The metrics that TrainingPeaks produces using data from your GPS and heart rate monitor have potent implications for monitoring, controlling, and adjusting your training. Anyone can write a training plan. The challenge is successfully applying it to the athlete so that it produces the best results, and one of the biggest problems for coaches and athletes is assessing training load—on both a short-term and long-term basis—and the effect that load has on the athlete.
TrainingPeaks was initially set up to serve triathletes and cyclists. Exercise physiologist Andrew Coggan developed the Training Stress Score system as a simplified version of an existing mathematical model by Eric Banister using the acronym TRIMP (TRaining IMPulse). The result was called the Performance Manager. For mountain athletes, these metrics require significant adaptation to be useful—but once adapted, they represent the best fitness and recovery monitoring tool I have come across in decades of coaching.
We frequently field questions about what these metrics mean and how to interpret them. This article explains the system, its limitations for mountain athletes, and the specific adjustments we use at Uphill Athlete to make it work.
What Do the TrainingPeaks Metrics Mean?
TSS: Training Stress Score
TSS is the foundational element of the entire system. Get this wrong, and every downstream metric will be inaccurate. Like any algorithm, poor input produces poor output.
TSS quantifies the training load imposed by an individual workout based on its duration and intensity. The model uses the maximum sustainable intensity you can hold for an extended period—30 minutes for a moderately fit amateur, up to 60 minutes for an elite—as the benchmark. This intensity is your Anaerobic Threshold, also called Lactate Threshold, Functional Threshold Power/Pace, or Critical Power/Pace. Establishing this intensity accurately is vital to getting a useful TSS.
There are three ways to calculate TSS, and they differ significantly in accuracy:
Power-based TSS (pTSS) is the most accurate because it uses data from a power meter. It measures the actual work the athlete is doing. This is the standard in cycling.
Pace-based TSS (rTSS) is also quite accurate because running pace is a nearly perfect proxy for power—except when it is not. Wind, hills, and rough ground all decouple pace from actual effort. If that sounds like most of your training in the mountains, you can see why rTSS is not helpful for mountain athletes.
Heart rate-based TSS (hrTSS) uses heart rate as a proxy for work. Heart rate tracks reasonably well with effort in steady-state conditions, but not perfectly. This makes hrTSS the least accurate of the three methods—and unfortunately, it is the one mountain athletes are stuck with.
CTL: Chronic Training Load
CTL measures how much training you have been doing over the past six weeks. It uses an exponentially weighted running average of TSS, which means the training you did in the past few days affects the number much more than the training you did six weeks ago. CTL reflects the positive effects of training: the fitness you have built. Athletes who train more will generally have a higher CTL. It is a proxy for fitness—relevant only to yourself, not an absolute measure that can be compared between athletes.
ATL: Acute Training Load
ATL is a backward-looking, exponentially weighted calculation based on TSS for the last seven days. Recent training has a more fatiguing effect than training done several weeks ago, so ATL captures the negative, fatiguing effects of training. Like CTL, ATL is a relative measure—it tells you about your fatigue level compared to your own history, not in absolute terms.
TSB: Training Stress Balance
TSB is the difference between CTL and ATL (CTL minus ATL). Think of it as the balance between the positive training effects (fitness) and the negative effects (fatigue). This number should reflect your current performance potential. As anyone who has trained consistently knows, fatigue masks fitness—you cannot express your full potential when you are exhausted. TSB quantifies this in a relative way.
How the Performance Manager Works
Your actual performance on any given day is the result of a constant tug-of-war between CTL and ATL, expressed by TSB. CTL indicates performance potential. ATL indicates how much of that potential is being masked by fatigue.
The statements about the relative nature of these metrics are not made lightly. The more history you have, the more helpful the numbers become. A weekly TSS of 1,000 is high for any athlete, but the CTL, ATL, and TSB it produces depend entirely on the athlete’s history. An elite with a CTL over 150 can handle a weekly TSS of 1,200 during routine training and manage an ATL of 200 without overreaching. An amateur with a CTL of 60 might struggle with a weekly TSS of 500.
