A comprehensive overview of Canova-style percentage-based training for runners

Percentage-based training is a mathematical approach to planning workouts for runners. Percentage-based methods are used by many top international coaches, most notably Renato Canova, to train runners at distances from 800m to the marathon. 

I could write a whole book about percentage-based training for runners (in fact, maybe I will someday), but the goal of this post is to give a clear, comprehensive, and readable overview of percentage-based training as a system—a set of principles that can be used to guide training decisions.

To this end, we’re going to focus on the concepts and rationale behind the percentage-based training method, as opposed to exact training calendars. This post does include an appendix with recommended workouts for every event from 800m to the marathon (even the 3k!), but event-specific full training calendars will be a project for another day.

Understanding how a training calendar comes together requires an understanding of the basic concepts anyways, so we might as well start from the foundational principles of percentage-based training.

What is percentage-based training anyway?

Percentage-based training is an approach to planning workouts that’s identifiable by its use of percent of race pace—e.g. “4 x 2 km at 95% 5k pace” is a percentage-based workout. Percentage-based training is unique in that it takes more of a mathematical approach to training, as opposed to a physiological one.

Percentage-based training is most well-known for its use by Renato Canova, though he is not the only coach who uses percentage-based training for runners—many others of the “Italian school” of training use a percentage-based approach.

Percentage-based training can look vaguely similar to typical physiologically-based training (e.g. Daniels Running Formula) when you examine 5k or 10k training, which is one reason why I think people are overly swift to dismiss percentage-based training as “basically the same” as a traditional VO2max + race pace + threshold + easy mileage approach.

Hopefully, this article will convince you that percentage based training is emphatically not the same as a physiologically-based approach; it has different principles, different goals, and uses different workouts to achieve them.

Percentage-based training gets a lot more interesting when it doesn’t resemble traditional American training, which is absolutely the case when you apply this system to events like the 800m, 1500m, 3000m, half marathon, and marathon.

Why should runners use Canova-style training?

The main appeal of percentage-based training is that it is a fully general system of training, not merely a few “magic workouts.” Moreover, this system works for running events of any distance—it works just as well for the 800m as it does for the marathon, and you can use it for any arbitrary distance in between, e.g. the 3000m.

American runners seem prone to “training fads” that get popularized by one or two successful runners. But you should be hesitant to copy any one individual runner’s training—everyone is different, after all. What you really want to use is a system that shows consistent success across different runners and different events. That’s exactly where percentage-based training shines.

Canova and other Italian school coaches have used this type of training in world-class runners in every event from the 800m to the marathon. I have likewise found it useful in my own coaching for athletes at this full range of distances for the far more modest goal of breaking personal records.

As usual, heed the following disclaimer: what follows is my own understanding and implementation of percentage-based training, and I don’t claim to speak for anyone else.

The ten basics principles of percentage-based training

You can boil down percentage-based training to ten key principles:

  1. Percentage-based training is primarily mathematical, not physiological
  2. The main problem to solve is supporting your race-specific workouts
  3. Adjacent speeds form a “ladder of support”
  4. Endurance support matters most
  5. Paces slower than 80% of race pace are general, not specific
  6. Order of operations: general training, race-supportive training, race-specific training
  7. Start from your current fitness, not your goal fitness
  8. Progress your workouts over time
  9. Pair greater stress with greater recovery
  10. Update your workout paces as you improve

We'll take a look at each in detail.

1. Percentage-based training is primarily mathematical, not physiological

Traditional American training systems usually build workouts around key physiological parameters: VO2max, lactate threshold, critical speed. Enhance these parameters, the logic goes, and you’ll get faster.

The percentage-based approach is different—it views training as a mathematical problem: given your current fitness, how do you become able to run x speed for y distance?

The answer is relatively straightforward: improve your speed at fractionally faster than x, and improve your endurance at fractionally slower than x.

One you can run long, faster, and run fast, longer, you’ll easily improve your race times.

Concretely, percentage-based training involves evaluating your current capabilities at a range of speeds surrounding race pace—often in five percentage point jumps (i.e. how long can you run at 80% race pace, 85%, 90%, 100%, 105%, 110%, etc…).

Then you construct workouts to improve your abilities at each of these pace increments.


📱 Reminder: I have a free web app for calculating pace percentages!

2. The main problem to solve is supporting your race-specific workouts

Regardless of their training system, virtually all runners agree that the most important workouts that you do—at least for short term performance—are your race-specific workouts: for example, 300m repeats at 800m pace for the 800; 1000m repeats at 5k pace for the 5k; mile repeats at 10k pace for the 10k, and so on.

So, if a 10k runner is comfortably doing, say, 6 x mile at 6:00 with 2-3 min rest, it’s clear that she’s ready to run 37:30 in the 10k.[1] And if she could improve her workout times to 6 x mile at 5:50 at the same effort level, she’d be able to run 36:30. Not a very controversial position.

The real question—the “hard problem” of training—is how can you get this athlete to go from 6:00/mi to 5:50/mi for the same specific workout?

Percentage-based training’s answer is to build support for race-pace workouts at 5% faster and 5% slower than race pace.

For our 10k runner (current 10k pace: 6:00/mi), this means doing specific speed workouts at 5:42/mi (105% 10k pace) and specific endurance workouts at 6:18/mi (95% 10k pace). If we use workouts that develop our runner’s capabilities at these two speeds, she should improve her capabilities at 100% 10k pace as well.[2]

To give a specific example, on the speed side, our runner could progress over time from 10 x 600m at 105% 10k pace to 8x800m, and later 6 x 1000m.[3]

On the endurance side, she might progress over several weeks from 5mi continuous at 95% 10k pace up to 8mi continuous at 95% 10k pace.[4]

You might object—“I can’t do eight miles at 6:18 pace!” but that’s sort of the point—no, of course you can’t do that right now; the goal is to become able to do it gradually, over time. Once you can do 8mi at 6:18 pace (and 6 x 1000m at 5:42 pace), you won’t be a 37:30 10k runner anymore. 

