Norwegian Sub-Threshold

The accounting changes: 20–25% of TOTAL weekly training time across all three sports, not per sport. Then cap each discipline at about 25% of its own weekly time. A discipline carrying none of it is fine — concentration is the point — and what breaks the arithmetic is a discipline carrying too much: two sub-threshold runs in a 3 h running week is a third of your running at quality.

The reference week in this entry totals 10 h 40 min. The three sessions are 131 min of sub-threshold work: 20.5% of the week, 24% of the running, 25% of the riding, none of the swimming. That the whole-week figure sits at the bottom of the 20–25% band is the swim's doing, and it is the right answer — do not add bike reps to lift the percentage.

The faithful version keeps all three sessions in one discipline. Spreading them one-per-sport is a departure, and worth naming as one: a single sub-threshold run a week is an ordinary threshold run, not this method. The practical compromise for most 70.3 athletes is two sessions on the bike and one on the run, because the bike carries quality at near-zero mechanical cost and the run cannot.

Per-discipline anchors. RUN: 90–98% LTHR, 90–92% HRmax as a ceiling, 1–3 km reps on 60 s. BIKE: 85–92% FTP in 10–15 min reps on 90–120 s — note that FTP approximates LT2, so 95% FTP is essentially at threshold and defeats the purpose. SWIM: at or just under CSS in 100s and 200s on 15–30 s. CSS overestimates maximal lactate steady state (Dekerle 2005: CSS at 92.7% of maximal aerobic speed vs steady-state at 88.3%), so CSS pace is not automatically sub-threshold.

The specific trap: the bike will happily absorb sub-threshold work, so triathletes routinely end up with five or six sub-threshold sessions across three sports and no genuinely easy day left. Cardiovascular and metabolic load stack even when mechanical load does not. And a triathlete's long ride is not an easy day in the sense this method requires — count its true intensity honestly.

If you use a lactate meter, re-derive the numbers per discipline rather than porting one target across sports. Beneke measured maximal lactate steady state in the same six athletes at 2.7 mmol rowing and 4.5 mmol cycling — a 1.8 mmol within-person gap, wider than the whole morning-to-evening spread the elite method turns on.

DOUBLE-THRESHOLD ESCALATION, 12 h/week and above only. One cross-discipline double day per week, never two: AM longer bike reps, PM shorter run reps, 6–8 hours apart (4–10 tolerable). Both sessions at Bakken's own band — 2–3 mmol, roughly 80–87% HRmax — with the PM at the top of it. The 2.5 AM / 3.5 PM figures widely quoted come from one illustrative table in Casado's paper, not from Bakken's own prescription. Never two same-day threshold runs for a triathlete running under 60–70 km/week. And the budget is a whole-athlete budget: adding a bike double day on top of an existing run threshold day gives you three threshold days, not one.

SUPPORTED. That controlled sub-threshold interval work accumulates large amounts of near-LT2 time at low fatigue cost, and that this is broadly what elite distance runners do (Casado, Foster, Bakken & Tjelta 2023, a review). That intensity discipline on easy days matters. That 10-minute sub-threshold reps sit at roughly 2–3 mmol (Talsnes 2024). That intermittent swimming at critical speed approximates steady state where continuous swimming does not (Nikitakis 2019).

CONTESTED. That three sub-threshold singles beat a conventional plan at matched volume. There is no randomised trial of this method. The evidence is one athlete's forum log plus a large, self-selected online community with obvious survivorship bias and near-universal confounding — most adopters increased frequency and consistency at the same time as they changed the sessions. The elite model it is named after is no better off: Casado et al. state outright that "no controlled studies have tested the efficacy of this training model."

Worth being precise about one study that gets cited on the wrong side. Talsnes 2024 compared one 6 × 10 min session against two split sessions 6.5 h apart, and found the single session drifted upward (HR 168→173, lactate 2.60→3.01 mmol) and cost more in next-morning fatigue. That is an argument for doubles, not for singles. It also ran on a treadmill at a fixed 10.5% incline with 11 of 14 subjects cross-country skiers, which largely removes the mechanical loading the method is supposedly protecting against.

LORE. The specific rep-and-rest menus — nothing distinguishes 10 × 1000 m on 60 s from 5 × 2000 m on 60 s at matched lactate. The claim that VO2max work is never needed. CTL targets. And the confident mapping of a given race pace onto a given mmol value: in recreational runners, fixed anchors estimate LT2 speed with a mean absolute error of 0.4–0.8 km/h and LT2 heart rate 2.8–5.2 bpm, and using estimated rather than measured maxima worsened that to 1.1 km/h and 6.7 bpm (Nuuttila 2025, n=165). A pace-prescribed "2.5–3.5 mmol" is an approximation, and erring slow is the documented safe direction.

One more claim to hold loosely: that your threshold lactate falls as you get fitter. Bakken says so from his own testing, and it is his stated reason for working at 2.3–3.0 rather than 4.0. But Beneke 2000, in 33 subjects, found maximal lactate steady state independent of performance. Do not ratchet an athlete's target downward across a season on that premise.