Two runners post the exact same training load number this week. One trained smart. One trained in a rut.
Their fitness score (CTL) can’t tell them apart. That’s not a bug someone forgot to fix. It’s baked into the math. A little-known TrainingPeaks tool called Chronic Intensity Load (CIL) was built a decade ago to patch that exact hole, and almost nobody has heard of it.
Before we go further, a correction. Some recent write-ups call CIL a “new 2026 TrainingPeaks metric.” That’s wrong. CIL was built inside WKO4, TrainingPeaks’ desktop analytics software, back when WKO4 launched in July 2015. Coach Chris Baddick built it using WKO4’s custom-chart tool, working with TrainingPeaks’ Dave Schell. It’s not a first-party platform feature and it’s not new. TrainingPeaks itself says plainly: CIL “is not designed as a replacement for CTL, but simply a useful addition.”
So why write about a decade-old, lightly-used custom chart in 2026? Because the problem it targets is still completely unsolved across the industry. Here’s what it fixes, and just as important, what it doesn’t.
The Problem Chronic Intensity Load Was Built to Solve
Your fitness score comes from TSS (Training Stress Score), which blends how long you worked and how hard. Two sessions with the same TSS can feel nothing alike.
Here’s the proof. 80/20 Endurance ran the numbers on three real running sessions: 1.5 hours in Zone 2, 53 minutes in the murky “gray zone,” and 45 minutes in Zone 3. All three land on a TSS of exactly 75. Three completely different training stimuli, one identical number.
In short: your fitness score can’t tell a long easy run from a sneaky moderate grind. They look the same on paper.
Here’s a sharper version of the same trap. Picture two athletes, “Rebecca” and “Rachel,” each training 10 hours a week. Rebecca does more moderate-pace work and racks up 603 TSS. Rachel trains truly easy on her easy days and truly hard on her hard days, and racks up 570 TSS. Rebecca’s fitness score reads 5 to 10% higher. Rachel still beats her on race day, because Rachel recovers better and her hard sessions actually push her limits.
A car engine doesn’t care how many miles are on the odometer if half of them were spent idling in traffic. Your fitness score works the same way: it counts miles, not what kind of miles they were.
The math backs this up. A 6-week trial split 12 trained cyclists into a polarized group (80% easy, 0% moderate, 20% hard, about 6.4 hours a week) and a threshold group (57% easy, 43% moderate, 0% hard, about 7.5 hours a week). The polarized group trained less and gained more: bigger jumps in VO2peak, longer time-to-exhaustion, and better breathing efficiency. The threshold group trained more hours for worse results. Same rough training stress, very different payoff.
The Real Formula: CIL vs CTL, Side by Side
CIL doesn’t reinvent the wheel. It borrows CTL’s exact math and swaps one ingredient.
CTL_today = CTL_yesterday x e^(-1/42) + TSS_today x (1 - e^(-1/42))
CIL_today = CIL_yesterday x e^(-1/42) + (1000 x IF_today) x (1 - e^(-1/42))
CTL runs a 42-day rolling average of your daily TSS. That’s roughly six weeks of accumulated training, smoothed out. CIL runs the identical 42-day rolling average, but instead of TSS it uses your Intensity Factor (IF), how hard a session was relative to your threshold, scaled by 1000 so the numbers land in a familiar range.
Translation: CTL answers “how much did you train?” CIL answers “how hard, on average, were your efforts?”
This is also exactly where CIL’s honesty runs out. Because IF strips out duration completely, a blistering 5-minute effort and a steady 3-hour ride at the same relative intensity add the same amount to CIL. That’s a deliberate design choice, not an oversight; Baddick built CIL for mountain biking, cyclocross, and criteriums, where short, punchy efforts matter more than how long you were out there. It was never meant to replace CTL for steady-state sports like Ironman or marathon training.