If the initial workout (5mi at 95% 10k pace) is too tough, that’s fine too—just make it 4mi, or break it into intervals.

The point is to start with whatever your fitness currently supports, and work towards improving your capabilities at that speed, i.e. by running it for longer without stopping (as in the 600m → 1000m progression above for a workout totaling 6 km) or running more total volume (as in the 5mi → 8 mi progression above).

The actual workouts you use at 95% and 105% of race pace will of course depend on the event—you’re not going to use continuous runs at 95% of 800m pace, after all—but we’ll get to that later.

3. Adjacent speeds form a “ladder of support”

Now, there’s a catch to this strategy of building support for race pace with specific speed (105% race pace) and specific endurance (95% of race pace). The catch is that your ability to run these speeds also relies on speed and endurance support for each speed in turn.[5]

Let’s take the example above of a continuous run at 95% of 10k pace. Your capabilities in that workout will depend strongly on your endurance at 90% of 10k pace, and your speed at 100% of 10k pace.

In fact, there’s nothing particularly special about race pace—all speeds exist on a ladder of support with adjacent speeds serving as speed and endurance support to one another.

For example, 105% 10k pace from above is endurance support for 110% of 10k pace, and speed support for 100% of 10k pace.

This ladder of support means that there is a cascading effect for speed and endurance support across a wide range of speeds. In practice, good training should encompass speeds ranging at least from 80% to 115% of race pace to sufficiently pad out this ladder of support.

Renato Canova uses the term “special” to refer to 90% and 110% of race pace—these are “special endurance” and “special speed,” respectively, in his nomenclature. As a brief aside, this is the etymology of the famous special block: it is “special” not in its specialness, but in its focus on special speed (110%) and/or special endurance (90%).

I tend to use the term “race-supportive” as it’s a nice complement to “race-specific” but that’s merely a difference in terms.

The 5% rule for race distance conversions

Using five percentage point increments for the ladder of support has an additional benefit, which is that it allows you to easily relate many speeds to common race paces via the 5% rule. That rule says the following:

Decreasing your pace by 5% allows you to sustain that pace for twice as long.

In other words, 10k pace is 95% of 5k pace, half marathon pace (21 km) is 90% of 5k pace or 95% of 10k pace, and so on. 

Now, this is an approximation! Real race prediction calculators will be a bit more precise.

But the 5% rule is still an extremely useful tool getting the effort right: when doing long repeats at 95% of 5k pace, the effort level you should be shooting for is “close to 10k pace.” Likewise, 105% of marathon pace should be “close to HM pace.”

4. Endurance support matters most

There is one way in which physiology creeps into percentage-based training, and that’s in the following observation—endurance support matters more than speed support. We can justify this statement physiologically or empirically.

Physiologically, we know that aerobic energy production is overwhelmingly important in all middle and long distance events. Especially for events from the 800m to the 10k, paces on the endurance side—i.e. fractionally slower than race pace—are the ones that most strongly target aerobic energy production.

On an empirical basis, most coaches know that it’s a far bigger mistake to neglect endurance than to neglect speed.

Let’s say our 37:30 10k runner had a twin sister with the same fitness level. Suppose our original runner puts far more emphasis on 95% and 90% of 10k pace in her training, using fast continuous runs at these speeds; while her sister puts far more focus on 105% and 110%, using shorter intervals at these speeds.

Several months later, the sister who focused on endurance would be in a far better position than the sister who focused on speed.

Of course, we don’t want to totally neglect speed or endurance, but when composing a comprehensive training plan, we should always lean more heavily on endurance at fractionally slower than race pace.

It might help to modify our concept of the “ladder of support” to a new analogy of a “pyramid of support,” though this metaphor is a little unusual in that what we really care about is in the middle of the pyramid (race pace), not at the top or on the bottom.

Volume and extension for endurance versus speed workouts

The pyramid analogy is also helpful because in most cases (the marathon and 800m are a little unusual in this regard) your workout volume and extension follow a pyramid-like distributions: speed support involves shorter reps and lower total volume, while endurance support involves longer reps and higher total volume.

Take our 10k runner from above: She might do 8–10 km of work at 100% of 10k pace, using extension (rep distances) of 1–2 km. At 105%, she might do only 6 km of work, using extension of 600–1000m. At 95%, she might do 12–18 km of volume (in interval sessions) with extension of 2–3 km, or 5–8 mi of volume but without stopping: so 5–8mi of extension also. 

Again, this is another spot where the specifics depend on your experience level and mileage, and there are enough edge cases—like the example earlier of 5mi at 95% 10k pace—where this principle doesn’t rise to the level of a hard and fast rule, but it’s still a good starting place to guide your workout planning.

5. Paces slower than 80% of race pace are general, not specific

The ladder of support does not extend forever. There comes a point at which paces are so slow (or so fast) that they are not directly relevant to performance in the event. Renato Canova maintains that 80% is where this ladder of support ends, and I’ve found this to be a useful floor as well.

That doesn’t mean that speeds slower than 80% of training are not worth doing at all—just that they should not be considered race-relevant speeds. At best, they can be thought of as “training to train,” but at worst, they can be a complete waste of time.

