SNAPSHOT - JANUARY 2026. ⸻

How I got here

Although I conceived the idea of kayaking around Britain some years ago, I only began training in earnest in October 2025. I’d just returned from a month of cycling through France and Spain with my wife — our first proper project after I retired as Director of Sport and Outdoor Education at Norwich School.

I came back bike-fit, cardio-fit, and buzzing to get back in a kayak. Earlier in the year I’d been paddling around 30–50 km per week, and my plan was to ramp this steadily to 100 km per week by Christmas. Alongside this I intended to:

Integrate regular sea paddling with river-based fitness paddling
Build confidence with overnight trips
Underpin everything with Core and Pilates sessions
Maintain bike fitness with one or two rides each week or fortnight
Add some running, simply because I had the time and thought it would help

With time on my hands and a clear goal, I got stuck in and expected steady progression in building fitness. With everything being recorded via Garmin, and also using Strava, I thought I had everything I needed to monitor and achieve the progress I was aiming for.

It didn’t happen!

Through October and November I went backwards. I missed training targets, became ill, and suffered a recurrence of a longstanding back injury. My HRV plunged and stayed stubbornly low for several weeks.

Eventually, taking advice from my physiotherapist wife and daughter and coaching friends, I accepted that pushing on wasn’t the answer. I stepped off the “treadmill”, rested for two weeks, and watched my recovery markers — particularly HRV — rebound.

That pause forced a rethink. I realised that both the nature of the challenge and my age (I’m 58) demanded a different approach. Fitness alone won’t get me around Britain. If my back goes, if I get injured, or if I become ill, I’m not going anywhere — no matter how fit I am.

It also became clear that my rest and recovery simply weren’t sufficient. As it was put to me, I needed to “treat sleep like miles, more than I treat miles, like miles!”

Resilience, not fitness, has to be central.

Since late December, I’ve been building-back progressively and deliberately: placing far greater emphasis on rest and recovery, replacing cardio sessions with strength work, and doing so very gradually to actively reduce the risk of illness and injury. Trying to build but not push too hard seems sensible but sometimes doesn’t feel like enough. Analysis and feedback are therefore crucial.

This post summarises the approach I’m now taking — one that continues to evolve. It still relies on logging data, but with a far more discerning view of which metrics matter, and with progress audited rather than justified.

⸻

Using AI - My ChatGPT Agent

Central to my experimental approach is a growing use of AI — specifically ChatGPT — to help analyse and inform my thinking. This sits alongside the human input from a few specialists I’m, lucky to have around me, namely Lisa, my physiotherapist wife, Eleanor, my physiotherapist daughter and international kayaking coaches, Tim Scott and Dyson Pendle.

This blog focusses on the use of AI for monitoring and guiding training. Over previous weeks I’ve built a “relationship” with ChatGPT and it’s getting to know me and my goals rather well. Indeed, the approach outlined here that I’m experimenting with, was honed in discussion with ChatGPT.

As I’ve used this system more, I’ve realised three key things:

Over time, AI really does get to know you as a person — your routines, strengths, weaknesses, and how these relate to your short- and long-term goals. It remembers stuff and brings it back when relevant. This emerges from the information that I have given it and from regular metric updates.

Its default mode is unwaveringly enthusiastic and supportive, which can be nice for the ego but less helpful when I need objective, neutral — perhaps even cynical — feedback. This is particularly true when assessing whether I’m under - or over-training. Towards the end of January, I reconfigured my agent to be much more neutral in its feedback, and I’m assessing whether this shift changes its content and tone in a helpful way.

It can be surprisingly nuanced in its advice and guidance, often pushing back on my proposals or suggesting options I hadn’t considered, with referenced justifications. This happens even in its most supportive mode.

I’m now moving into a slightly different relationship with AI, where it plays more of an auditor role than that of an enthusiastic ally. I’m hoping this will tighten up my training decisions and overall preparation.

That’s the theory, anyway! Let’s see!

⸻

What should I aim for?

As I move deeper into preparation for Britain by Kayak 2026, I’ve realised that the biggest risk in preparing for a long expedition isn’t a lack of motivation or effort — it’s quiet overreach.