What the Evidence Actually Says
CIL itself has no peer-reviewed research behind it. None. A search of the sports science literature for “Chronic Intensity Load” turns up exactly one source: the original 2015 TrainingPeaks blog post. That puts CIL in the same evidence tier as CTL and TSS themselves, useful coaching tools built by practitioners, never tested in a randomized controlled trial against race outcomes.
That’s not disqualifying. TSS-family metrics, called TRIMP in their original form, do have a body of research behind them, and it’s a mixed bag. Older Banister-style models correlate with real performance at about r = 0.65, which is decent but far from precise. And the underlying blind spot CIL targets is well documented outside CIL itself. A 2020 study in PLoS ONE had 10 runners complete five interval sessions with different work-to-rest ratios. Standard load formulas like TRIMP swung by 48 to 76% higher on longer sessions compared to alternative formulas, even when the actual physical demand was similar. The measuring stick, not the effort, was moving.
Here’s the short version: the flaw CIL targets is real and well studied. CIL itself, as a fix, has not been studied at all.
What CIL Still Doesn’t Fix
This is the part most write-ups skip, and it’s the part that matters most.
Durability is invisible to both metrics. Researchers now define durability as how well your pace, power, or efficiency holds up late in a long race, as a distinct trait from raw fitness. A 2023 study found durability improves with a mix of easy and hard training, and a 2025 paper formally separated durability from fatigue resistance and repeatability as its own measurable thing. Neither CTL nor CIL comes close to capturing it. You can have a huge fitness score and a high CIL and still fall apart in hour four of an Ironman.
One number can’t hold a training week. Whether it’s CTL or CIL, you’re compressing weeks of swims, rides, runs, lifts, and life stress into a single scalar. That’s always going to lose information. It’s like judging a whole movie by its average brightness.
It doesn’t transfer across sports. CIL is built from cycling power data. There’s no validated version for running pace or swim pace, and certainly none for a HYROX workout or a strength session. If your training week includes a heavy squat day alongside your bike rides, neither CTL nor CIL sees it.
Everybody responds differently. The same IF, the same TSS, the same drift ratio hits two athletes in different ways depending on age, training history, and recovery capacity. No formula, including CIL, adjusts for that on its own.
Take a rider I’ll call Marcus, 34, training for a cyclocross season on 8 hours a week. His fitness score sat at a steady 62 for two months, barely moving. But his coach started tracking his intensity average alongside it and noticed it climbing fast, from an IF-scaled reading in the low 60s to the high 70s over six weeks, as his short race-pace efforts got sharper. His overall load looked flat. His intensity was quietly spiking. That gap is exactly what a duration-only number misses, and exactly what pushed his coach to back off two weeks before his first race instead of loading more volume on top.
Hard Numbers to Keep Nearby
Whatever combination of load metrics you track, the risk guardrails around CTL still apply. TrainingPeaks’ own published guidance: 3 to 5 fitness-score points per week is a sustainable build rate for most athletes. 5 to 8 points a week is aggressive and should be short-lived. Past 7 to 8 points a week, injury and illness risk climbs sharply.
Bigger fitness score isn’t automatically better training.
That single sentence matters more than any formula here.
How AthleteOS Handles This Without a Bolt-On Metric
AthleteOS doesn’t ask you to add a second obscure chart on top of your fitness score. The training plan already tracks your fitness score (CTL) and fatigue score (ATL) per discipline, alongside a separate view of your actual zone-time distribution, so two sessions with identical training stress show up differently on your dashboard instead of collapsing into one indistinguishable number. If you want to see the same idea done the free, DIY way outside a paid platform, tracking CTL and TSB without TrainingPeaks walks through the options.
None of this claims to solve durability or cross-sport load. Nobody has cracked that yet, CIL included. But you don’t need a decade-old custom chart to at least see intensity and duration as two separate signals. You can sign up for AthleteOS and see your own fitness score alongside your zone mix on the next long block.
If you’re weighing whether a threshold-heavy or a polarized approach fits your season, that decision matters more than which load formula you’re staring at. The formula just tells you what happened. It doesn’t tell you what to do next.