This is where the art of coaching comes in. Even with the Performance Manager, the trick is balancing chronic and acute training load to arrive at the best performance on a given day. That can only be done with experience and a good data set. The combination of hard data from your watch with subjective notes about how you feel—which you should be recording in the comments section of every workout—is the most potent tool in your training toolbox.
Why These Metrics Fall Short for Mountain Athletes
Unfortunately, heart rate is not a great proxy for work output because many factors beyond effort affect it. In steady-state running or cycling, heart rate tracks well with pace or power. But many real-world conditions routinely encountered by mountain athletes cause a decoupling of heart rate from actual work.
Steep uphills. Heart rate usually does not fully reflect the effort and muscular fatigue of hiking steeply uphill, especially with a pack. The local muscular endurance load is enormous, but the heart rate may be disappointingly low relative to how hard the work actually is.
Downhills. Heart rate drops on descents at a rate that does not reflect the significant neuromuscular fatigue that comes from running or hiking downhill. Your heart rate says you are resting. Your quads say otherwise.
Strength training. Heart rate is essentially meaningless as a measure of how hard a strength workout is. hrTSS for a max strength session will dramatically underestimate the actual training load and recovery demand.
The single-threshold problem. The Performance Manager recognizes only one benchmark—your Anaerobic Threshold—to anchor training intensity zones and calculate TSS. But another critical metabolic point, the Aerobic Threshold, differs significantly between athletes and must be accounted for to give individualized training recommendations. The Performance Manager assumes this point to be about 80 percent of AnT. In reality, the spread between thresholds ranges from 7 percent for the very fittest to over 30 percent for the aerobically deficient. This formulaic assumption introduces error for every athlete who does not match the model’s default.
To address the single-threshold problem, we strongly recommend that athletes perform an aerobic assessment and set their training zones based on both their Aerobic Threshold and their Anaerobic Threshold, rather than relying on TrainingPeaks’ default zone calculations.
Our Solutions: The TSS Fudge Factors
To make hrTSS more reflective of the actual training effect and the resulting fatigue, we developed a set of manual adjustments—fudge factors—for workout types where heart rate systematically underestimates load. These are not scientifically derived. They come from hundreds of interactions with athletes over many years and reflect our best attempt to acknowledge the training stress that heart rate alone misses.
For the sake of comparing your workouts consistently, it is important to apply these fudge factors the same way every time. That way you are always comparing apples to apples.
Aerobic Runs, Hikes, and Ski Tours
Without significant weight: Use the TrainingPeaks hrTSS and add 10 TSS for each 1,000 feet (300 meters) of elevation gain and loss.
Carrying more than 10 percent of body weight: Add an additional 10 TSS per 10 percent of body weight carried, per 1,000 feet (300 meters) of gain.
Muscular Endurance Workouts
A hard muscular endurance workout typically has a very high local muscular load at a disappointingly low heart rate. To determine TSS, pick a number that reflects the recovery time before you feel ready for another such workout.
Slow-twitch dominant athletes (those who find ME workouts take a big toll and require 72 hours of recovery with only easy aerobic work in between): score ME workouts at 150 to 200 TSS.
Fast-twitch dominant athletes (those with significant strength training or sprinting backgrounds who can bounce back in 48 hours): score ME workouts at 100 TSS.
Strength Training
General strength and core workout: 50 to 70 TSS per hour
Max strength with core warm-up: 80 to 90 TSS per hour
Climbing Workouts
We consider climbing sessions to be strength training. Heart rate is a particularly poor proxy for climbing effort.
General strength climbing: 50 TSS per hour
Max strength climbing: 75 to 80 TSS per hour
ARC training below onsight level: 50 TSS per hour (count only time on the rock)
ARC training at or near max level: 80 to 100 TSS per hour (count only time on the rock)
These fudge factors may seem arbitrary and too formulaic to apply across broad swaths of athletes. But they reflect actual workload and recovery times well enough to make the Performance Manager genuinely useful for mountain athletes. While it takes time to manually adjust TSS for every workout, we believe it is worth it—especially when assessing an athlete’s fitness and preparedness for a specific objective.
