training 10 aug – 16 aug 2026

Week ending 16 aug 2026

On track for 2:40-2:43. The limiter is specificity: goal-pace finally appeared, but never for longer than 5 km.

Every run in the week of 10 aug 2026 to 16 aug 2026
day	session	km	time	pace	hr	load
mon 10 aug	Afternoon Run	5.2	26:36	5:05	139	28 e
Rain and hail. 5:05/km.
tue 11 aug	Rep ladder + 15min MP	15.3	1:03:44	4:10	167	97 e
5-5-4-3-2-1-30s-30s off 60s, then 15min MP. Reps 3:42 / 3:41 / 3:43 / 3:34 / 3:31 / 3:23 / 3:06 / 3:12 per km. Closing 15min block 3980 m at 3:46/km, HR 179.9 - first goal-pace-band running of the block, though 79 s of stoppage sat inside it.
wed 12 aug	Herdsman + 4x400, 5x800	20.3	1:38:09	4:50	162	140 e
11.65 km approach at 4:57 (194 m vert), then 4x400 in 3:55 / 3:50 / 3:40 / 3:37 and 5x800 off 200 m jog in 3:39 / 3:40 / 3:39 / 3:44 / 3:41. HR fell across the 800s: 176.3 / 171.1 / 168.9 / 164.9 / 167.6.
fri 14 aug	Bridges	12.2	58:44	4:48	157	79 e
4:48/km whole-run. Undulating rather than flat-easy - Mount Street climbs inside it ran 183-184 bpm, which is where the 187 max comes from.
sun 16 aug	Long run, 4x5k @ MP	36.4	2:33:18	4:12	160	214 e
11 km approach, then 5k reps at 3:50.0 / 3:55.2 / 3:53.4 / 3:56.6 per km off 1 km floats, HR 170.4 / 170.3 / 168.1 / 168.8, climb 9 / 31 / 43 / 87 m. 20 km of MP at 3:53.8 average with a net HR decline. Whole run 4:12/km at 160.4 bpm.
sun 16 aug	Double	0.7	3:43	5:10	146	4 e
5:10/km shakeout.

Goal-pace entered the data for the first time in this block. Sunday's first 5k rep ran 3:50.0/km at 170.4 bpm - inside the 3:43-3:50 window and 1.6 bpm below the 172-177 marathon band floor. Tuesday's closing 15-minute block ran 3:46/km. Before this week no long-form effort had run faster than 3:52.
Twenty kilometres of marathon-pace work on Sunday at 3:53.8/km average with essentially zero drift: 170.4 / 170.3 / 168.1 / 168.8 bpm across the four 5k reps, a 2.3 bpm spread and a net decline. The identical 20 km at the identical average pace two weeks earlier drifted 9 bpm and finished at 175.8. The last rep this week ran 7 bpm cheaper while carrying 87 m of climb against 54 m.
Wednesday's 5x800 held 3:39 / 3:40 / 3:39 / 3:44 / 3:41 with heart rate falling across the set - 176.3 down to 164.9 - 24 hours after Tuesday's session peaked at 188. No recovery penalty from stacking the two days.

Volume fell 19% week on week, 111.6 km to 90.3 km over 6 runs and 5 run-days, while work above 165 bpm rose from roughly 17% of volume to roughly 40%. Thursday and Saturday were full rest. Two weeks out this reads as taper onset with quality sharpened rather than a shortfall, but that is inference from the shape of the week, not from anything stated.
Tuesday's 15-minute block at 3:46/km carried 179.9 bpm average and 188 max - about 3 bpm above the marathon band ceiling - and sat at the end of an eight-rep ladder, so it is a fast block rather than a marathon-pace datum. It also contained 79 s of stopped time inside 979 s elapsed, so 3:46 is running pace, not continuous pace.
Sunday's rep pace decayed across the run at matched heart rate: 3:50.0 at 170.4 (km 11-16), 3:55.2 at 170.3, 3:53.4 at 168.1, 3:56.6 at 168.8 (km 29-34). Climb per rep ran 9 / 31 / 43 / 87 m, so most of that decay is terrain rather than fade, but the grade-adjusted remainder is the number that matters on race day.

The 2:40 criterion - 15 km or more inside 3:43-3:50 on fatigued legs at 177 bpm or less with under 10 bpm of drift - moved from not shown to approaching. Heart rate and drift are now comfortably satisfied over 20 km. The only unmet leg is duration at goal-pace: the longest goal-pace block on record is 5 km. With the race on 30 August there is roughly one quality week left, and it will be taper-length, so this criterion will most likely go untested to the line and the call will rest on the evidence already in hand.
Sunday's long run sat on 53.0 km and two hard days of prior weekly load, against 71.6 km and three hard days into the equivalent session two weeks ago. Part of the 7 bpm improvement on the final rep is reduced accumulated fatigue rather than fitness. The comparison is real but not clean.
Projected range moves to 2:40-2:43 from 2:41-2:44, centre of mass 2:41-2:42. The 2:44 end no longer fits 20 km at 3:53.8 with heart rate falling. The 2:40 end is now carried by two independent lines - marathon-pace heart-rate cost and an actual goal-pace rep - rather than one, but it still rests on extrapolating 5 km, not on any sustained block.

Quiet. No symptoms reported at either dormant site, nothing cut short, every rep completed: Tuesday's full eight-rep ladder plus the 15-minute block, Wednesday's 4x400 and 5x800, Sunday's 4x5k.
Friday's 187 max on an otherwise easy 12.2 km is real rather than a strap artifact - the Mount Street climb segments inside that run average 183-184 bpm, so the peak has a cause. The default artifact reading does not apply here.