The science that used to need a lab, a coach and a spreadsheet — now reading your own runs, remembering how your body responds, and telling you what changed and what to do next. Most apps hand you a dashboard; this hands you an answer.
The best endurance science should no longer be reserved for the sponsored elite. Our mission is to turn lab-grade physiology into clear, trustworthy signals you can act on — the moment you sync your watch.
We don’t just move data around. We build longitudinal intelligence: a living picture of how your body actually responds over weeks and months. We learn your patterns, your breaking points, your breakthroughs — and we don’t forget them. That memory is the edge, and it compounds the longer you train with us.
The science is open, and it should be. The craft is making it hold up on a rainy Tuesday run — turning noisy, imperfect data into a signal you can trust. We do that in the open, alongside the runners who use it, and it’s never quite finished.
Questions about a method, or a paper we should be reading? support@ryun.app
Deterministic before generative
Every number comes from a transparent, deterministic pipeline. The AI explains it — it never invents it.
Peer-reviewed or it doesn’t ship
Each metric traces to published, citable research. When the literature is split, we say so — in the product.
Your baseline, not population norms
Rolling individual baselines and smallest-worthwhile-change bands decide what counts as a real change — not generic charts.
Signal quality before interpretation
Windows that miss artifact standards get flagged and excluded. Clean signal first, confident answer second.
Longitudinal memory by design
We don’t forget. The longer you run with Ryun, the better it understands you.
DFA-α1 0.75 → aerobic · 0.50 → anaerobic
A threshold isn’t a fixed number — it drifts with fitness, fatigue and heat. Read from beat-to-beat data across many sessions, so no single run defines your zones.
Rogers, Giles, Draper, Hoos & Gronwald (2021) · Frontiers in Physiology — Sempere-Ruiz et al. (2024) — Altini (2024)
Field note →lnRMSSD vs 7-day rolling baseline ± SWC
Last night’s HRV means little on its own. What matters is the deviation from your own rolling baseline — the smallest change worth acting on, not a red/green score.
Plews, Laursen, Stanley, Kilding & Buchheit (2013) · Sports Medicine — Esco & Flatt (2014) — Hopkins (2000)
Field note →CTL · ATL · TSB — impulse-response tradition
Fitness and fatigue are the same training seen on two timescales; their balance is your form. The acute:chronic injury ratio, though, is contested — so we read it as one signal, not a diagnosis.
Banister et al. (1975) — Gabbett (2016) · BJSM, with its documented limitations
~80/20 easy–hard · HR zones × DFA-α1
Most “easy” runs aren’t. Classifying a session’s true intensity against your own thresholds exposes the moderate-intensity rut that quietly stalls recreational runners.
Seiler (2010) · IJSPP — Kenneally, Casado & Santos-Concejero (2018) — Casado et al. (2023)
Field note →Pw:HR decoupling · within-session Δα1
Two runners with identical fresh thresholds can diverge completely at hour three. Durability — resistance to drift deep into long efforts — is endurance’s fourth dimension, and its evidence is still being written.
Maunder et al. (2021) — Smyth et al. (2022) vs Hunter & Muniz-Pumares (2025) — Rothschild & Maunder (2025)
Field note →Block vs traditional — what transfers
Periodization has strong evidence in elites and thin evidence in amateurs. For most runners, consistency and a sane progression rate move fitness more than any block structure.
Rønnestad, Hansen & Ellefsen (2012) — Rønnestad et al. (2014) — Issurin (2008)
Field note →VDOT → race-equivalent paces
Race predictions age. Recalibrated continuously from recent efforts and threshold estimates, your paces track the fitness you have now — not last season’s.
Daniels & Gilbert (1979) — Bakken (2020) on threshold-centered practice
Every session runs the same path — from raw beat-to-beat data to a read you can act on. Signals in, understanding out.
Raw input
Gate before interpret
Deterministic, from papers
Against your own baseline
In plain language
What the lab is building right now — shipped, in validation, and on watch.
Everything on this page runs against your beat-to-beat data the moment you connect a watch. Free during the beta.