Why the paces are separated at all
Running everything at a moderately hard effort lands in a gap: too demanding to recover from properly, not demanding enough to force adaptation. Sessions accumulate fatigue without producing fitness, and the block ends where it started. The standard answer, and the one nearly every coaching system converges on, is polarisation — run most of the volume genuinely easily and a small fraction genuinely hard.
The two ends do different jobs. Easy volume builds capillary density and mitochondrial content, and lets you accumulate mileage without breaking. Hard work stresses cardiac output and running economy. Blending them into a middle intensity gets neither.
How each pace is derived
The threshold pace is anchored to the distance you could race in exactly one hour, back-calculated through the Riegel model from your entered result. That is the conventional operational definition of the lactate threshold pace: the fastest you can hold for around sixty minutes. Marathon pace is the Riegel prediction for 42.195 km. Interval pace uses the 3K prediction and repetition pace the 1500 m prediction, which is how the Daniels system positions them.
Easy pace is the exception and is not derived from anything. It is threshold pace plus 60 to 90 seconds per kilometre, about 1:37 to 2:25 per mile, which is a rule of thumb with a long track record and no formula behind it. If your easy runs currently sit faster than the top of that band, the band is more likely to be right than your habit.
The Riegel model and where it stops working
Peter Riegel's 1977 formula states that t2 = t1 × (d2 / d1)^1.06. The exponent encodes how much pace decays as distance grows: at 1.06, doubling the distance multiplies the time by 2^1.06, which is 2.085, so pace slows by about 4% per doubling.
A 10K in 50:00 predicts a marathon of 3:50:01 by this model. That prediction spans a factor of 4.2 in distance and should be read as an upper bound on what is possible rather than a target, because the marathon is limited by things a 10K does not test: weekly volume, long-run history, glycogen availability and pacing discipline over three-plus hours. Predictions between adjacent distances — 10K from 5K, half from 10K — hold up considerably better.
The exponent is adjustable for that reason. A high-mileage runner whose long-distance results consistently beat the model can lower it to 1.04 or 1.05; a runner with track speed who falls apart past 10K can raise it to 1.07 or higher. Fit the exponent to two of your own results and it becomes a description of you rather than of the population.
Running the easy days easy
The single most common failure in amateur training is that easy runs are not easy. The calculated easy pace often looks embarrassingly slow, particularly to someone training alone with no external check on effort, and the correction people make is to speed it up until it feels like running properly. That converts the recovery volume into moderate-intensity work and removes the recovery the hard sessions depend on.
Two checks are more reliable than the clock. You should be able to speak in complete sentences. And your heart rate should sit in the lower zones — the heart rate zone calculator gives the numbers for that cross-check. If pace and heart rate disagree, on a hot day or a hilly route, follow the heart rate.
How much fast running the week takes
Threshold, interval and repetition work together conventionally stay under about 20% of weekly mileage, and under 10% for runners newer to structured training. The frequency ceiling matters as much as the volume: two hard sessions in a week is the usual maximum, three only for experienced runners on high mileage, and never on consecutive days.
Weekly mileage itself is the other constraint. Increases beyond about 10% a week are where injury rates climb, and adding fast work and mileage in the same week compounds two stresses at once. Build volume first, then add intensity to it.
What the model does not know
It has one input. It does not know that your reference race was run into a headwind, that you were three weeks into a cold, that the course had 800 feet of climbing, or that you have not run more than six miles in a year. Each of those changes the correct answer and none of them are visible to the arithmetic. Recalculate from a fresh result every six to eight weeks rather than treating one set of paces as fixed for a season.
It also has no view on whether you should be training at all today. Pain that changes your gait, an Achilles that is sore on the first steps out of bed, or a shin that hurts through a warm-up rather than easing during it, all mean the session comes off the schedule regardless of what these tables say. Nothing on this page is capable of noticing an injury.
Questions people ask
What if my most recent race was months ago?
The paces will describe the runner you were then. If training has continued and gone well they will be conservative, and if you have been out or unwell they will be too fast — the more dangerous direction. Either run a hard time trial over 5K or 10K to refresh the input, or use a recent well-executed hard session as an approximation. Recalculating every six to eight weeks during a block keeps the paces attached to reality rather than to history.
Why does the marathon prediction look faster than I could actually run?
Because the Riegel model assumes only that pace decays with distance at a constant rate, and the marathon breaks that assumption. Beyond about two hours, glycogen depletion, fuelling, thermoregulation and the cumulative mechanical cost of the distance become limiting in ways a 10K never tests. Runners with high weekly volume and a history of long runs come close to the prediction; runners without it commonly finish ten or more minutes behind. Treat it as a ceiling, and set a target from your own long-run evidence.
What exactly is threshold pace?
Operationally, the fastest pace you could hold for about an hour, which is why this calculator derives it from the distance you would cover in exactly sixty minutes. Physiologically it sits near the intensity at which lactate production and clearance stay in balance, and above which lactate accumulates and the effort becomes unsustainable. For most trained runners it lands somewhere between 10K and half marathon race pace. It is the intensity that responds most readily to training, which is why threshold sessions appear in nearly every plan.
My easy pace looks absurdly slow. Is that right?
Almost certainly, and it is the finding most runners resist. Easy running exists to accumulate aerobic volume with minimal recovery cost, and it does that job at an intensity that feels unremarkable. Speeding it up converts recovery volume into moderate-intensity work and takes the recovery away from the sessions that need it. Check it against conversation — full sentences, not gasped phrases — and against heart rate rather than against how a run ought to feel.
Should I change the Riegel exponent?
Only if you have evidence. The default of 1.06 fits the population reasonably well. If you have two honest race results at different distances, solve for the exponent that connects them and use that — it describes your own endurance profile rather than the average one. Runners with high mileage and strong long-distance results often sit at 1.04 to 1.05; those with track speed and poor endurance sit at 1.07 or above. Changing it on a hunch just moves every derived pace without improving anything.