Ironman Training Diary - June 24, 2026

🤖 Apollo (AI) analysis

Metric	Value
Distance	5.00 mi (8.05 km)
Duration	62:55
Pace	12:34/mi (7:49/km)
Avg HR	163
Max HR	185
Cadence	156 spm
Stride	81.9 cm
Elevation gain (corrected)	6 m
Z4 time	37:32 (60%)
Z5 time	9:09 (15%)

HR drift: first 5 min 125 bpm → last 5 min 181 bpm (+56 bpm)

What the data shows

This was your first recorded run since the Saturday June 20 double (8.05 km long + 3.24 km hard). Today covered the same 8.05 km as Saturday morning’s long run — same neighborhood, four days apart — which sets up a clean side-by-side:

	Sat AM (Jun 20)	Wed (Jun 24)
Distance	8.05 km	8.05 km
Duration	59:55	62:55
Pace /mi	11:59	12:34
Avg HR	159	163
Max HR	185	185
Z4 time	36:19 (61%)	37:32 (60%)
Z5 time	2:49 (5%)	9:09 (15%)
Cadence	147 spm	156 spm
Stride	91.5 cm	81.9 cm
Splits recorded	33 (fragmented)	1 (continuous)
HR drift (first 5 → last 5 min)	+22 bpm	+56 bpm

The headline number is the +56 bpm drift over the 63 minutes. First 5 min averaged 125; last 5 min averaged 181 — within 4 bpm of your recorded max HR. Effort was pinned at the top by the end.

Saturday’s run logged 33 splits, with several short, slow segments (0.5–1.5 m/s) interleaved between running segments. Today logged exactly 1 split — uninterrupted from start to finish. That single structural change explains a large fraction of what you’re seeing:

• No walk-break HR recovery → higher cumulative cardiac load → bigger drift.
• More Z5 time (15% vs 5%) is consistent with HR being allowed to climb without being shed.
• The avg-HR gap (+4 bpm) is small for the same nominal effort because the average is dragged up by the late surge, not by the early miles.

What the data does not show

This is not direct evidence of a fitness regression. A continuous run that ends at 181 bpm can mean: cardiac drift from sustained effort with no walk breaks, a deliberately progressive finish, heat, accumulated fatigue, or something physiological off-baseline (poor sleep, low fuel). The data can’t separate these.

Confounders worth naming:

• Temperature was not recorded (sensor fields are null). Dublin in late June is warm; a 60-min continuous run mid-morning is exposed.
• Sleep architecture last night was unusual — 9 h 16 m total time in bed, but only 26 min of deep sleep. Light + REM filled the rest. Deep-sleep paucity is associated with reduced perceived recovery, though one night is noisy and not conclusive.
• No training_load or training_readiness rows are populated for the week. Whatever Garmin’s algorithm said about acute:chronic load, it’s not in the database to check.
• Elevation was actually flat today (6 m gain, corrected). Saturday’s 55 m was almost certainly altimeter drift, not real terrain. Rule out “Saturday was easier because flatter” as an explanation.

The cadence-stride change is observable, not diagnostic. Today: shorter steps (81.9 cm) taken more frequently (156 spm). Saturday: longer steps (91.5 cm) at lower turnover (147 spm). The shift can come from many sources — fatigue, conscious focus, different shoes, different surface — and one run isn’t enough to attribute it.

Honest read

This is one continuous run versus one fragmented run. Until you replicate today’s continuous-effort shape on a fresh, well-rested, cool day, you can’t tell whether the +56 bpm drift represents a problem or just what any of your runs would look like without walk-break interruptions. Strong prior: the continuous structure is doing most of the work in this comparison.

You also haven’t run in four days. Whatever the next run shows, that gap is a variable to control for — not a baseline to assume away.

Recommendations (low confidence)

• Run the same loop again within 5–7 days, continuously, at the same nominal pace, in similar conditions. Compare drift. If it’s still +50 bpm, you have a signal worth investigating. If it drops to +25–30, today was conditions- and freshness-dependent.
• Cap the next two runs at HR 150 using the watch alert. If you can’t hold 150 at any pace, that’s a more important data point than today’s average.
• Get temperature into the dataset if your watch supports it, or log start-time temp/humidity manually. Heat-confounded analysis is mostly storytelling.
• Don’t read 12:34/mi as a slowdown yet. Saturday’s 11:59/mi was a stops-included average; the actual running pace between walk segments was closer to 11:00/mi. Today’s 12:34 was every minute of running. The two paces aren’t directly comparable.

VO2 max still reads 39.0, same as Saturday. As before — don’t watch it day-to-day.