0
Tests passing
Causal performance engine
Know what changed, why it changed, and when the relationship holds.
AthDash is a causal engine for elite endurance and hybrid coaches. It separates a real effect from a coincidence — and when the evidence isn't enough, it says so plainly instead of guessing. It's also the grounding layer that keeps your coaching agents honest.
Reads from the tools you already use
WHOOPOURAGARMINTRAININGPEAKSINTERVALS.ICUAPPLE HEALTHFIT
01 / The problem
Correlation isn't a coaching decision.
Stack enough signals and something always lines up. A dashboard will happily show you the pattern. It won't tell you whether it's real, or whether you have enough data to act.
Everything correlates
The cloud of false signals
Sleep, HRV, load, mood, weather — track a dozen series and dozens of pairs will appear related by chance. Most of it is noise wearing a trend line.
One number, no context
A score that flattens everything
A single readiness figure collapses a dozen mechanisms into one digit. It tells you something moved — never which lever to pull, or by how much.
No sense of enough
It never admits uncertainty
Tools render a verdict on three data points as confidently as on three hundred. So when the sample is thin, you're guessing — and you can't tell that you are.
02 / Effect modifiers
The same input helps — or hurts. Depending on the conditions.
Most tools report one average effect. AthDash finds the conditions under which a relationship flips its sign, and shows you the threshold where it turns.
readout · training load → next benchmark
effect (HRV suppressed) −0.056
effect (HRV recovered) +0.022
modifier HRV vs. baseline
evidence exploratory · n=19
conditional slope
HRV ↗ rising
−0.056
Load hurts
HRV < threshold
Load helps
HRV ≥ threshold
03 / One engine
One set of numbers. Three surfaces.
The engine runs once. What changes is how you read it — a signed report, a live console, or a typed API. Same evidence, same uncertainty, everywhere.
SIGNAL FLOW · one estimate, three readouts
Deliverable
Audit
A signed causal report per athlete. Every claim traced back to the data behind it and the confidence interval around it.
Explore
Console
The interactive console. Drill from a driver to its evidence, its interval, and the caveats that bound it.
Integrate
API
The same numbers over a typed API — verdicts, intervals, and licenses you can drop straight into your own stack.
04 / The rigor
How the engine earns a claim.
Five steps, in order, every time. A claim only leaves the engine once it has survived all of them.
01
Ingest
sourcesPull from the wearables and tools you already trust — aligned to one timeline, units reconciled, gaps marked rather than filled.
02
Model
causal structureFit the structure behind the data, not just the correlations on its surface — controlling for the confounders that fake a relationship.
03
Test
held-outRun each relationship against held-out data and every known confounder. A pattern that only holds in-sample never ships.
04
Gate
honestyIf the evidence is thin, the engine says so: insufficient — don't claim it yet. Declining to answer is a valid output.
05
Deliver
with uncertaintyShip the verdict with its confidence interval and caveats attached — never a bare number stripped of the conditions that bound it.
05 / Proof
The discipline, in numbers.
0
Sources unified
0
Causal relationships modeled
0%
Claims shipped with a CI
0
Effect modifiers found
0
Overclaims shipped
06 / For AI coaches
Keep your coaching agent honest — by contract.
An LLM gives the same confident answer whether it knows or is guessing. AthDash gives your agent the one thing it can't generate for itself: a verdict it's allowed to make, the evidence behind it, and a hard stop where there's nothing to say.
› athdash_can_i_claim("sleep_regularity", "benchmark_workout", athlete="A.R.")
evidence: SUPPORTED (n=19, p<.001)
license: ADVISE
interval: [+0.78, +1.93]
› athdash_can_i_claim("ctl", "ftp", athlete="A.R.")
evidence: INSUFFICIENT (n=3)
license: DECLINE
evidence: SUPPORTED (n=19, p<.001)
license: ADVISE
interval: [+0.78, +1.93]
› athdash_can_i_claim("ctl", "ftp", athlete="A.R.")
evidence: INSUFFICIENT (n=3)
license: DECLINE
License ladder · low → high authority
DECLINENot enough of their data to judge yet
BORROWLean on the cohort prior — flagged as borrowed
WITHHOLDSignal is within the noise — don't claim it
FLAG_WEAKWeak signal; mention only if asked
HYPOTHESIZEWorking hypothesis — propose an N-of-1
ADVISEState it, but lead with the uncertainty
ACTEstablished — may base a recommendation on it
The contract
Bound the moment it connects
Your agent doesn't promise to be careful — it's constrained. On connection it's bound: never claim an insufficient relationship, never invent a number, offer borrowed evidence only as borrowed. The rules ship with the tools, not as a prompt you hope it obeys.