Workouts that are too slow for race-specific benefits

An extreme example: I once knew a long sprints coach who regularly—as in multiple times a week, during the track season—had his (male) 400m runners doing three- to four-mile runs at 8:00-9:00/mi pace. It’s not hard to see why this training was virtually irrelevant for athletes trying to run 400m at 3:20–3:40 mile pace (50–55 sec for 400m) in the near future.[6]

Almost everyone needs to do some “training to train,” though, and given the importance of endurance, it shouldn’t be surprising that a lot of your schedule ends up filling up with “training to train” type workouts. Or maybe even “training to train to train.”

A 15:37 5k runner (5:00/mi) might do easy mileage at 7:30-7:00/mi pace. That’s only 50-60% of race pace! But it’s absolutely a mistake to think you can cut out this easy running and only focus on workouts at 80-115% of race pace. Why? Because you need endurance support for your ability to run 80%, 85%, and 90% of race pace.

More importantly, though, percentage-based training explains why it’s also a mistake to think you can go straight from easy mileage into 5k-specific training and expect to run your best. When doing so, you’ve left a huge hole in your ladder of support; there are no rungs to get you from 50-60% of 5k pace up to 100%.

In Canova’s parlance, these speeds at 80% of race pace and slower are “fundamental training”—not specific to the event, nor sufficient for top performance on their own, but essential nevertheless for supporting your race-specific endurance work.

Workouts that are too fast for race-specific benefits

The same ladder of support framework also explains why some workouts are too fast for race-relevant performance. Take the surprisingly common example of marathoners doing fast 400s during their final few weeks of marathon training.

The percentage-based view suggests that this is a mistake: during this key period of training, a 2:30 marathoner (5:43/mi) shouldn’t be spending much time doing workouts faster than 5:10/mi (110% of race pace).

Outside of strides or hill sprints, which have more of a mechanical or neuromuscular goal, and perhaps the rare maintenance workout that touches on 115% of race pace (4:52/mi), fast 200s and 400s at 3k or mile pace are not relevant for marathon performance under the percentage-based framework.

6. Order of operations: general training, race-supportive training, race-specific training

Traditional training cycles are often presented as a pyramid: build a base of aerobic fitness, then use interval workouts at race pace, then peak with short speedwork—base, pace, race (in that order), working up the pyramid from the bottom.

Percentage-based training works instead more like a funnel: you start at either end of the ladder of support and work inward over time, with speed and endurance support converging on race pace.

General training

The initial base-building for a percentage-based training cycle uses continuous runs and low-key interval work (depending on the event) at around 70-90% of race pace, plus some shorter mechanically-focused work at 115+% of race pace. Because endurance matters most, the endurance work is emphasized more often during this phase of training.

Canova calls this phase the fundamental phase under his system, sometimes using that term interchangeably with “general training.”

Building general fitness can also encompass building up overall strength and athleticism to support running-specific workouts; this takes the form of drills, hill sprints, and circuits or other strength work.

Again, sprint training is a useful lens here: lifting weights is not the most important training that a 400m runner does, and it’s not specific to the event. However, there is some level of general strength that’s required for the 400m specific workouts to come later in the spring, hence the need to do lifting over the winter.

Importantly, and unlike traditional American training, general training does not involve completely avoiding 90–115% of race pace. Instead, these speeds are touched on occasionally, in shorter repeats and moderate-volume workouts, coexisting among the full ensemble of other paces used in training.

Race-supportive (special) training

Race-supportive training—sometimes called “special” training in Canova’s parlance—involves putting increased focus on workouts at 110% of race pace and 90% of race pace. These are race-supportive (or “special”) speed and race-supportive endurance, respectively.

The specific workouts you use will of course depend on the race distance in question, but both the endurance and speed support you build should be stepping stones to later workouts at 95% and 105% of race pace (which in turn are stepping stones to 100% of race pace).

As with general training, the emphasis during this phase of training is not exclusively on 90% and 110% of race pace. Rather these speeds are emphasized to a greater degree.

Race-specific training

Finally, when all the pieces are in place, you can spend several weeks with your major emphasis being race-specific workouts: sessions at 95%, 100%, and 105% of race pace. Some of these sessions might be actual races themselves!

As with the other phases of training, you still want to touch on fundamental and race-supportive workouts—as Canova has said, “training is to add, not to replace.”

Race-specific training really isn’t that complicated—as noted earlier, almost everyone agrees on how to construct race-pace workouts. It’s getting strong enough to do these race-pace workouts that’s tricky. 

The ladder of support across time, in one image

Here's how I conceptualize the three phases of training:

Tap to enlarge

This figure shows the main elements to keep in mind with percentage-based training: the ladder of support, the frequency at which you touch on each speed, and the volume of the workout (in relative terms).

7. Start from your current fitness, not your goal fitness

Training is like a bridge with one end grounded in the current state of your body, and the other end leading—hopefully—to your racing goals. Traversing this bridge requires planning your initial workouts using a realistic estimate of your current fitness, not where you hope to end up.

Because the entire project of percentage-based training is based on increasing your ability to run further at every point along the ladder of support, it clearly doesn’t make any sense to start up with workouts that are already above your current capabilities.

You will eventually need to update your fitness estimates, but we’ll cover that in a moment.

8. Progress your workouts over time

 Traditional American training is often quite monotonous—every week has 6 x 1k at threshold and a 13 mile long easy run, for weeks or months on end.

Percentage-based training, in contrast, virtually never uses the same workout twice.

Every workout has a purpose, and every workout must evolve: either by adding more total volume, increasing the extension, or by reducing the recovery between repeats. Each of these is a lever you can pull to evolve a workout.

Increasing workout volume

This is the most straightforward method to increase the stimulus of a workout: add more reps (for intervals), or increase the distance (for a continuous run).

