Training Log: The Metrics That Actually Matter

Anyone who logs their training stops training purely by feel — and that is exactly the point. A training log is not a chore for data nerds, it is the precondition for a plan being able to react to reality. Without records there is no trend, and without a trend every plan adjustment stays a gut feeling. The good news: the effort is about 60 seconds per session, and the eight metrics that matter are quick to note.

What a training log is supposed to do

A training log answers three questions no single session can answer: Am I developing in the right direction? Why does this week feel harder than last week? And what did I do differently four weeks ago when things were going well? That does not require an endless flood of data — it requires consistent, comparable entries over months.

The decisive difference lies in combining objective measurements with subjective perception. A watch delivers duration, distance, elevation, heart rate and power, all reproducible quantities. It does not deliver how a session felt, whether sleep was poor, or whether a stressful block at work held recovery back. Those are exactly the pieces that explain the outliers on which rigid plans break. A log without a comment is like a ship's log without weather notes: it shows where you were, but not why it looked like that.

Start small. A complete log with twenty fields per session gets abandoned after two weeks. Eight fields that are filled in reliably are worth more than thirty that stay empty. Consistency beats completeness — true in training control as much as in data analysis.

The eight metrics that matter

• Duration. Pure moving time, not the time between start and finish. In intervals without the rest periods, on rides without the café stop. Duration is the basis of every load calculation.
• Distance and elevation gain. The classic volume measures in running and cycling. Elevation is not a side column: 500 metres of climbing cost considerably more than their distance equivalent suggests.
• Average and peak pace or power. Note both, because only the peak shows whether the stimulus actually landed. If you wanted a tempo run at a correct average but never reached the target zone, the session missed its purpose.
• Heart rate. Average and maximum, plus time in the target zone for structured sessions. What counts is not the daily value but the comparison at the same load across weeks.
• Intensity distribution. How much time was easy, how much at threshold, how much hard? The distribution decides the training effect more than total kilometres do.
• RPE. Rating of perceived exertion on a scale of 1 to 10. The only metric that also reflects equipment failures, heat and stress.
• Resting heart rate and sleep. In the morning, before getting up, plus estimated sleep duration. The best free early-warning indicator you have.
• Short comment. One sentence is enough: "Heavy legs", "strides felt smooth", "side stitch at kilometre 12". Three months from now that is the most valuable entry in the whole log.

Why RPE belongs next to heart rate and pace

Training science uses RPE because it closes a gap no sensor can fill. Heart rate is sluggish: during short, hard intervals it lags the actual stimulus by 20 to 30 seconds, and in heat and dehydration it climbs without the muscular load having risen accordingly. Pace in turn says nothing about the internal cost — when tired, the same pace costs considerably more.

The scale is simple. RPE 1 to 2 is walking or very light movement, RPE 3 to 4 is the classic base run where you can hold a conversation without trouble. RPE 5 to 6 is the range where sentences are still possible but conversation stutters — that is threshold. RPE 7 to 8 is hard but controlled, RPE 9 to 10 is racing or a sprint finish.

What makes it attractive is a study from the running world: experienced runners who steered their sessions by RPE over a training block achieved practically identical performance gains to a group with fixed heart rate targets — with noticeably fewer missed sessions and misjudgements, because RPE automatically accounts for day-to-day form. The practical consequence for your log: record RPE right after the session, before you look at the analysis. Whoever checks the watch first unconsciously shifts their rating towards the numbers.

Heart rate and RPE apart — the most important comparison

The strongest information in a training log does not come from a single value but from the relationship between values. Two standard comparisons cover most diagnoses:

• High heart rate at the same pace. If pace is on target but heart rate on an easy session sits 5 to 8 beats above its usual range, heat, dehydration, sleep loss or an oncoming infection are usually involved. Do not panic, but watch the next two days.
• Same pace at a higher RPE. If the same session suddenly feels harder than two weeks ago while the numbers are unchanged, that is a fatigue signal — even when no objective value looks unusual. This constellation is usually weeks ahead of overtraining.
• Low RPE with improving pace. The nicest entry in the log: the session comes easier than before at the same heart rate. That is adaptation, not chance — and a reason why logs also work as motivation.

The second big advantage of a log is cardiac drift: comparing the first and last 30 minutes of a long run at the same pace shows aerobic stability. Under 5 percent is very good, 5 to 10 percent is normal, and above that either endurance is still limited or fuelling was insufficient. Neither shows up anywhere later if it does not make it into the log.

