A triathlete swims a steady 3km set with paddles. Comfortable, controlled, nothing close to threshold. They climb out, check the numbers, and find the session has logged a Training Stress Score that suggests they just did something brutal.
They didn't. The metric did. And if that number feeds into how the next few days are planned, the athlete is about to be rested for fatigue they never accumulated — or worse, the inflated swim score masks a genuine spike somewhere else.
Triathlon has a load problem that pool swimmers and single-sport runners never face: three disciplines, three different ways of measuring effort, and no honest common currency between them.
Why one number can't hold three sports
The appeal of a single training-load metric is obvious. Three sports, one busy week, one figure that tells you whether you're building fitness or digging a hole. Training Stress Score became the default answer, and for cycling — where it was born, calculated from power — it works well.
Then triathletes tried to stretch it across all three disciplines, and the seams showed.
The swim is the worst offender. Swim load gets estimated from threshold pace, and it tends to overestimate — especially when paddles or other tools are involved — because you can cruise along well below threshold, barely working, and still rack up a score that implies a savage session. The reason is physical, not a software bug: water resists you far more than air, so the physiological cost of swimming rises disproportionately with speed — the power required climbs roughly with the cube of velocity — far faster than it does in running. A pace-based model borrowed from land sports simply doesn't map onto it. And unlike the bike, there's no swim equivalent of a power meter doing the measuring — the people who build these metrics say so themselves — which is why so many triathletes end up estimating their swim load by hand and hoping.
Running sits in between — more reliable than swimming, less clean than cycling, and thrown off by terrain and gradient that pace alone can't capture. So you end up with a "unified" load number assembled from one trustworthy input (bike), one rough one (run), and one that actively misleads (swim) — then averaged together as though all three were equally sound. The single number feels authoritative precisely because it hides which parts of itself you can't trust.
The problem isn't too little data — it's too much, unacted upon
Here's the part that should sting. A 2024 pilot study — Procida and colleagues, published in Sports — looked at how age-group triathlon coaches actually manage load when handed the data. Across a training block, weekly load spikes — jumps of more than 10% — showed up in 83% of swim weeks, 74% of bike weeks and 87% of run weeks. Spikes everywhere, in every discipline.
And the coaches? They were given the load reports. They acted on them in roughly a quarter of cases — and when they did adjust, most of those changes still failed to bring the week's load variability back under control. The training lost to injury clustered, predictably, after the biggest jumps, and it was running load — the weight-bearing discipline — where the largest spikes did the damage.
It's a small pilot, so hold the exact percentages loosely. But sit with what it points at. The information existed. It was in front of the coach. And it still didn't translate into a calmer, better-managed training week — because three disciplines' worth of raw load data, scattered across tools and muddied by a swim metric you can't trust, is not the same as clarity. It's noise wearing the costume of insight. A coach buried under three streams of partly-unreliable numbers doesn't automatically make better decisions than one going on feel. Sometimes they make worse ones, because the numbers carry false confidence.
What actually helps a triathlete
The fix isn't a cleverer single score. It's three honest things at once.
First, treat each discipline's load on its own terms rather than forcing them into one flattering average — and be explicit about which inputs you trust and which you don't. A swim load you know is inflated is more useful than a combined number that quietly buries the inflation.
Second, watch the spikes, per discipline. The study's clearest signal was that sudden jumps in a single discipline — especially run load — were where injuries lived. A swimmer-cyclist-runner can be sitting on a sensible total while one discipline lurches up 30% underneath it. The total hides exactly the thing you most need to see.
Third — and this is the part land-based load models leave out entirely — fold in how the athlete is actually recovering. Sleep, soreness, energy, readiness. Three disciplines means three sets of demands on one body and one nervous system, and that body keeps a single account regardless of how many sports are drawing on it. The swim doesn't fatigue a separate athlete from the one who runs. A readiness signal cuts through the metric-mismatch problem by measuring the thing all three disciplines ultimately feed into: whether the athlete in front of you can absorb what you've got planned.
You don't need a perfect unified load number. No such thing exists, and chasing it is how triathletes end up rested for phantom swim fatigue while a real running spike goes unnoticed. What you need is to see each discipline honestly, catch the spikes before they catch you, and keep one clear eye on whether the whole athlete — not the swim leg, not the bike split, the athlete — is actually ready.
Three sports. One body. One readiness picture that doesn't let the swim lie to you.
We built our load and readiness tracking to handle the multi-discipline athlete honestly — per discipline, spike-aware, and anchored on how the whole athlete is recovering. Join the founders list and we'll show you what that looks like across all three.