Intervals.icu uses different methods depending on what data is available: *If you have power data (cycling):* - Calculates *TSS* (Training Stress Score) same as TrainingPeaks: `TSS = (seconds × NP × IF) / (FTP × 3600) × 100` - NP = normalized power, IF = intensity factor (NP/FTP) - "Not moving" time is excluded (paused segments don't count) *If you only have heart rate (running, no power):* Two models you can choose between: 1. *Time in HR zones* — weights time spent in higher zones more heavily 2. *Average HR* — maps avg HR to a load-per-hour estimate, calibrated against your previous power+HR rides Power always takes priority when available; HR is secondary. *For planned/library workouts (before you do them):* This is where it gets tricky. Estimated load comes from the structured workout grammar — intervals.icu parses the step targets (zone, % FTP, % pace, duration) and computes expected NP/pace, then runs the TSS formula against it. That's why workouts without proper `workout_doc` structure show no load estimate — there's nothing to compute from. The key levers: *duration* and *intensity relative to your threshold*. Longer or harder → higher load, roughly as you'd expect. Source: [forum.intervals.icu](https://forum.intervals.icu/t/how-is-calculated-training-load/1661)

*Intensity Factor (IF)* = NP / FTP It's a ratio expressing how hard a workout was relative to your threshold: - IF = 1.0 → you rode exactly at FTP the whole time - IF = 0.75 → moderate aerobic (typical Z2 ride) - IF > 1.0 → harder than FTP (short, punchy efforts) Typical ranges: - Recovery ride: ~0.55–0.65 - Z2 endurance: ~0.65–0.75 - Tempo: ~0.76–0.90 - Sweet spot: ~0.84–0.97 - FTP test/race: ~1.0–1.05 It's a cleaner signal than average power because it uses *normalized power* (NP), which accounts for variability — a ride with hard surges and coasting registers higher than a flat steady ride at the same average wattage.

Two pretty different philosophies: *TSS (TrainingPeaks/intervals.icu)* - Power-first: `(seconds × NP × IF) / (FTP × 3600) × 100` - HR fallback when no power - Anchored to your FTP — 1 hour at FTP = 100 TSS, always - Deterministic and reproducible given the same power file *Garmin (Firstbeat EPOC-based)* - Estimates *EPOC* (excess post-exercise oxygen consumption) from HR data using their proprietary Firstbeat algorithm - EPOC is a physiological proxy for how hard your body worked, not just mechanical output - Incorporates HR dynamics (how quickly HR rises/falls, HR drift) to infer actual metabolic stress - Expressed as "Training Load" points, which don't have a clean mathematical anchor *Key differences:* | | TSS | Garmin | |---|---|---| | Primary signal | Power (or HR) | HR only | | Anchor | FTP (objective) | Estimated VO2max (can drift) | | Transparency | Formula is public | Black box | | Cross-sport | Works well | Also cross-sport via HR | | Variability | Low (power is stable) | Higher (HR varies with heat, fatigue, caffeine) | *Practical implication for you specifically:* since your HRV and HR don't reliably reflect fatigue, Garmin's EPOC model is probably noisier for you than TSS is. TSS from power data doesn't care that your HR stays suspiciously low — it just sees watts. That's another reason to lean on the CTL/ATL/TSB math over Garmin's readiness scores. Source: [Firstbeat whitepaper](https://www.firstbeat.com/en/science-and-physiology/white-papers/)

Yes — *running power* is a thing, and the main options are: *Stryd footpod (~$230)* - Clips to your shoe lace - Measures ground contact time, cadence, vertical oscillation → estimates mechanical power - Also accounts for wind resistance - Widely considered the gold standard for running power - Pairs with Garmin, streams power into your activity file - intervals.icu can then compute proper TSS from it *Garmin native running power* - Built into newer Garmin watches (FR965, Fenix 7+, etc) - No extra hardware needed - But it's wrist-based, uses accelerometer, and runs ~30% *higher* than Stryd — the two are not interchangeable - Afaict it's less trusted by the community for structured training *The caveat:* running power is still somewhat the "Wild West" — unlike cycling power which measures actual torque at the crank, running power is always *estimated* from motion metrics. The numbers mean less in absolute terms (Stryd's 250W ≠ cycling's 250W) and more as a *self-consistent* relative signal. As long as you anchor your training zones to Stryd power specifically, it works well. The TSS calculation in intervals.icu will work the same way once Stryd data flows in — just needs a running FTP (called rFTPw in Stryd's ecosystem). Do you have one of the newer Garmin watches? If it has native running power already, could be worth testing before spending $230 on Stryd. Sources: [Reddit comparison](https://www.reddit.com/r/strydrunning/comments/16d6aem/garmin_running_power_vs_stryd/) | [Stryd 10k km review](https://the5krunner.com/2023/11/16/stryd-review-discount-duo/)