Running Power Meters: An Emerging Alternative
Running power meters have improved significantly in recent years. Most watch companies now offer this feature, and several claim their algorithms account for elevation change, making them potentially useful for trail running.
These devices work by measuring variables like vertical oscillation, ground contact time, and elevation change using accelerometers in a chest strap, foot pod, or the watch itself. They feed this data into a proprietary algorithm and produce a running power number in watts.
In exploring this space, I came across physicist and mountain runner Markus Holler, who developed his own algorithm (for Garmin) that processes the same input data with full transparency about what goes into the model versus what he believes the major brands’ models may be missing. I was impressed enough to buy his e-book and do a podcast with him. His model can even account for the weight of a pack you are carrying—important for ultra runners carrying 2 to 3 kilograms on very long efforts.
However, Markus does not think running power meters work well for mountaineers carrying heavy packs, where the muscular load is substantially higher and the training effect is qualitatively different. For heavy-pack mountain training, the fudge factors described above are likely still the best approach.
For trail runners operating on moderate terrain with minimal pack weight, running power meters represent a significant step toward more accurate TSS. The result is a more accurate CTL and a better picture of fitness and fatigue. I am very intrigued by the progress in this area and believe it will continue to improve.
What CTL Can and Cannot Tell You
At a coaching conference years ago, I listened to experienced triathlon coaches explain that they had enough data to say with conviction that completing the Ironman world championship in Kona in under 9 hours required a CTL of 150. They had similar benchmarks for other races and marathon times. This got me thinking: could we establish a similar yardstick for mountain athletes?
In our earlier years, I posted specific CTL numbers on the Uphill Athlete website as rough guidelines for different mountain objectives. My intention was to encourage people to be well prepared—at the time, there was not much of a culture around training for big climbs, and many people were undertaking mountain objectives for which they were dangerously underprepared.
I opened a can of worms. I had intended those numbers as rough guidelines. Instead, athletes treated them as hard rules. Our coaches and I spent years explaining away the apprehension of climbers who believed that not having a CTL of 100 meant they were unprepared for Denali.
What We Know
After many years and thousands of data points, some trends are visible. A CTL of 100 is roughly the price of admission for an oxygen-assisted ascent of an 8,000-meter peak. Athletes with a CTL of 40 to 50 are not ready for big mountain adventures. Holding a CTL above 100 for a month will make an athlete fitter than holding a CTL above 80 for two weeks after a week of overreaching.
What We Do Not Know
Whether you need a CTL of 120 or 130 for success on Everest, I cannot say with certainty. The accuracy of our predictive model—predicting fitness levels for specific objectives—is probably below 75 percent. We are relying on hrTSS, the least accurate method of calculating training stress. When you layer on fudge factors, the numbers need to be taken with a large grain of salt. They are rough estimates that only work when comparing apples to apples.
CTL is a decent proxy for fitness, but it is still a proxy. It distills a tremendous amount of information into a single number. Use it as one input alongside subjective feel, training history, and the judgment of an experienced coach. Do not use it as the sole arbiter of whether you are ready for an objective.
I should have known better than to post those specific numbers. I will no longer try to predict fitness for athletes whose training I am not directly familiar with.
The Bottom Line
The TrainingPeaks Performance Manager, with appropriate modification, is the best fitness and recovery monitoring tool available to mountain athletes. It is not perfect—hrTSS is an imperfect proxy, fudge factors are estimates, and CTL is a simplified model of a complex human system. But used with experience, honest self-assessment, and consistent methodology, it provides a level of training intelligence that was simply unavailable a generation ago.
The lesson is this: the combination of hard data from your watch, manually adjusted TSS using the fudge factors above, and your subjective notes on fatigue and fitness recorded in the comments section of every workout creates a training feedback loop that is far more powerful than any single metric alone. Use the data. But never stop listening to your body.