The main downside of increasing workout volume is that the biomechanical stress on your body goes up—8mi at 95% 10k pace imposes double the mechanical stress as 4mi at 95% 10k pace.

Increasing extension

Extension—the distance of each rep in an interval workout—is a vastly underused tool for evolving workouts. Manipulating extension is an excellent option because it increases the physiological stimulus of a workout without affecting the mechanical stress on your body.

As an example, let’s say a 16:40 5k runner is doing 8 x 1 km at 5:36/mi (95% 5k) with 1.5 min jog recovery as a specific endurance workout. She could, over time, progress to 2k-1k-2k-1k-2k, then a few weeks later, 4 x 2k, then later, 3k-2k-2k-1k.

All of these workouts total 8 km at 95% of 5k pace, but this progression represents a significant boost in the training stimulus with no increase in injury risk.[7]

Notice one trick with extension used in the progression above: you don't necessarily have to extend all of the repeats. Even 3-2-2-1k is a new and different stimulus than 2-2-2-2k.

Decreasing recovery

A third way to progress the same workout is to take less recovery. Working with the same 5k example, a midpoint between 8 x 1k and 4 x 2k might be 4 sets of 2 x 1k, with 45 seconds rest within sets and 3 min between sets.

In the limiting case, decreasing recovery ends up being synonymous with increasing extension: once our 5k runner progresses to 4 x 2k, she’s essentially just decreased the recovery to zero between half the repeats in her workout. 

Decreasing recovery has the same mechanically-favorable properties as increasing extension: you can make the workout more demanding without an increase in biomechanical loading.

Workout progression is necessary even when your fitness is improving

 Remember my critique above about runners who do 6 x 1k at threshold for weeks at a time?

You might have protested—“Ah, the workout is the same, but I’m getting more fit! Three weeks ago I did 6 x 1k at 3:30; now I’m doing it at 3:25 with the same effort!

And it’s true, you did get more fit. But there’s a problem: even with the updated paces, the workout is not a new stimulus in terms of the internal load on your body!

Threshold is threshold, whether that’s 3:30 or 3:25/km.

Even as your fitness grows, your workouts need to evolve in one of the three directions above: more volume, longer reps, or less recovery.[8] Otherwise, your body will not be experiencing a new stimulus.

The converse is actually not true: if you’re adding volume, increasing extension, or decreasing recovery, it’s totally fine if your pace stays the same: 3:30/km for 8 x 1k is still a definite improvement from 3:30/km for 6 x 1k.

9. Pair greater stress with greater recovery

One important counterpart to progressing your workouts over time is to pair greater stress—longer, faster workouts—with an increased number of easy to moderate days to recover in between workouts.

Recovery can progress within a season and across your career.

Within a season, a 5k runner might do 8mi at 80% 5k pace early on during the winter, which might only require one easy day afterwards. By early spring, though, they might be doing 15mi at the same pace, which could require three to four days of easy running and light fartlek before doing a real workout again.

Across your career, younger and less-experienced runners can use more frequent, lower-volume workouts, often doing three workouts and a long easy run in a week and distributing the intensity fairly uniformly throughout the week—some of your non-workout days can be runs at more of a moderate to steady effort.

As you become more experienced, though, you’ll move to less-frequent, higher-volume workouts, with a greater number of strictly easy days between high-quality workouts.

Renato Canova calls this strategy increasing modulation, and remarks that after the highest-quality workouts his top Kenyans will often take four or five days of easy recovery before they work out again.

When I see people who have “burned out” trying to do Canova-style workouts, it’s almost always because they did not use this strategy of increased modulation, and started doing long, fast workouts with far too little recovery afterwards.

10. Update your workout paces as you improve

Over time, you’ll be able to run a faster pace for the same effort level. That’s what it means to improve.

There are actually three pretty different ways to progress your paces under a percentage-based program as you improve. I call these methods “fitness updates,” “workout evolution,” and “bridging to goal pace.”

Fitness updates

This is the simplest method to use and it’s the one I recommend for almost everyone.

As you improve your individual workouts, you can update your estimate of your current fitness level.

If you race often, this is trivial: just use your most recent race result as your fitness estimate.

But what if you aren't racing much? In this case, you can infer your fitness from your workouts.

Let’s return to our 10k runner from earlier—suppose she incorporates a few of our recommended workouts, and soon is able to do a 10k-specific workout (say 5 x 2 km) at 5:56/mi at the same effort level that used to yield 6:00/mi. Now she can update her fitness estimate to 5:56/mi and use it for her new workouts. 

Fitness updates do have a few tricky aspects, though.

First, you don’t always improve equally across all workout types. If our 10k runner is a true distance specialist, she might find that she can absolutely crush the fast continuous runs at 95% 10k pace, but struggles with repeats at 105% 10k pace.

In this case, you might need to update your fitness estimate for your 95% 10k pace workouts, but not for your 105% 10k pace workouts, leading to our runner (or her coach) having a bit messier of a time planning workouts and strategizing for races.

Another potential source of trouble is the fact that you have to be honest with yourself about your effort level.

Some runners seem to be miscalibrated in this regard, almost always in the direction of being overeager about their fitness based on their workouts. For shorter distances, this is an easy fix—you just get out and actually race—but it can be tougher for the 10k, HM, and marathon. 

To emphasize again, until you really know what you're doing, this is the way you should be updating your workout paces! However, there are two more techniques for more advanced runners and coaches.

Workout evolution

Workout evolution is a more sophisticated and long-term oriented approach to progression in workouts.

Under this approach, you plan out your initial “rungs” on the ladder of support using your current fitness, but then you completely let go of the idea that these paces are “attached” to race pace thereafter.