Fenix 7x does have native running power built in — but you may need to verify it's actually recording to your activities. *Check if it's enabled:* On the watch: Settings → Activities → Run → Data Screens → add "Running Power" as a field. If it shows a number while running, it's active and will be logged. *Then in intervals.icu:* Go to Settings → Sports → Running → set your *rFTPw* (running FTP in watts). Without this anchor, intervals.icu can't compute proper TSS from the power data — it'll fall back to HR. *Caveat to keep in mind:* Garmin running power reads ~20-30% higher than Stryd. So don't compare your running watts to anyone using Stryd, and don't cross-reference against cycling watts. But *internally* it'll be self-consistent, and that's all you need for TSS to work. Worth checking your recent runs on intervals.icu to see if power is already in the data — if you see a power curve on any run activity, you're good.

A few ways to estimate rFTPw: *Method 1: Best 20-min power test* Run hard for 20 min (flat surface, all-out sustainable effort), then: `rFTPw = avg_power × 0.95` Classic FTP estimation — same as cycling. Most accurate but requires a proper effort. *Method 2: From existing data (quick and dirty)* In intervals.icu, go to your Fitness tab → click any run → look at the *Power Curve*. Find your best ~20-min power from recent runs. Multiply by 0.95. If you mostly run short (you mentioned ~5km currently), your 20-min best from treadmill runs is probably your best proxy. *Method 3: Estimate from pace + VO2max* Based on your VO2max (~43 from Garmin) and running at Z3/tempo, we can back into it. Typical running power at threshold is roughly: - rFTPw ≈ VO2max (ml/kg/min) × body weight (kg) × ~0.074 - At 227 lbs (103 kg), VO2max 43: `43 × 103 × 0.074 ≈ 328W` That's a rough ballpark — Garmin's native power tends to read high, so your actual number is probably *270–350W*. *Recommendation:* pull your power curve from a recent hard treadmill run in intervals.icu and look at the 20-min mark. That'll give you a real data point to anchor from. What pace are you typically running those runs at?

Got the API key, pulling your recent runs with power data now...

Got it. Here's what the data shows across your last 9 runs with power: *Best efforts across all runs:* | Duration | Best Power | |---|---| | 1 min | 480W | | 5 min | 395W | | 10 min | 371W | | 20 min | 353W | | 30 min | 347W | *Estimated rFTPw:* `353W × 0.95 = ~336W` This came from your Apr 10 Easy Run 30m, which had the best 20-min power at 353W. A few caveats worth noting: - These are all *easy runs* — none were hard efforts. Your 20-min best from an easy run will underestimate your true FTP. The real number is likely higher. - Garmin native power reads ~20-30% high vs Stryd, so *don't compare this to cycling watts or anyone using Stryd*. - The ~330–336W estimate is internally consistent though — use it to anchor your running zones in intervals.icu and TSS will start working for runs. I'd recommend setting rFTPw to *336W* for now, and refining it after you do a proper tempo or hard effort run. A 20-min time trial on the treadmill at max sustainable pace would nail it down precisely.

Nice — this confirms the estimate was close. A few things worth noting from this chart: *Recent fitness (42/84 days):* - Best 20m: 343W → 343 × 0.95 = *326W* by the simple formula - But intervals.icu's *eFTP model* says *334W* — it fits the whole power curve (W' / critical power model), which is more accurate than just the 20m point - The 336W I suggested is essentially spot-on with eFTP *All-time curve (orange):* - Best 20m was *407W* on Jan 27, 2025 → eFTP 423W - That's a big gap from current 334W — you were significantly fitter (or running harder efforts) in early 2025 *Recommendation:* set rFTPw to *334W* (the eFTP value from recent data) rather than the 336W estimate — it's a better-fit number from the full power curve. Trivial difference but cleaner. The eFTP method also means you don't *need* a formal time trial — intervals.icu will auto-update it as you accumulate more runs. Worth checking back after a harder effort to see if it shifts up.