When you’re intrinsically motivated, enjoy training, and have a big objective on the horizon, it’s surprisingly easy to explain away fatigue:

“It’s fine — this is what I’m training for.” When fatigue compromises immune response, you get ill. When fatigue rules, adaptations aren’t made as positively as they should be. Performance gains are made in the rest periods. Not keeping a balanced, symmetrical body leads to muscular-skeletal injury.

So, the purpose of my training system isn’t to constantly push harder. It’s to try to build strength and durability, while staying well enough to keep training and replicate what will be needed on the expedition itself. That requires something different from most fitness-tracking setups and its different to preparing for a specific race.

⸻

The expedition reality

My Britain by Kayak project isn’t about peak performance, a specific race or event or breaking records. It’s about durability and being able to enjoy, what should be an amazing adventure, safely.

On expedition I’ll need to:

Paddle for multiple hours per day, predominantly in Zone 1–2

Push and handle a 60–70 kg loaded sea kayak in awkward conditions

Carry that boat and kit up Long Beaches, over soft sand, rocks, stone and cliffs to find a campsite.

Set up camp, sort/find food and water, and manage life at the end of each day

Do this day after day, week after week

There will be short bursts of higher intensity (Zones 3 & 4) — tide races, foul-tides, headwinds, surf landings — but success depends on resilience to fatigue.

Good planning and decision making will hopefully avoid the need for heroic efforts. On expeditions, the mantra is “keep moving forward, safely”.

“It isn’t the mountains ahead to climb that wear you out; it’s the pebble in your shoe.” — attributed to Muhammad Ali

Small, persistent, or overlooked issues (the pebble) are usually more debilitating than large, obvious obstacles (the mountain or sea, in this case). That means kit, nutrition, health and wellbeing are central if the goal is to be achieved.

That context matters, because it defines what good training actually means.

⸻

The problem with most fitness platforms (e.g. Garmin and Strava)

Most training platforms are designed to:

Reward intensity
Smooth out variability
Encourage consistency without questioning cost
Stroke egos (Segments, badges, insights etc)

They’re excellent at answering:

“How fit am I?”

They’re much worse at answering:

“Can I repeat this for weeks without breaking?”

For expedition preparation, that second question is the only one that matters.

⸻

Separating data from judgement

The first principle of my system is simple:

Logging and judgement must be separated.

Daily: I log, I don’t (now) analyse

Each day I record a repeatable set of metrics on a spreadsheet (one row per day), listed below. These I add to ChatGPT as a daily training log.

HRV - Overnight and 7-day average
Sleep score and duration
Overnight resting heart rate
BOLT (Body Oxygen Level Test) scores (two standard, one maximal breath-hold time after exhaling)
Activities performed
Aggregated duration of activity for the day
One aggregated Relative Perceived Effort (RPE) score for the whole day
Step count (for background load that can be significant but often unaccounted for)
Brief factual notes (fatigue, soreness, late finish — no justification)

If I do multiple activities in a day — paddling, Pilates, strength — I aggregate them. One duration. One RPE. One honest answer to the question:

“How hard would it feel to repeat this entire day tomorrow?”

This captures stacking fatigue without creating admin.

⸻

Weekly: evidence, not interpretation

On top of my daily inputs, once per week I add to ChatGPt one screenshot from Garmin (desktop) showing my activities for the week.

That’s it. No totals. No analysis. No fitness scores.

The screenshot provides:

Session frequency
Multi-session days
Modality mix (paddling, mobility, strength)
Visual clustering that spreadsheets hide

Garmin is my truth source. Strava (free version) is optional context — I’ve let my Strava subscription lapse while I test this system.

⸻

Why I’m not chasing (certain) popular metrics

I deliberately avoid building decisions around:

Relative Effort scores (Strava)
Fitness/Freshness charts
Intensity Minutes (Garmin)
Algorithmic training-status labels

These are derived metrics. Useful for motivation, but too optimistic for expedition risk management.

Instead, the backbone of load is:

Duration × honest RPE

Simple. Transparent. Hard to manipulate.

⸻

Strength training: judged by expedition relevance

Strength work isn’t assessed by gym numbers or soreness.

For me, it’s judged by one question:

“Does this make it easier to handle a loaded kayak when I’m tired?”