Sound confusing? Here’s how it works. Our 37:30 10k runner would use her initial fitness estimate of 6:00/mi to set up her ladder of support, with five percentage point increments from 80% to 115% of race pace.

Then she would evaluate what level of workout (in terms of volume, extension, and recovery) she is currently able to do at each rung on the ladder of support. Here’s how that might look: 

Current workouts

PercentPaceCurrent workout
115% 10k5:06/miStrides
110% 10k5:24/mi16 x 200m
105% 10k5:42/mi10 x 600m
100% 10k6:00/mi6 x mile
95% 10k6:18/mi5mi continuous
90% 10k6:36/miNone
85% 10k6:54/mi4 mi progressive at the end of a 13 mi long run
80% 10k7:12/mi4 mi progressive at the end of a 13 mi long run

Looking at the ladder of support this way immediately reveals some avenues for improvement—add a long moderate run with a greater proportion at 80-85%, add a long fast run at 90%, increase the length of the 5mi fast continuous run at 95%, add some more substantive mechanical work at 115%, and so on. 

Here’s the key to the workout evolution approach: the workouts themselves evolve, but the paces do not.

It’s a bit unintuitive, so let’s zoom in on the 5mi fast continuous run at 6:18/mi.

Over several months, our runner might be able to progress this workout to 8mi at 6:18/mi. Surely she’s in better than 37:30 10k pace, but under the workout evolution approach, we don’t change the paces—the workout itself evolves into a different type of training session.

You’re not even focusing on “how long can I run at 95% 10k pace”; you’re merely trying to improve how long you can run 6:18/mi, whatever percent that corresponds to.

Sometimes, with dramatic fitness improvements, workouts can evolve almost in the Darwinian sense, from one “species” of workout to another. Say our runner also progresses her 6 x mile session, moving to 5 x 2k, then 3 x 2mi, then 4mi continuous, then 5mi continuous—all at 6:00/mi pace.

This would take several months, of course, but by the time she’s doing 5mi at 6:00 pace as a workout, that workout has morphed into a completely different session in terms of the internal load on her body. Now, it’s probably more like 95% of 10k pace!

Workout evolution lets your ability at each pace increment on the ladder of support grow on its own. It also lets you more carefully monitor your weak points, shifting the frequency with which you do different workouts to fine-tune your fitness.

If, for example, our 10k runner noticed that she was stagnating or struggling at 6:36/mi, she could redirect more training time to that pace and those adjacent to it (6:54 and 6:18/mi).

I’ve found that the the workout evolution approach works particularly well in three scenarios:

1. Long time horizons

When you’re planning out a five- or six-month training cycle, hopefully paces at the beginning are very different from paces at the end. For instance, a marathoner’s lactate threshold in June might easily be their marathon pace by December.

2. Highly imbalanced runners

I don’t mean “imbalanced” in the biomechanical sense, but in the fast-vs-slow-twitch sense. You might have a 2:20 800m runner whose endurance is abysmal at 84 sec/400m (80%) but whose speed is incredible at 63 sec/400m (110%).

Using workout evolution allows you to track her progress at these speeds totally independent from one another. A static fitness estimate, in contrast, would make her 80% workouts too hard, and her 110% workouts too easy.

The inverse situation comes up with ultramarathoners: when doing marathon training, their endurance at 80-85% MP is often essentially infinite, but their abilities at 115% MP are lamentable. In this situation, you have to build up the speed side of marathon pace much more patiently and slowly, not letting nominal marathon pace dictate these workouts. 

3. Runners who are rapidly improving

If you’re returning to running after a long break, or if you happen to be an incredible talent, your fitness can be on such a steep upward trajectory that it’s very hard to accurately gauge your current fitness from week to week.

The workout evolution approach frees you from having to worry about fitness estimates at all: you just see what progress you're able to make at each pace increment.

For these runners, I often recommend a range of workout volume, e.g. 4–6mi at 95% 10k pace, allowing them to decide how far to go (perhaps in 0.5mi increments) based on how they feel.

Bridging to goal pace

The most advanced method of progressing workout paces, and the most difficult to use with any consistent success, involves comparing your current fitness against your goal race pace, calculating how long you have until your goal race, and then linearly updating your workouts in between.

I call this “bridging to goal pace” because your workouts serve as a “bridge” to get you from your current fitness to your goal fitness, and you specify exactly what time you want to run.

This is what top coaches often do, but I have very rarely used this approach—it’s very tough to know a priori what an athlete is capable of, and it's also difficult to tell what rate of improvement is realistic unless you deeply know an athlete's abilities. Bridging to goal pace also requires extremely precise pacing in workouts.

Nevertheless, this approach could make sense if you have a definite time you need to run—especially if you’re pretty sure that time is within your abilities. Obvious examples include Boston Qualifiers, London Good-for-Age times, elite field entry standards, and conference meet qualifying cutoffs for track athletes. 

Conclusion: percentage-based training in one sentence

Here’s the best one-sentence summary of percentage-based training that I can manage:

Use your workouts to gradually and progressively improve your ability to run at speeds fractionally faster and fractionally slower than race pace.

That’s really all there is to it.

If you want a tiny bit more detail, these are the ten principles of percentage-based training from above:

  1. Percentage-based training is primarily mathematical, not physiological
  2. The main problem to solve is supporting your race-specific workouts
  3. Adjacent speeds form a “ladder of support”
  4. Endurance support matters most
  5. Paces slower than 80% of race pace are general, not specific
  6. Order of operations: general training, race-supportive training, race-specific training
  7. Start from your current fitness, not your goal fitness
  8. Progress your workouts over time
  9. Pair greater stress with greater recovery
  10. Update your workout paces as you improve

And see also the ladder of support in one image, above.