Yes — this is a clean math problem. Let me model it. *The decay model:* CTL and ATL are exponentially weighted moving averages. The standard formulation: ``` CTL_new = CTL_old × (1 - 1/42) + load × (1/42) ATL_new = ATL_old × (1 - 1/7) + load × (1/7) ``` Form (TSB) = CTL - ATL *The steady-state question:* If you train at constant daily load `L`, both eventually converge: - `CTL_ss = L` (by definition — 42-day avg of constant L = L) - `ATL_ss = L` - `TSB_ss = 0` So *constant daily load always converges to TSB = 0*. You can't maintain negative form in steady state — the body adapts. *The practical model:* To stay in TSB -10 to -30, you need ATL persistently above CTL. That means you need to be in an *ongoing build* — training load increasing over time, so ATL (fast-reacting) is always chasing a CTL that hasn't caught up yet. The question becomes: *how fast does CTL need to be growing?* If CTL is growing at rate `r` CTL units/week: - ATL leads CTL by roughly `r × (42 - 7) / 7 ≈ 5r` in the transition zone - To maintain TSB ≈ -20: you need ATL - CTL ≈ 20, so you need CTL growing at ~4 CTL/week *Let me compute this precisely with your current numbers and model it out — what's your current CTL and ATL?* I can pull them from the API and run the simulation.

Modeled this out fully. Here's the complete picture: *The fundamental math* CTL/ATL are EWMAs with decay constants 1/42 and 1/7. At any *constant* periodic load, both converge to the same value → TSB → 0. You *cannot* maintain -10 to -30 at steady state. It's only achievable during a build phase. The key relationship: during a ramp where weekly load increases by `r` units/week: ``` TSB ≈ -r × 5 ``` (ATL is 7-day lag, CTL is 42-day lag — the gap is 35 days ÷ 7 ≈ 5 weeks of offset) So to maintain TSB = -20, you need load increasing ~4 units/week total. *Your current numbers* CTL=21.5, ATL=39.5, TSB=-18 — you're right in the zone *now*, but ATL decays fast. After 2 rest days ATL drops from 39 → ~29. By end of a full rest week: ATL≈13, TSB≈+5 (completely out of range). *The practical model: 5x/week, increasing load* | Week | Load/session | Mid-week TSB | Notes | |------|-------------|-------------|-------| | 1 | 60,60,50,60,60 | -20 to -24 | ✓ | | 2 | 65,65,55,65,65 | -15 to -22 | ✓ | | 3 | 70,70,60,70,70 | -13 to -22 | ✓ | | 4 | Recovery (40s) | -4 to -9 | ✗ (intentional) | *19 out of 28 days stay in range* with this schedule. Weekend rest days and recovery week drift above -10 — that's normal and expected. *Key takeaways* 1. *-10 to -30 is a BUILD PHASE metric, not a steady state.* Trying to live there permanently is a category error. It means "you're applying training stimulus." 2. *Rest days will always push you above -10.* ATL's 7-day time constant means two rest days drops TSB ~6-10 points. You'll always oscillate in/out. 3. *The lever is weekly total load ramp rate.* Adding 4 load-units/week to your total maintains TSB ≈ -20. For 5 sessions/week that's just +0.8 load/session/week — tiny increments. 4. *Recovery weeks are outside the zone by design.* TSB rises to 0 or positive → that's when adaptation happens. Then the next build block reloads the fatigue. 5. *At your current CTL=22, you need 60-70 load sessions (not 45) to stay deep in the zone.* 4

Created a new skill at `~/skills/training-plan.md`. It covers: - *Weekly template* — Mon/Tue run+strength, Wed/Fri fasted walk/cycling + OMAD fish&veggies, Sat key session (long run), Sun rest - *Long run progression* — 8-week example table building from 40m → 65m with 4th-week recovery, ~30% of weekly volume rule - *TSB/load targets* — how to read CTL/ATL to size sessions, recovery week triggers - *Intensity distribution* — 80/20 polarized, which library workouts to use - *Weekly planning workflow* — step-by-step checklist for when we plan each week - *Nutrition context* — training day vs fasted day guidance I left the `intervals-icu.md` skill intact (it has the API calls and formulas). This new skill is the "what to do" layer; that one is the "how to do it technically" layer. One thing worth flagging: your current run library (id:2 Long Easy) is 90 min. For the early progression weeks (40-50 min long runs), we'll just schedule ad-hoc runs or shorten the description. Want me to add a shorter long-run template to the library?

On it — pulling your current CTL/ATL and next week's scheduled workouts now.