The gate
It asks before it asserts
Before the agent tells an athlete one thing moved another, it calls can_i_claim. Back comes a verdict, an interval, and a license — or DECLINE. It cannot speak past the evidence.
The license
Authority, capped at the evidence
Every finding carries a rung, DECLINE → ACT. The agent acts within its license, never above it. That's not a guideline — it's the API's answer. Alignment you can point to.
The trace
Every claim is on the record
Each thing the agent says traces back to the data, the interval, and the license that allowed it. When an athlete asks why, or your reviewer asks how, the answer already exists.
Put your name on what your agent ships.
07 / Pricing
Pricing.
Per coach, billed monthly. No data resale, no per-athlete surcharge surprises. Cancel anytime.
Solo
Solo
For the independent coach running a tight roster.
$49/mo
- Up to 5 athletes
- Console & signed Audit reports
- All wearable & file integrations
- Email support
Most chosen
Coach
For the full roster, with the engine behind your agent.
$149/mo
- Up to 30 athletes
- Everything in Solo
- Typed API & license queries
- Cohort priors & borrowed evidence
- Priority support
Team
Team
For squads, federations and performance departments.
Custom
- Unlimited athletes & staff
- SSO & role-based access
- On-prem & custom integrations
- Dedicated support
08 / FAQ
Questions, answered honestly.
What does AthDash actually claim?+
Only what the data supports — each claim carries its confidence interval and the caveats that bound it. When the evidence is thin, the engine returns insufficient rather than a guess dressed as a verdict.
Where does the data come from?+
The wearables and tools you already use — intervals.icu, WHOOP, TrainingPeaks, Oura, Garmin, FIT files and Apple Health. We don't sell hardware, and we don't resell your athletes' data.
Is this a medical device?+
No. AthDash is a performance-analysis tool for coaches, not a diagnostic or medical device. It does not diagnose, treat, or prevent any condition, and nothing it returns is medical advice.
How is this different from a readiness score?+
A readiness score is one number with no mechanism behind it. AthDash gives you the drivers of performance, the conditions under which each one holds, and the uncertainty around every estimate — so you know which lever to pull, not just that something moved.
Can I export the numbers?+
Yes. Signed audit reports for the record, and a typed API for everything else. The same estimate, interval and license are available on every surface.
What happens when there isn't enough data?+
Nothing gets claimed. The driver is gated as insufficient and held there until enough evidence accumulates to cross the threshold — at which point it's promoted with its interval attached.
Is AthDash causal or correlational?+
Causal. A correlation only says two things moved together; AthDash estimates whether one actually moved the other, after accounting for the training and confounders that could fake the link.
Can it keep my AI coach from hallucinating?+
That's the point. Every finding carries a license — DECLINE through ACT — and an interval, so your agent acts within the evidence instead of inventing a confident answer. When the data is thin, the answer is DECLINE — not a guess.
What is an effect modifier?+
A condition that changes a relationship's strength or sign — for example, whether HRV has recovered. The same training load can help or hurt depending on it, and AthDash finds the threshold where the effect flips.
Does it replace a coach?+
No. AthDash is decision support: it tells you what the evidence says and how sure it is. The coaching call — and the relationship with the athlete — stays yours.
09 / In plain terms
The vocabulary, without the jargon.
The terms behind the engine, defined the way we'd explain them to an athlete — exact, and no longer than they need to be.
- causal, not correlational
- A correlation says two things moved together. A causal estimate says one moved the other — after accounting for the training that could fake the link.
- within-athlete
- Population studies tell you what's true on average. AthDash works inside one athlete's own history, so the answer is about them — not a cohort they may not resemble.
- effect modifier
- A condition that changes a relationship's strength or sign. The same load can help or hurt depending on, say, whether HRV has recovered. AthDash finds the threshold where it flips.
- confidence interval
- The range the true effect plausibly sits in. A wide interval is a quiet way of saying not sure yet. AthDash ships the interval, never a bare number.
- insufficient
- A first-class answer, not an error. When the sample is too thin, the engine says so — declining to answer is a valid output.
- license
- For an AI coach: how far it's allowed to act on a finding, DECLINE → ACT. The estimate carries its own permission.