A few years ago I was coaching a woman who was training for Denali. She had the fitness, she had the time, and she had the work ethic. What she didn’t have was patience for what I was asking her to do, which was a lot of slow running in a city park near her home. About eight weeks into her program she called me, frustrated. She told me she was running circles around the same loop, going slower than she’d ever run in her life, and she had no idea whether any of it was working. She wasn’t lazy and she wasn’t doubting me, exactly. She was doing the work. She just couldn’t see the work doing anything.
That call stayed with me. Because she was right to be frustrated. One of the most important variables in an endurance athlete’s program is the aerobic threshold, the ceiling under which most aerobic base training takes place. In terms of heart rate, it’s the number that tells you whether your easy runs are easy enough, whether your training zones are calibrated to your current physiology, and whether the slow miles you’re putting in are actually building the engine you want. And yet, for most athletes, this number is invisible. It gets established once, in a heart rate drift test or a lab visit, and then it sits there as an assumption while everything else in their training shifts around it.
That gap between what we know matters and what athletes can actually see is the gap I have spent a long time trying to close. Continuous AeT, the capability now live in the Uphill Athlete training group dashboards, is the closest I’ve come. This is the story of how it works, why I trust it, and how it came to exist in the first place.
The problem with one-and-done testing
If you’ve trained with us for any length of time you already know the heart rate drift test. It is the standard field method we recommend, it correlates well with lab-based metabolic efficiency tests, and we include it as one of the first workouts in nearly every aerobically based training plan we offer. The heart rate drift test is the tool that the manual aerobic assessment guide on our site is built around, and it remains the right tool for many situations: establishing an initial baseline, validating a training block, confirming a suspected aerobic base deficit, or any case where an athlete wants a clean, dedicated read on their threshold.
But the heart rate drift test has costs. It takes about an hour and a half once you account for warm-up. It needs to be done well-fueled and hydrated, on consistent terrain, ideally on a treadmill so the variables stay clean. It needs to be done when the athlete is rested, not chasing it after a hard week. And then it needs to be done again, four to six weeks later, to actually be useful, because the aerobic threshold is not a static number. It rises with consistent base training, falls with overtraining or illness, and shifts in response to almost every meaningful change in an athlete’s program. The whole point of measuring it is that the measurement is supposed to track something that is changing.
In practice, what I see is that athletes do the test once. They get a number. They set their zones from that number. And then they train against those zones for months without retesting, because the test is logistically annoying enough that it keeps getting deferred. Six months in, the zones may no longer reflect the athlete’s actual physiology. The stronger and more consistent the athlete becomes, the more outdated those zones tend to be, and the more likely they are to be training too easy in zone two without realizing it.
The athletes who do retest regularly are usually working with a coach who keeps it on the schedule. And what I noticed, after enough years of coaching, was that I almost never needed the formal test result to know what the answer was going to be. By the time my athlete was ready to do the test, I had usually already decided what their AeT was, just from looking at their training data.
What a coach actually does
When I review an athlete’s training file, I am looking for patterns. I am looking at heart rate responses across many sessions of varying intensity and duration. I am looking at how the heart rate behaves when pace is steady. I am looking at recovery between intervals, drift during long efforts, the relationship between perceived exertion and heart rate, and the slope of heart rate decline at the end of sessions. None of these signals on their own gives me an aerobic threshold. Together, across enough workouts, they give me a confident estimate.
This is the same logic that underpins the manual heart rate drift test, but applied to data the athlete is already producing. The drift test isolates the signal by holding pace constant for an hour and watching what happens to heart rate. A coach reading a training log is doing the same thing in pieces, distributed across many workouts, with the noise filtered out by volume rather than by protocol. If you have enough sessions, the signal is in there. You just have to know how to look.
The trick is that knowing how to look took me a long time to learn, and most athletes will never have the time or inclination to learn it. They shouldn’t have to. The right place for that knowledge to live is not in the head of one coach who has been doing this for years. It’s in software.
How we built it
The path from “I do this in my head” to “the software does this for everyone” was not short. I sat down with a computer scientist and walked through, in as much detail as I could, what I actually do when I read a training file. We tried to codify that process into something a computer could execute. The early versions did not work. They produced numbers that looked plausible but didn’t agree with what an experienced eye would have produced from the same data. We kept iterating, and through 2025 we expanded our engineering team to refine the inputs, the filtering logic, the way the system handled outlier sessions and missing data, and the weighting it gave to different kinds of workouts. For a long stretch the project felt like it might never converge. Then it did.
