Give any two of pace, time and distance and this works out the third, in minutes per kilometre and per mile. It also predicts what the same fitness should produce over the standard race distances, and prints a split table you can read off during the race.
Riegel's formula, T2 = T1 x (D2/D1)^1.06, scales a known time to another distance. The exponent just above 1 is the whole idea: pace degrades as distance grows, so doubling the distance takes slightly more than double the time. It is accurate within a few percent from 5 km to the half marathon, and reliably optimistic for the marathon unless you have trained the distance specifically - the wall is a physiological event, not a mathematical one.
Pace is time per unit distance and speed is distance per unit time - reciprocals of each other, but pace is what runners think in because it is what a watch shows and what you hold during a race. The split table is even pacing. Most personal bests over 10 km and above are actually run with a slightly negative split, the second half a little faster than the first.
Most training should be conversational and notably slower than race pace; the commonly cited distribution is around 80% easy and 20% hard. Running everything at a moderate effort is the most reliable way to plateau, because it is too hard to recover from and too easy to drive adaptation.
Riegel's formula: T2 = T1 x (D2/D1)^1.06. The exponent slightly above 1 encodes that pace inevitably degrades over longer distances. It is accurate within a few percent between 5 km and the half marathon, and gets optimistic for the marathon unless you have specifically trained the distance.
Most of it should be comfortably conversational, considerably slower than race pace - the widely cited split is around 80% easy, 20% hard. Running every session at a moderate effort is the most common way to plateau, because it is too hard to recover from and too easy to drive adaptation.
Steuervergleich Kanton Zürich — Gemeindesteuern vergleichen (Deutsch)