Long run · Sunday, 16 August 2026

Long run, 23.3 miles

32 weeks to MDS

My write-up for this run isn't up yet. Everything the watch recorded is below.

Distance: 23.3 mi; 37.5 km
Moving time: 5h 31m; started 05:41
Average pace: 14:13; per mile
Heart rate: 118; average · max 139
Climb: 1430 ft; 436 m
Conditions: –; 71% humidity

How well the running held up

I compare how efficiently I was running early in the run (speed for each heartbeat, corrected for hills) with the late part. The drop was 10.8%. Over 5% is what I call a late fade. That band is a coaching convention, not a validated threshold.

A sustained drop began around 3h 06m into the run.

Late on, pace slowed 4.4% and heart rate rose 7.2%.

Confidence: medium (the late section had more climbing; the hill correction changed this reading by about 5 points, and it is an estimate).

Efficiency across the run against the early baseline (100%). The dashed line is 95%.

First half against second half

Comparing pace against heart rate in the first half with the second, the gap was 11%. Under 5% is the convention I coach to: it suggests the aerobic side coped with the length of the run. It says nothing about why.

Mile by mile

Mile	Pace	Avg HR	Climb
1	13:59	114	112 ft
2	14:03	115	92 ft
3	12:28	115	13 ft
4	14:03	114	92 ft
5	13:09	115	26 ft
6	13:00	115	7 ft
7	13:32	113	20 ft
8	13:21	108	13 ft
9	14:10	113	39 ft
10	14:35	117	102 ft
11	14:20	114	85 ft
12	15:02	116	102 ft
13	13:45	118	52 ft
14	15:15	118	102 ft
15	12:56	121	0 ft
16	14:20	116	36 ft
17	14:48	123	98 ft
18	14:59	124	72 ft
19	15:38	124	115 ft
20	14:54	123	82 ft
21	15:22	118	39 ft
22	15:22	119	39 ft
23	14:32	124	46 ft
0.29	12:37	124	0 ft

Figures from my watch via Intervals.icu, read by the Ultra Event Training app. Heart rate from a watch is a proxy for effort, not a lab measure. The efficiency figures compare speed per heartbeat early and late in a run with a hill correction; they show what changed, not the cause.