I now separate activities clearly in Garmin:

Pilates / mobility → control, stabilisation, tissue quality, postural balance
Strength → press-ups, chin-ups, cable-machine work, farmer’s carries, and load-bearing patterns such as deadlifts / Romanian deadlifts

High-rep upper-body endurance work scores positively if it remains sub-maximal and doesn’t erode recovery.

Strength that compromises repeatability is treated as less helpful.

I even record some dog walks as “Rucking” when I also carry a 10kg rucksack.

⸻

The audit: where judgement happens

Crucially, analysis only happens when I explicitly ask (ChatGPT) for it.

Once per week, I trigger ChatGPT to perform a training audit that:

Prioritises objective trends over narrative
Assumes risk when evidence is ambiguous
Flags monotony, stacking, and quiet accumulation
Judges training against expedition durability, not peak fitness

The output is deliberately blunt:

🟢 Green / 🟡 Amber / 🔴 Red verdict
Clear identification of risk
A conservative, non-negotiable recommendation for the following week

No encouragement. No spin — although justifications can be requested if needed. A copy of the very first audit is printed at the end of this blog to give you an example of the feedback given.

(Note-the AF (Atrial Fibrillation) issue mentioned is a long-term issue and follows an ablation procedure I had in 2009. It occasionally recurs and training stress may be a trigger but I am beginning to think it may also be linked to some food intolerances. Another thing to monitor!)

⸻

Why this approach (might) work

This system may benefit my preparation because it:

Keeps daily effort in check
Prevents me negotiating with myself
Surfaces problems early, not when I’m injured or ill
Aligns training decisions with expedition reality

It isn’t exciting. It doesn’t chase popular fitness numbers. But it might just keep me on track.

And it answers the only question that matters:

“Will this training still be working for me three months from now, on a wet west coast, with no rest days?”

That’s the standard I’m training to.

Will it work? I can’t run a controlled experiment, but I do know that what I chased in autumn 2025 didn’t work — so I’m trying this approach, and I’ll let you know how it goes.

The potential benefit of using AI, is the endless ability to interrogate data, obtain insights specific to me and my goal and “discuss” a way forward. The art is to keep an open mind, triangulate responses and feedback and ask for justifications if it isn’t automatically offered, which. mostly it is.

⸻

This approach forms part of my wider preparation for Britain by Kayak 2026, where the goal isn’t to be fast — it’s to be there every day, strong enough to keep going. To keep moving forward - if only a little!

⸻

Example of Weekly Training Audit produced by ChatGPT (Verbatim)

Week ending Sunday 1 February 2026
(Garmin activity log + daily diary context)

This audit reviews what actually happened during the first full week of my new metrics-based logging system.
The purpose is risk assessment, not justification.

1. What I actually did (objective)

Total recorded activity (Garmin only): ~20 hours

Breakdown:

Cross-country skiing: ~15.5 hours
Kayaking: ~3 hours
Strength + Pilates: ~1.5 hours
Walking + swim: ~1.5 hours

This is a large training week, regardless of how controlled or steady much of it felt.

2. Intensity reality (not intention)

Heart rate evidence

XC skiing average HR:
- ~105–118 bpm
- Upper Zone 1 to mid Zone 2
Kayaking (26 Jan):
- Avg HR ~140 bpm → clear Zone 3
Strength session (29 Jan):
- Avg HR ~114 bpm
- Max HR ~171 bpm

Key point

Even when the intention was low intensity:

Long duration
Cold exposure
Technique learning

…combined to produce real autonomic and sympathetic load.

This was not a low-stress week.

3. Density and stacking (where risk lives)

Observed stacked stressors

Mon 26 Jan:
Zone 3 paddle + walking + Pilates → atrial fibrillation symptoms later in the day
Thu–Sun:
Consecutive 4–6 hour ski days
Cold exposure daily
Travel mid-week
Cognitive load (learning skate technique, biathlon skills)

There were very few genuinely low-load days.

4. HRV behaviour (critical insight)

What HRV did

Stayed relatively high through much of the week
Did not crash immediately
Symptoms appeared after load, not during

Interpretation

This is a classic lag response:

HRV appears “fine”
System compensates
Symptoms emerge later

This confirms that HRV alone is not sufficient as a decision tool for me.