Weekly review: where the real value sits

Entering data is half the job, reviewing makes the difference. Once a week, ideally on the last rest day, five questions belong in the look-back:

• What was the distribution? Target under an 80/20 approach: around 80 percent of time easy, around 20 percent at threshold or high intensity. Deviations beyond ten percentage points are a planning signal.
• How did volume progress? The most common error is the jump. More than 15 percent weekly volume increase without an easy week in the third or fourth week measurably raises injury risk.
• What was the resting heart rate trend? Ignore single outliers, the weekly line counts. A rise of 5 beats across several days is a robust warning sign.
• What was missed — and why? Missed sessions with a reason are more valuable than completed sessions without a note. Three "no time" entries reveal a planning problem, three "tired" entries a recovery problem.
• Which session surprised you? Positive and negative surprises deserve a second line in the comment. They are the raw material for the next plan.

A monthly look at the bigger picture is worth it too: how many weeks were structured, how often did the plan get spontaneously rebuilt, how did the TSS trend develop? Anyone who sees that a cascade of rebuilds always ended in a weak training week will stop changing the plan after every poor session. That is the real return on keeping a log.

TSS, CTL and TSB in the log

Anyone combining several sports quickly hits the limit of kilometre lists: 60 minutes of swimming and 60 minutes of running are not the same thing. This is where the Training Stress Score (TSS) helps, condensing duration and intensity of a session into a single number. 100 TSS roughly equals one hour at threshold. An easy distance run sits at 40 to 60, a long run at 120 to 180, a hard interval session likewise above 100.

The daily series produces three metrics that make the log analysable: CTL as a rolling 42-day average of fitness, ATL as a 7-day average of acute fatigue and TSB as the difference between them — form. A TSB between −10 and +5 counts as the productive training zone, sustained below −20 as an overload signal, above +15 as a sign of an approaching race phase. How these curves work in detail is explained in the article on CTL, ATL and TSB.

The practical advantage: once every session lands in the log with a TSS value, the debate about whether a week was "hard" disappears. The trend shows it. And anyone logging TSS per session also sees which sessions carry the largest share of total stress — usually not the long runs, but the combination of long runs and the interval week that follows.

Digital tools and data protection

Training data is health data and therefore falls under the GDPR. If you choose a tool, it should be clear where the data lives and what happens to it. Three minimum requirements: servers in the EU, encrypted transfer, and the ability to export or fully delete all your own data.

A sports watch alone is not enough as a log. It delivers raw data but no assessment, no analysis and no context. Only when values are merged with planning and perception does the information that carries training decisions emerge. With Peakora, TSS trend, form curve and intensity distribution appear automatically alongside the planned sessions — which is why it makes sense to keep the log where the plan lives. Details on data handling are in the article on GDPR and fitness apps.

The smallest effort that works

A log does not fail because of missing fields, it fails because of too many. The pragmatic start looks like this:

• Right after the session (30 seconds): duration, distance, RPE, one sentence on how it felt. Done while you are still taking your shoes off.
• In the morning (20 seconds): resting heart rate before getting up, sleep duration in hours. Both effortless, both extremely informative in trend.
• Once a week (10 minutes): run through the five review questions, plan or confirm an easy week, check next week's plan.
• Once a month (20 minutes): review the TSS and resting heart rate trends plus the number of spontaneous plan changes — and derive one rule for the next month.

Stick to only those four rhythms and after three months you will have more reliable information about your training than most athletes with a full sensor package. The log is not an end in itself: it is the memory of your training and the justification for every adjustment you make.

Common mistakes with a training log

• Too many fields. Anyone trying to enter 25 values per session quits after ten days. Eight relevant fields filled consistently beat thirty filled occasionally.
• Entering RPE only after checking the watch. That unconsciously shifts your rating. Order: feel and note first, then add the numbers.
• Kilometres only, no intensity. Two weeks with identical volume can contain completely different stimuli. Without intensity distribution the review stays blind.
• Not documenting missed sessions. They carry the most important clues about schedule stress, recovery deficits or overload.
• Collecting but never reviewing. Data without a look-back is archival material. The value comes from the weekly comparison, not the entry.

Frequently asked questions about training logs

How to raise load correctly and when an easy week is due is covered in the article on adjusting your training plan. The basic terms around form, fatigue and intensity distribution are in the lexikon (German only). And if you want to connect your log to the bigger picture of a race build, the structure is in the 16-week marathon training plan.