Of course, the implementation of these principles into specific training schedules makes things more complicated—mileage, experience level, racing opportunities, and so on will all influence how you translate these guiding principles into an actual training calendar. I’ll write more on that in the future.

Nevertheless, I’ve found that having a bedrock foundation of principles to return back to is incredibly useful for dealing with novel situations in training.

Support my work

If you enjoyed this article, sign up for my free newsletter below! You’ll be the first to know when I release more articles like this one in the future.

You should also check out my book on Amazon.com - it includes info on a simplified version of percentage-based workouts for 800m to 10k training where you do race-pace sessions at events longer and shorter than your primary race distance.

In the future, I’ll write up full-fledged guides to percentage-based training for each track and road race distance so you can see how the principles from above actually shake out in real training schedule. For now, though, I’ll leave you with some examples of percentage-based workouts for a range of different events that address general, race-supportive, and race-specific fitness. 

Remember, you can use my free web app to calculate your paces for these workouts!

A few notes on the recommended workouts

The following sessions are “my” workouts, in that I don’t claim that they are endorsed by anyone but me.

In many cases I didn’t actually come up with the workout de novo (and no workout ever really exists in a vacuum), but this is a list of "John Davis-endorsed workouts," not “Renato Canova-endorsed workouts,” or Claudio Berardelli-endorsed, Gabrielle Rosa-endorsed, or anyone else who uses percentage-based training, for that matter.

These examples are mostly “middle of the road” workouts suitable for top high schoolers, most college runners, and post-collegiates doing mileage that’s reasonable for their event.

I did my best to keep these workouts straightforward. I love sophisticated workouts as much as anybody else, but it’s very easy to galaxy-brain yourself into something overly complicated. When working with a new training system, keep things simple at first.

Remember: these are just examples! You’ll of course need to adjust workout volume and difficulty for your fitness, mileage, and experience level. If any of them seem too aggressive, or too conservative, that’s fine—adjust them to fit your needs.

Experience and mileage level affect workout volume quite a bit. I’d have no problem with a 27-year-old 14:30 5k runner working up to 12mi at 85% 5k pace if he was doing 90–100 mi/wk; I’d never prescribe that session to a high schooler, even if he was also in 14:30 shape (and even if he was running similar mileage!). 

All percentages below are percent of current race pace for the event in question. Some instances have multiple different example workouts. Please double-check for sanity—it’s easy to overlook typos with so many workouts listed!

Percentage-based workouts for the 800m

80%Basic endurance4–6 x 500m at 80% w/ 3–4 min walk/jog
85%General endurance8–10 x 300m at 85% w/ 1.5 min walk
90%Race-supportive endurance4–5 x 500m at 90% w/ 4–5 min walk/jog
95%Race-specific endurance3–4 x 600m at 95% w/ 4–5 min walk/jog
100%Race pace3 x 500m at 100% w/ 6–8 min walk/jog
105%Race-specific speed6 x 300m at 105% w/ 2–3 min walk/jog
110%Race-supportive speed10–12 x 200m at 110% w/ 2–3 min walk/jog
115%General speed6–10 x 60m at 115% w/ 3–4 min walk

Notes on recommended workouts for 800m runners:

Good 800m training is a lot more variable from one runner to the next depending on the degree to which they are a “fast twitch” runner with good 400m speed versus a “slow twitch” runner with good 1500m endurance.

This is the set of workouts I feel least strongly about—It’s been about six years since I’ve worked with 800m runners on a regular basis, so I’ve sent the above workouts to a coaching friend to get feedback on them. I’ll likely update the recommended workouts after I hear back.

Because “800m pace” often involves a ~5% range (e.g. running 57.0 - 60.0 splits for a 1:57 800m), you can often mix 100% and 105% in workouts, e.g. with the 3 x 500m session—that could easily start at 100% and progress to 105% by the final rep.

For a lot of 800m runners, 115% is close to top speed, so you might categorize that speed as pure sprint work.

Percentage-based workouts for the 1500m / 1600m / mile

80%Basic endurance6–8 x 1 km at 80%, 1–2 min jog; or 3–5 mi at 80%
85%General endurance5–7 x 1 km at 85%, 1.5–3 min walk/jog
90%Race-supportive endurance5–6 x 800m at 90% w/ 2–3 min walk/jog
95%Race-specific endurance6–8 x 600m at 95% w/ 2–3 min walk/jog
100%Race pace5 x 600m at 100%, 3 min walk/jog
105%Race-specific speed8–10 x 300m at 105% w/ 2 min walk
110%Race-supportive speed12 x 200m at 110% w/ 2 min walk
115%General speed8–10 x 120m at 115% w/ 3–5 min walk

Notes on recommended workouts for milers:

Depending on your aerobic fitness, 80% of mile pace can wind up pretty close to lactate threshold, so you can use your favorite Daniels-style “T” workout here.  Do keep in mind that speed-oriented 800/1500 types will not have the same LT as an endurance-oriented miler, and these speed-oriented runners will require more recovery and shorter rep distances after endurance-focused sessions.

Again, depending on endurance, 85% can be pretty close to critical speed (a.k.a. “critical velocity” or “CV”, roughly 8k-10k pace) so you can use Tinman-style CV workouts for this pace if that’s your preference. Same caveats about speed vs endurance milers apply.