There came a point where I started running Continuous AeT against athletes whose AeT I already knew, either because I had coached them through a heart rate drift test or because I had been reading their data long enough to have a strong prior, and the numbers started lining up consistently across athletes and across data sets. Continuous AeT was producing estimates that matched what I would have produced from the same data, and matched what the athletes’ own drift tests had produced when they did them. That was the moment I knew we had something usable.
What Continuous AeT is doing, in essence, is reading the athlete’s recent training the way a coach reads it. It looks at heart rate behavior across multiple workout types and durations. It identifies the heart rate region where the aerobic system is operating at its sustainable maximum without significant drift. It filters out sessions that aren’t informative. It updates that estimate week over week as new data arrives. The methodology is the methodology of an experienced coach reading a file. The advantage is that the software does it on every athlete, every week, without having to waithaving to wait for the test to be scheduled.
Why I trust it
There are two questions that should be asked of any tool that claims to measure something physiological. The first is whether the method is sound. The second is whether the method, in practice, produces results that agree with established methods. I’ll take them in that order.
The method is sound because it is built on the same physiological principle that the heart rate drift test is built on, which is the same principle that the manual coach interpretation is built on. At an intensity at or below the aerobic threshold, a well-fueled and hydrated athlete can hold a steady pace with a stable heart rate for an extended period of time. Above that intensity, heart rate drifts upward over time even at constant pace. This is not a controversial claim. It is the foundational observation behind every credible field method for aerobic threshold determination. The drift test isolates the observation in a single workout. Continuous AeT extracts the same observation from many workouts.
The method agrees with established methods because that is how we developed it. Every iteration was checked against athletes whose AeT had been established by drift testing or by lab work. When the result disagreed with the established result, we asked why, and we either improved the method or, in a few cases, identified that the original test had been compromised by something we were correctly identifying. The development process was a long, slow conversation between the software and the existing methods. It is not a black box that produces a number from somewhere we can’t see into. It is a software implementation of methods that already exist and that we have already validated, applied at a frequency that manual testing cannot match.
That last point is worth dwelling on. The drift test is more accurate than Continuous AeT on the day the test is performed. This is well-established. But the drift test gives you one number, on one day, every four to six weeks at best. Continuous AeT gives you a number every week, drawn from the actual training the athlete is doing, in the actual environment they train in. For tracking change over time, which is what most athletes actually need, Continuous AeT is the better instrument. For establishing an initial baseline or confirming a suspected aerobic base deficit, the drift test remains the right tool. Both methods are in our toolkit, and they are complementary, not competitive.
What you actually see
In the training group dashboard, Continuous AeT surfaces in two places. The first is on the progress view, where there is a trend graph showing your AeT over time. This is the graph I most wanted to exist. It is the answer to the question my Denali athlete was asking on that phone call years ago: am I getting fitter? When the line on that graph rises, the slow base work is doing what it’s supposed to do. When the line plateaus, that’s information too. It might mean you’ve reached a temporary ceiling. It might mean you’re not training enough volume, or you’re modulating poorly, or you’re missing your strength workouts and accumulating residual fatigue that’s suppressing aerobic adaptation. The graph itself doesn’t tell you which of those things is happening. But it tells you that something needs attention, and that’s a conversation worth having with your coach or in the live coach Q and A session in the training group.
The second place Continuous AeT surfaces is in the threshold delta. Continuous AeT gives you a current AeT every week. Your anaerobic threshold (AnT) comes from your TrainingPeaks data (which is often pushed from the device maker such as Garmin). The dashboard surfaces the delta between the two, expressed as a percentage. A large delta, more than ten percent of your AnT heart rate, suggests that your top-end fitness has outrun your aerobic base and that more base work is likely the right next step. A narrower delta indicates a well-developed base and is the signal that an athlete is ready to begin incorporating higher-intensity work without compromising their aerobic foundation. This means an athlete can see, in real time, whether they are ready for the next phase of their training, or whether they need more base work first. That is information that used to require two separate tests and a coach to interpret.
The dashboard has one more capability still in development that is worth describing here. As your AeT rises, Continuous AeT will suggest updates to your training zones to reflect your current physiology. As the gap between your AeT and AnT narrows, the boundaries between your upper and lower intensity zones shift accordingly. The athlete will be presented with the suggested updates and can accept, reject, or modify them. Nothing happens to your training without your consent, but the work of figuring out whether your zones are still accurate gets done for you, automatically, every week.