5. Steps: the missing load

Adding daily steps to my log explains a lot:

Several days exceeded 12,000–18,000 steps
Often layered on top of:
- Long ski sessions
- Strength work
- Cold exposure

These were not rest days, even if labelled as such.

6. Cardiac sensitivity context (facts only)

Observed patterns:

Irregular / elevated heart rate occurred:
- After Zone 3 paddling
- Around travel days
- With reduced sleep
- With cold stress
Not triggered during steady skiing
Not triggered during low-intensity work

This suggests intensity density and stacking, rather than lack of fitness.

7. What this week actually was

This week was:

❌ Not a recovery week
❌ Not a light transition week
❌ Not “easy aerobic base”

It was:

✅ A high-volume, moderate-intensity, multi-stressor week
Handled reasonably well
But close to my tolerance ceiling

8. What changes going forward

Non-negotiable rules

No Zone 3 paddling when:
- Sleep <7 hours
- Travel day ±1
- Cold exposure is high
One true low-load day per week:
- <8,000 steps
- No training >30 minutes
Treat skiing days as:
- Primary training stimulus
- Everything else optional

9. Reframing success

Success = absence of symptoms, not volume
Fitness is already sufficient
Risk management is now the limiter, not motivation or discipline

Bottom line

I did not undertrain.
I did not significantly overreach.

But I operated close to the edge while HRV masked accumulating stress.

This audit justifies:

Restraint when HRV dips
Avoiding “squeezed-in” sessions
A calm, conservative re-entry after skiing and travel

This is the shift from training by effort to training by risk management.

ACTION PLAN FOR WC 02/02/26 (Post ski camp and covering Oxygen Advantage Course, Shannon, Ireland)

Revised Plan: Travel + Oxygen Advantage Week (Shannon)

This week is not a training block.
It’s a regulation and consolidation phase wrapped around travel and a cognitively demanding course.

The objective is:

Arrive settled, stay calm through the course, and return home ready to build again.

Constraints (by design)

Travel on Tuesday (UK → Ireland)
No training kit in Shannon
Oxygen Advantage course Thu–Sat (mentally and physically demanding)
Home late Saturday
Training time limited to:
- Walking
- Bodyweight / core
- Breathing work

That’s not a limitation — it’s the plan.

What I Will Include

Daily Walking (Every Day)

Walking is the anchor this week.

30–60 minutes most days
Easy, conversational pace
Flat routes where possible

Target:

~6–12k steps/day
No “step chasing”

Walking keeps circulation moving and helps regulate the nervous system without stimulation.

Bodyweight / Core (3–4 short sessions)

These are form sessions, not workouts.

20–40 minutes
Trunk, hips, shoulders
Slow, controlled tempo
Plenty of rest between sets

Examples:

Pilates-style sequences
Press-ups, dead bugs, side planks
Light mobility between movements

If it feels like training, it’s too much.

Breathing Work (Daily, light)

This is not about pushing BOLT scores.

Gentle nasal breathing
Longer exhales
Calm rhythm

The course itself will provide enough stimulus.

What I Will Not Include

No cardio sessions
No intervals
No Zone 3 work
No volume targets
No “squeezing something in”

The course days are already a load — just a different kind.

Day-by-Day Shape

Tuesday (Travel Day):

Walk when possible
Mobility in the evening
Early night

Wednesday (Free Day in Shannon):

Longer walk (45–75 min)
Short core session
Otherwise rest and settle

Thursday–Saturday (Course Days):

Course attendance = primary load
Optional short walk in the evening
Very light mobility only

Saturday Night:

Travel home
No training

Sunday (First Day Home):

Walk + gentle mobility only
No decisions about “getting back into it”

How I’ll Know This Week Worked

HRV steadier by the end of the week
No cardiac symptoms
Better sleep consistency
Feeling keen — not obliged — to train again

That’s the green light for February.

Bottom Line

This is not a week to add fitness.

It’s a week to:

Reduce noise
Improve regulation
Let January’s work settle
Learn skills that support long-term resilience

Training doesn’t stop this week —
it just changes form.

This is still training by audit, not ego.