Percentage-based workouts for the 3000m / 3200m / two mile

80%Basic endurance4–8 mi at 80%; or 3–4 x sets of (1.5 mi at 80%, 0.5 mi at 60-65%)
85%General endurance6–8 x 1 km at 85% w/ 1 min jog; or
3.5–6 mi at 85%
90%Race-supportive endurance4–6 km at 90%; or 3–4 x mile at 90% w/ 3 min jog
95%Race-specific endurance8 x 800m at 95% w/ 2 min jog
100%Race pace8 x 600m at 100% w/ 2–3 min jog
105%Race-specific speed6-7 x 500m at 105% w/ 2–3 min walk/jog
110%Race-supportive speed16 x 200m at 110% w/ 2 min jog
115%General speed10–12 x 150m at 115% w/ 2 min walk

Notes on recommended workouts for 3k / 2mi runners:

80% of 3k pace will be a bit below the lactate threshold for most runners, so it’s suited either for fast continuous runs or long repeats with recovery at more of moderate, uptempo pace, like you might do in half marathon training.

85% can end up close to lactate threshold, so as above, feel free to use Daniels-style workouts to hit this pace.  Ditto for 90% and critical speed (“CV”). 

Percentage-based workouts for the 5k

80%Basic endurance8–15 mi at 80%
85%General endurance5–9 mi at 85%
90%Race-supportive endurance4–7 mi at 90%
95%Race-specific endurance4 x 2 km at 95% w/ 3 min jog; or
4–6 km at 95%
100%Race pace5-6 x 1200m at 100%, 3 min jog
105%Race-specific speed2 sets of (5–6 x 500m at 105%) w/ 45 sec & 4–5min walk/jog within/between sets
110%Race-supportive speed16 x 200m at 110% w/ 1–2 min jog
115%General speed10–12 x 200m at 115% w/ 2–3min walk/jog

Notes on recommended workouts for 5k runners:

5k training will look pretty similar to traditional workouts you might be familiar with already. 85% of 5k pace is close to “aerobic threshold” or Daniels’ “M pace”; 90% is pretty close to half marathon pace; and 95% will be 8k–10k pace / critical speed / “CV.” 105% might be “goal 3k pace” or thereabouts, depending on how speed-oriented you are.

If you’re looking for new 5k sessions to add to your repertoire, I recommend adding fast continuous runs at 95% and long fast runs at 80%. These are underused elements in most 5k runners’ training and can make a big difference in fitness. If you don’t already do long repeats at 95% (2k to 3k in length), add these too–they are a huge value-add.

Percentage-based workouts for the 10k

80%Basic endurance12–18 mi mi at 80%
85%General endurance10–18 mi at 85%
90%Race-supportive endurance7–10 mi at 90%
95%Race-specific endurance4–7 mi at 95%; or
5 sets of: (3 km at 95%, 1km at 85%)
100%Race pace5 x 2 km at 100%
105%Race-specific speed6 x 1000m at 105% w/ 3–5 min jog
110%Race-supportive speed8–10 x 500m at 110% w/ 1.5–2 min jog
115%General speed16 x 200m at 115% w/ 1.5–2min walk/jog

Notes on recommended workouts for 10k runners:

As above, the 5% rule applies here—105% is very close to 5k pace, 95% is very close to HM pace, and 90% is close to aerobic threshold / Daniels’ “M” pace. 

Top NCAA XC programs often use long tempo runs or “aerobic threshold runs” at close to 90% of 10k pace; if you don’t already do these, they should be at the top of your list of workouts to add.

Remember that you can include races to hit some of these paces–e.g. a 5k to hit 105% 10k pace, or a half marathon to hit 95%. 

Percentage-based workouts for the half marathon

80%Basic endurance10–18 mi at 80–85%
85%General endurance10–18 mi at 80–85%
90%Race-supportive endurance10–18 mi at 90%
95%Race-specific endurance10–15 mi at 95%
100%Race pace5–6 sets of: (3 km at 100%, 1 km at 90%)
105%Race-specific speed5 x 2 km at 105% w/ 3 min jog; or
5–6 km at 105%
110%Race-supportive speed8 x 800m at 110% w/ 2 min jog
115%General speed12 x 300m at 115% w/ 100m jog

Notes on recommended workouts for half marathoners:

The distinction between 80% and 85% gets a little fuzzy for half marathoners, so sometimes it makes more sense to do them together as a progressive long run (or long-ish run of more like 10 mi, early in the training cycle).

Usually the right move for progressing your long run over time is to start with runs at 80-90% as a “base,” then move to long fast runs at 95% during the final several weeks of half marathon training. If you touch on 80-90% in a shorter mid-week session once in a while (e.g. 5mi at 80-85%) you don’t need to continue doing long runs at 80-90% once you move to long fast runs at 95%.

The “5% rule” applies here; 105% HMP is very close to 10k pace, and 115% HMP is close to 5k pace, so you can use standard 5k/10k workouts for these speeds. 

Percentage-based workouts for the marathon

80%Basic endurance15-22 mi at 80-85%
85%General endurance15-22 mi at 80-85%
90%Race-supportive endurance15–22 mi at 90%
95%Race-specific endurance15–20 mi at 95%
100%Race pace5 sets of: (4 km at 100%, 1 km at 90% )
105%Race-specific speed8–10 sets of: (1 km at 105%, 1 km at 90%); or
4–5.5 mi at 105%
110%Race-supportive speed3 x 3 km at 110% w/ 3 min jog
115%General speed8 x 800m at 115% w/ 1.5–2 min jog

Notes on recommended workouts for marathoners:

As with the half marathon, the 80-85% range is a little fuzzy since it gets close to some runners’ easy run pace. In practice I actually don’t ever prescribe workouts at these paces; I train marathoners almost like 5k runners for the base phase, then switch to long fast runs at 90% of marathon pace (which is no problem, since they’ve been doing shorter long fast runs at 80% of 5k pace already, plus long easy runs).

