.

It sounds like your idea of a well designed training plan has a single “proper distribution of training” intensity for an athlete. I don’t agree with that, I think a given athlete needs several periods with, sometimes extremely, different intensity distributions. But, plan design aside I think that is exactly where CTL is so needed.

For instance: A rider who is following a fairly stagnant very polarized distribution of intensity, with volume changing during periods of base, build, peak and taper might have very predictable weekly TSS and CTL ramp rates. You can drive the PMC just fine that way but you probably don’t need to. Once that athlete gets into a period of heavy racing however, that predictability is shot. TTs and crits, don’t fall easily into the polarized approach and road races won’t either if they are hard or our athlete is active in the race. So it is that during the race season your pmc isn’t being built on a polarized distribution but CTL and the other metrics still give guidance and quantification. They basically help you to decide if you want or need to do recovery or endurance work mid week to either prop up CTL and hold on to fitness or spend some CTL for better freshness at your next race.

As far as Xert, Armando built a really cool tool, you might find that its a bit derivative but hey, he saw limitations of the work done by Coggan and addressed them. I watched some of the discussions between the two unfold on the wattage forum years ago. To your specific example, I speak the language of TSS I understand 2 hours endurance pace in terms of TSS 90-100ish is where I’d like to see that value. Comparing it to 139 in the Xert software needs a translation because I look at those numbers and don’t think bigger is better its just a different scale.

Another great episode thank you.

Apart from experience and history of working with an athlete, are there signs you look out for about who would do better with a higher or lower training load?

If the athlete has a longish history of power data you can get a lot of clues from the PMC history.

Some age ranges work better with generally lower CTL plateau/peaks. Juniors, some masters…

The amount of physical activity and work/life stress lower the CTL plateaus. Though sometimes if peaking for something important a bigger peak CTL and longer taper works really nicely.

Thanks @hammervalley - agreed.

I was thinking about when you don’t have data or history together, what might guide you initially in deciding whether to work towards a higher or lower CTL. Age and training age are definitely factors.

Hi @hammervalley,

Thanks for the message and for starting this conversation. @chris and I really do appreciate this sort of feedback. It keeps us on our toes and ensures that we ultimately provide the best information. You had some very good points. I don’t know if you’ve heard today’s episode or not, but we did post a clarification that was motivated by your message.

You are right. I botched TSS. I’d like to blame it on the fact that we recorded while I was on vacation, but the fact is I care deeply about our standards. I always want to have our ducks in a row before we enter the recording studio. So there was no excuse for getting that wrong. As you mentioned, I got TSS mixed up with Bannister and Lucia’s earlier TRIMP concepts. I should have revisited the various load/stress metrics prior to our recording.

I’m just thankful that that mix-up doesn’t impact the primary message we were trying to convey. It does sound, however, like the message also needs clarification.

I really don’t want you to think we were attacking CTL. I think it’s a valuable metric that can be used in very powerful ways. We have actually, as you suggested, dedicated whole episodes to the PMC and how to use it effectively. It goes back a bit, but we addressed it with Dirk Friel in Fast Talk Episode 19: Training as a Numbers Game, in Fast Talk Episode 119: How to Use Data to Make Better Training Decisions with Tim Cusick, in a An Introduction to Training Metrics Part 2, and in How to Use TrainingPeaks Performance Management Chart this winter that at 26 minutes only allowed me to skim the surface.

Chris and I did discuss if we wanted to address the PMC in this episode, but we decided that including PMC would turn it into a 3-hour Fast Talk episode and detract from the primary message. That’s why we made the choice to leave it out but reference those previous shows where the PMC is addressed.

I do have a lot of respect for Dr. Coggan and his concepts. He has made an enormous contribution to the science and practice of training. If Dr. Coggan and I disagree on anything, it’s that I’m a little more focused on heart rate (hence TRIMP was on my mind) while he’s more focused on power. But that’s about it. We’ve had Dr. Coggan on Fast Talk and I was happy to see that we had little to argue about.

The message we were trying to convey in this recent episode on Chronic Training Load was actually a pretty simple one: CTL is a very useful metric but shouldn’t be the goal in our opinion. Treat it as the goal and there’s a real danger of getting off track. We decided to do an episode focused on that message because in the last couple months, we’ve received a surprisingly large number of messages from listeners and members who have been very concerned about their CTL. There was an honest belief from those riders that if CTL dropped or if it wasn’t high enough, they wouldn’t be able to perform. That’s what we wanted to address. Our hope was to both calm their fears and to help them fixate on that singular number a little less.

Also one quick clarification. I fully agree with you about the value of dropping CTL and raising TSB as a peaking strategy. But that wasn’t what was going on with the athletes I described in the episode. Of course, it’s hard to show that on a podcast. Here’s the CTL graph of that first athlete:

His ride was in early August. As you can see, we actually brought him down below a CTL of 65 in July. So we actually had to build him up to a 72 before the event. You can also see from May onwards, we were keeping him at a lower CTL than we were in the spring. He was performing better at that lower level.

Thanks again for the message! As I said, we really appreciate this feedback.



@gerrard I was really trying to formulate a response but I didn’t have a great answer.

I think the CTL you can reach is really a function of the ramp rate you can sustain and how much time you’ve got. Which just changes the question to how do you know if an athlete can sustain a smaller or bigger ramp rate? I think you just follow sound training principles, and add or back off as the athlete gives feedback (and you build a dataset).

As an example check out the graphic @trevor posted, looking at January you see a normal maybe conservative ramp rates, then a week of roughly +13-15 then a flat couple weeks to another +13ish jump over 1.5weeks. That example would smoke a lot of athletes, but a steady build of +3-4 TSS per week over that 11 week period would have put that athlete at 100 TSS (minus and rest weeks needed) very conservatively he could have reached the peak of ~87, without getting smoked.

It’s not just the number but how you get there.

The ramps in May and July look way more progressive, I don’t think it’s at all surprising that the athlete rode great in August.

@hammervalley couldn’t agree more with you! To be honest, I probably use the rolling ramp rate graph for my athletes more than the PMC itself. I find it very valuable. Here’s the ramp rate for that same athlete:

What I have found in the couple years I’ve been using this graph is that the ramp rate is highly individual and learning what’s optimal for each athlete can make a big difference in their training. To that point that I’ve considered creating a custom version of this graph for each of my athletes with horizontal markers for what I know works best for them.

@gerrard I think @trevor and I are coming to the same conclusion which is; you don’t KNOW how an athlete is going to respond until you try.

I finally listened to this podcast, and I thank you for it. I have at times become addicted to CTL and TSS. With limited time, I often worry when my CTL is going down, and the worrying is not helpful. I feel some relief.
I do think TSS/CTL are useful ALONG with a well structured plan. One thing that wasn’t discussed was motivation. Seeing my predicted CTL and actual CTL if I follow my plan, sometimes keeps me going and pushing to do that next workout. I also really appreciate ramp rates as a metric. I’m still going to ask the following questions:

  1. In the episode, you routinely mentioned CTLs of 70-100+. I’m a Cat 4 Master’s athlete and I usually get to a CTL in the 50s, occasionally 60 if I do a training camp. Does this mean that my training plan is necessary all wrong, because my CTL is too low? (I think not, but I’m asking anyway.)
  2. I once heard a rule that went something like, “You want to build your CTL to the point where it will be 1/4th of your target race TSS.” Have others heard that? Is there any truth to this? (Not as a substitute for a good training plan, but as a tool for constructing one.)