Where Continuous AeT is available today
For now, Continuous AeT is rolling out only to training group athletes. We want to verify it works in the wild, across many athletes, many sports, and many training environments, before we extend it. Training groups give us the right population to validate against. The reads get checked against drift-test results, against athlete-reported sense of effort, and against coaches hand-scoring athletes’ AeT.
What it doesn’t do, and what it requires
Continuous AeT needs data. Specifically, it needs about four weeks of training with heart rate data before the trend stabilizes into a confident read.. If you’ve just started using the platform and you don’t have historical workouts to import, you’ll see a blank space where the AeT trend will eventually appear. That’s by design. The system will not produce a number it isn’t confident in, on the principle that silence is better than a misleading reading.
It also needs the heart rate data to be of reasonable quality. Wrist-based optical heart rate has gotten dramatically better over the past few years, and Continuous AeT handles it well in most cases, but it still works better with a chest strap when chest strap data is available. If your heart rate file has obvious artifacts, missed beats, or sustained sections of clearly wrong data, the system will filter those out where it can, but garbage in is garbage out. If you want the most accurate read, train with a good heart rate signal.
Continuous AeT does not replace the heart rate drift test in every situation. If you are coming back from a long layoff, a major illness, or a significant change in your training environment, a manual drift test gives you a clean reset and a fresh baseline. Continuous AeT will catch up over the next few weeks of training, but if you want a confident number on day one of a new program, the drift test is still the cleanest path. The two methods are designed to work together. The drift test is your scheduled checkpoint. Continuous AeT is your week-to-week monitor.
Why this matters
There is a particular frustration that endurance athletes know well: not knowing whether the work is working. You commit to aerobic base training. You stay disciplined, keep the effort honest, and skip the gratification of fast workouts because you trust the long view. Then, weeks in, you have nothing concrete to point at that tells you the trust is being repaid. The fitness gains from aerobic base building are real but invisible without measurement. They are slow, cumulative, and structural, showing up eventually as the ability to go longer, climb higher, and suffer less. But during the building phase, they don’t announce themselves.
The job of Continuous AeT is to make those gains visible while they are happening, not to add another data point to an already crowded screen, but to answer one specific question that athletes deserve an answer to: is the slow work paying off? When the trend line rises week over week, the answer is yes, and you can keep going with confidence. When it plateaus, the answer is that something needs attention, and that conversation can begin with your coach or in the training group. Either way, you know. That’s what I wanted to build for the woman training for Denali, years ago, who was running circles in the park and couldn’t see the work doing anything. I couldn’t build it for her at the time. We have built it now.
The training group dashboard is where Continuous AeT lives, alongside the four input scores that anchor the rest of the system: consistency, training load, work:rest balance, and strength. Continuous AeT is one piece of a larger architecture whose purpose is to make training intelligible to the people doing it. None of this replaces the work of training itself. The slow miles still have to be run. The strength sessions still have to be done. The hard weeks still have to be recovered from. What the dashboard does is show you whether the work is producing the adaptations it’s supposed to produce, in plain language, at a frequency that matches the actual rate of change in your fitness.
If you are an athlete who has been training on your own and wondering whether your aerobic threshold estimate is still accurate, whether your aerobic base is developed enough to start adding intensity, or whether the patient work of the past three months has actually paid off, Continuous AeT is built for you. The training groups are where you can use it, alongside coaching, community, and the live coach Q and A sessions. The slow path is still the right path. We’ve just made it easier to see that you’re on it.
Train with us
Continuous AeT is one of several capabilities we’ve built into the Uphill Athlete training group dashboards to make the work of training more visible and more actionable. The training groups are designed for athletes who want structured, progressive training and an engaged community of mountain and endurance athletes, with access to coach input when they need it. If you want to see Continuous AeT in action, track your aerobic threshold over time, and train within a group framework that helps you understand what your data means, the training groups are the place to start. (Link to webpage for more info)
For athletes looking to learn more about manual aerobic and anaerobic threshold testing, our complete guide to aerobic and anaerobic fitness assessment covers all of them. The manual methods and Continuous AeT are designed to work together, each pointing at the same underlying physiology from a different angle and at a different frequency.