Some approximate 5% rule conversions: 105% MP is about half marathon pace; 110% MP is about 8k–10k pace. In theory 115% MP should be about 5k pace, but many marathoners have trouble actually running that fast for a 5k when they’re in marathon shape.

If you add one pace to your workout rotation, it should be long fast runs at 90–95% MP. These are enormously helpful workouts, yet I almost never see non-professional marathoners doing this kind of workout.


Click the footnote number to return to the main text

[1]  I’m using “mile” here to mean 1600m; it makes the math easier.

[2] For reference, 5:42/mi is pretty close to 5k pace (= 17:49 for 5k) for a 37:30 10k runner. 6:18/mi is pretty close to half marathon pace (= 1:20:04 for the HM) and is a bit slower than what most calculators would predict as lactate threshold pace. This is the happenstance similarity that makes percentage-based 5k/10k training look familiar to American runners. 105% and 95% of 800m pace, though, don’t correspond to any familiar physiological or race paces. See also the note on the 5% rule.

[3] The specific workout progression might be 10 x 600m at 105% 10k pace with 1.5min jog recovery; then 10-15 days later, 8 x 800m at 105% 10k pace with 2 min jog recovery; then another 10-15 days later, 6 x 1000m at 105% 10k pace with 2-3 min jog recovery.

[4] You can also do intervals, e.g. as part of a 13 mi long run every other week. An example progression that’s lower in volume and intensity: 7mi easy + 3 sets of (1.5mi at 95% 10k pace, 0.5mi moderate run); then two weeks later, 5mi easy + 4 sets of (1.5mi at 95% 10k pace, 0.5mi moderate run); then in another two weeks, 3mi easy + 4 sets of (2mi at 95% 10k pace, 0.5mi moderate run).

[5]  I’m using “speed” a bit loosely here, and throughout this article. Really I mean pace: time per unit distance. Percentage-based training uses percentages of pace, not percentages of speed. I have an entire article about the distinction here, and there are notes on percent of pace vs. percent of speed on my pace app here.

[6] Is it always a mistake for a 400m runner to do slower easy running? Not always—Clyde Hart would have his world-class 400m runners at Baylor do two miles hard over cross country as a fall-season workout. And to get ready for that workout, you might do some 20–45 minute easy runs. But even this two mile cross country run is not directly relevant to 400m performance: it’s several layers deep in the “training to train” approach. Two miles hard over cross country terrain in the fall supports 800m repeats early in the winter, which support 600m repeats later in the winter, which support…and so on up the ladder of support.

[7] Unless you believe that fatigue changes gait mechanics in a way that increases injury risk; I’m skeptical of this argument in its most general form.

[8]  Unless you’re merely doing a workout to maintain fitness, versus improve it, which is fine as long as you’re honest with yourself about your workout goals!

Related articles

Getting the interval workout recovery right

I’ve previously written about getting the warm-up right and getting the cool-down right. Getting these components of training right is a relatively minor component of the design of an overall training program. Still, if you’re planning on having a long running career (or if you’re a coach working with a team), you’re going to be ... Read more
Digital art image of a runner with math equations in the background

Problems with the critical speed model: Can power laws predict running performance better?

The critical speed model—also known as critical velocity, CV, or critical power—is a powerful concept for understanding what running speeds are sustainable at a metabolic steady-state and what speeds are not. Critical speed is not without its detractors, though, and the critical speed model is certainly not without its flaws. I just posted a huge ... Read more
Female athlete running in front of scientific graphic paper

The science of critical speed, critical velocity (CV), and critical power training for runners

Critical speed is the boundary that separates running speeds that can be sustained at a metabolic steady-state from speeds that cannot. Sometimes called critical velocity or “CV,” critical speed is known in the running world in partly due to its popularization by Tom “Tinman” Schwartz and his proteges, including Drew Hunter.[1] Critical speed is increasingly ... Read more

The Keys to Marathon Training: Modern changes to Renato Canova’s elite marathon training methods

While researching my blog post on Renato Canova’s marathon training book, I came across a lecture that Canova gave at a coaching conference put on by Spanish marathoner and coach Antonio Serrano in 2017. The talk, called The Keys to Marathon Training[1] was held in conjunction with the 2017 Valencia Marathon. This lecture directly answers ... Read more
Photo of the book Marathon Training - A Scientific Approach by Renato Canova and Enrico Arcelli

Review and summary of Marathon Training - A Scientific Approach by Renato Canova

How do the best marathon runners in the world train? While you might catch a workout or two on Instagram or hear rumors about epic training weeks on message boards, there’s precious little information on the systematic approaches that elite coaches use with top marathon runners–and even less information on the science that backs up ... Read more
Stylized image of treadmills in a gym

A scientific guide to treadmill training and workouts for runners

Though many runners treat treadmills as a necessary evil, treadmill access is a must-have if you live anywhere that gets extreme cold or extreme heat and want to train seriously year-round. Being from Minnesota (and having coached many runners in the Midwest), I’ve had plenty of experience modifying workouts and training sessions for the treadmill.  ... Read more

About the Author

John J Davis, PhD

I have been coaching runners and writing about training and injuries for over ten years. I've helped total novices, NXN-qualifying high schoolers, elite-field competitors at major marathons, and runners everywhere in between. I have a Ph.D. in Human Performance, and I do scientific research focused on the biomechanics of overuse injuries in runners. I published my first book, Modern Training and Physiology for Middle and Long-Distance Runners, in 2013.

Leave a Comment

Did you know I have a book? Check it out here!