colin dismuke
Search esc to close

training

Training

A running log kept for the long view. A week of running says almost nothing on its own, so this is the same record at three lengths: the week in full, then the weeks around it, then the year they sit in. It moves once a week.

goal race
Sydney Marathon
date
30 aug 2026
out
34 days · 4 wks
target
2:40 to 2:45

The week

week ending 26 jul 2026

On track for 2:42-2:44 with 2:40 upside indicated but unconfirmed. The limiter is specificity, not fitness: zero marathon-pace exposure this week.

runs
7
distance
105.7 km
time
8h 21m
elevation
774 m
Every run in the week of 20 jul 2026 to 26 jul 2026
day session km time pace hr load
mon 20 jul Jog 8.8 43:42 4:59 132 130
tue 21 jul 200's + 1k's + 400's 13.1 1:01:33 4:42 156 224
wed 22 jul Herdsman 21.3 1:48:50 5:07 140 317
wed 22 jul 4x2k threshold 12.3 52:16 4:16 159 221
fri 24 jul Upper Zone 2 10.3 53:16 5:11 131 175
sun 26 jul Long run 39.1 2:56:38 4:31 158 663
sun 26 jul Run home from the station 0.8 4:20 5:14 161 12

good

  • 4x2k threshold averaged 3:37/km (3:40 / 3:33 / 3:38 / 3:37) off 60s recoveries, run as the second session of a double after 21.3 km easy that morning. HR climbed 169 to 175 and topped out at 183. Eight kilometres at 6s/km faster than goal pace, on tired legs, with heart rate still inside the calibrated marathon band.
  • 39.1 km long run, the longest of the block, in 2:56:38 (4:31/km) with 356 m of climb. First 16.4 km at 4:23/km for an average HR of 159.6; final 22.7 km at 4:36/km for 157.4 with 289 m of the vert. Heart rate fell as the run went on. Essentially no aerobic decoupling over three hours.
  • Interval session held its shape end to end: 4x200m at 2:55-3:00/km, 3x1k at 3:24 / 3:25 / 3:26 (HR 176-178, max 188), 3x400m at 3:25 / 3:15 / 3:15. Every rep completed, no fade across the 1k set.

note

  • Distribution: roughly 51% of weekly volume easy or recovery, 37% the long run sitting in Z3, 12% at threshold or faster. Every standalone easy run landed at the recovery end (4:59, 5:07, 5:10/km) rather than in the easy band, so the aerobic base work this week was genuinely easy.
  • No run this week fell in either marathon-relevant pace band. Nothing at goal pace (3:43-3:50) and nothing at steady (3:55-4:08). Quality was threshold and VO2 only; the long run was unstructured endurance. Marathon-pace-specific work was absent for the first time in the block.
  • Load sequencing: the two push-off-heavy sessions landed on consecutive days. Tuesday's reps (200s, 1ks, 400s) then Wednesday's double, 21.3 km easy in the morning and the threshold set at lunch, making Wednesday a 33.5 km day. Both FHL tendinopathy and the left hamstring are provoked by the speed end, so consecutive placement concentrates the shared risk. Thursday and Saturday were full rest days.

watch

  • The 2:40 marker is still not demonstrated. It requires a fast-end goal-pace effort on rested legs at HR at or below 176 with no FHL or hamstring response, and that specific data point has now been absent for two straight weeks. Wednesday's threshold is supportive but is not the same thing: faster than goal pace, intermittent, and run on fatigued legs. Aerobic proxies are strong; the specific confirmation is missing.
  • Long-run specificity did not advance. Two weeks ago the long run carried 3x4k at marathon effort; this week's 39 km was continuous endurance with no discrete goal-pace blocks. Worth tracking whether goal pace returns to the long run and whether it lands later and more continuously.
  • No symptoms reported from either standing niggle. Nothing was cut short: all interval reps completed, threshold 4/4, and the 39 km finished clean. Tolerance is inferred from completion rather than directly reported, so treat it as awareness-or-below rather than confirmed silent.
  • Volume ran 105.7 km across 7 runs on 5 run-days, the second consecutive week below the typical 116 km / 8 runs, but concentrated into larger single days (33.5 km Wednesday, 40.0 km Sunday). Worth noting next week whether frequency rebuilds or this concentration is the shape of the peak phase.

5 weeks

22 jun 2026 to 26 jul 2026 · 35 days
runs
30
distance
438.3 km
time
33h 42m
per week
87.7 km
weekly distance avg 87.7 km/wk
Weekly distance, week by week
week endingkmΔ kmrunstime
28 jun28.9·32h 18m
05 jul116.0+87.288h 51m
12 jul84.9-31.266h 30m
19 jul102.9+18.067h 42m
26 jul105.7+2.878h 21m

12 months

01 aug 2025 to 26 jul 2026
runs
271
distance
4258 km
time
338h 39m
per month
355 km
monthly distance avg 352 km/month, whole months only
Distance by calendar month
monthkmrunstime
aug 2025313.31727h 23m
sep 2025359.32428h 04m
oct 2025325.92325h 12m
nov 2025262.32319h 49m
dec 2025324.41826h 02m
jan 2026300.81923h 46m
feb 2026351.52228h 14m
mar 2026473.12837h 34m
apr 2026465.02537h 30m
may 2026350.52329h 01m
jun 2026346.42426h 32m
jul 2026 to 26 jul385.32529h 32m
fitness and fatigue modelled tss, 42-day and 7-day averages
fitness
80
fatigue
93
form
+8
as of
26 jul 2026
  • fitness, 42-day average
  • fatigue, 7-day average
  • the gap, fitness minus fatigue
The model at the end of each month. Form is carried from the previous day, which is why it does not always match the gap on the chart.
month endfitnessfatigueform
aug 20257061+21
sep 20256761+5
oct 20256054+4
nov 20255346+1
dec 20256183-21
jan 20265653-4
feb 20266255-1
mar 20267682-8
apr 20268387-7
may 20267978+12
jun 20267252+25
jul 20268093+8

The curve runs on a heart-rate estimate recomputed the same way for every session, not on the stored load column. That column carries the numbers each source reported, and the two sources are on different scales, so feeding the mixture to the model would put a step at the seam between them and draw a fitness jump that never happened.