Here's the assessment: *Current fitness*: CTL=21.5, ATL=39.5, TSB=-18 (in range ✓) *Next week as scheduled:* | Day | Scheduled | Plan says | Status | |-----|-----------|-----------|--------| | Mon Apr 20 | Easy Run 30m | Run + strength | ✓ (no strength noted) | | Tue Apr 21 | Easy Run 30m | Run + strength | ✓ (no strength noted) | | Wed Apr 22 | *Z2 Ride 90m* | Fasted walk/easy cycling | ❌ | | Thu Apr 23 | Easy Run 30m | Run + strength | ✓ (no strength noted) | | Fri Apr 24 | *Z2 Ride 90m* | Fasted walk/easy cycling | ❌ | | Sat Apr 25 | Long Easy 50m | Key session: long run | ✓ | | Sun Apr 26 | (nothing) | Rest | ✓ | | Mon Apr 27 | Easy Run 30m | *This is next-next week* | — | *Issues:* 1. *Wed/Fri are wrong* — Z2 Ride 90m is way too much for fasted days. That's 90 min of real work with no fuel. Should be replaced with 60m easy walk or very easy Z1 spinning (no targets). 2. *No strength blocks* — Mon/Tue/Thu should have KB Simple (44m) after the run. Not critical but it's part of the template. 3. *Load check*: 3 × Easy Run 30m ≈ 3×32 = 96 load. Long Run 50m ≈ 52. Two 90m rides ≈ 2×50 = 100. Total ~248. If we swap rides for walks (~2×8 = 16), total drops to ~164. ATL will fall from 39 → ~25 by week's end (coming off vacation), TSB will rise to ~-3. That's fine — this is your first week back, recovery before build is appropriate. *Recommendations:* - Replace Wed/Fri Z2 Ride 90m → "Fasted Walk 60m" (type: Walk, 60m, Z1 HR only) - Add KB Simple after Mon/Tue/Thu runs (can schedule or just do it ad-hoc) - Sat 50m long run is perfect for week 1 Want me to update Wed and Fri now?

Good questions. Let me address each: *1) Jump rope* Yes, adding it in — perfect fit for running economy (ankle stiffness + elastic return). Best placed as a warmup or between strength sets, not a standalone block. 10-15 min is plenty. *2) The "2m KB swings 10 reps" format* Yes — the Garmin timer format means "2 minute *block*". You do the 10 swings (~12 sec), then the rest of the 2m is *rest*. It's there so the watch beeps you to the next movement. Kind of a clunky format but it's how intervals.icu structured workouts work for weight training. We'll clean this up. *3) Running-focused strength split* Here's what I'd propose — a 2-day rotating split, always done *after* runs: *Day A — Posterior Chain + Core* (Monday / Tuesday) Focus: glute-ham strength, hip stability, running posture - Jump rope warmup — 3 min - KB swings — 3×15 - Single-leg RDL — 3×8 each side (hip hinge, hamstring load) - KB goblet squat — 2×10 - Copenhagen plank — 3×20s each side (adductor/core) - Dead bug — 3×8 each side (deep core, anti-extension) - Pallof press — 2×10 each side *Day B — Single-leg + Core* (Thursday) Focus: running-specific stability, hip/knee control, rotational core - Jump rope — 3 min - Bulgarian split squat — 3×8 each side - KB clean & press — 3×5 each side - Step-up with knee drive — 2×10 each side - Side plank with hip lift — 3×20s each side - Bird dog — 3×8 each side - Suitcase carry — 2×30m each side *Total time: ~30-35 min* Rotation: Mon=A, Tue=B, Thu=A (or B, alternating). You never do the same split two days running. Fasted Wed/Fri: re your question — treadmill walk + sauna at Glenmoor is honestly the better call on rainy days. More enjoyable, and you get the sauna/heat adaptation benefit which has real science behind it for endurance. Easy spin is fine too if you prefer the convenience. Either works. Want me to update the `training-plan.md` skill and rebuild the KB library workouts to match this split?

Honest answer: no, the plan skews heavily posterior/functional. It'll make you a better runner but won't do much for the mirror. What's missing: - *Chest* — zero push/press work. No chest, no triceps. - *Shoulders* — clean & press hits some, but no lateral raises or front work. Shoulders are 80% of the summer look. - *Biceps* — nothing. Carries hit forearms but not bis. - *Abs (visible)* — the core work is all stability (dead bug, Pallof, plank). Good for running, not for definition. Need some flexion work too. *What I'd add — a "beach finisher" (~10 min) at the end of Day B:* - KB floor press or push-up variation — 3×10 - KB bent-over row — 3×8 each side (hits bis + upper back) - KB lateral raise — 2×12 (most visible shoulder muscle) - Ab wheel or hollow body hold — 3×10 / 3×20s That's it. Keeps Day B as the "vanity day" without adding a whole new session. Day A stays performance-focused. The KB is actually solid for this — floor press, rows, and lateral raises all work fine with a single bell. Only thing it can't really do is isolate biceps well, but rows get them secondarily and honestly that's fine unless you're going full bro mode. Want me to fold this into the training-plan.md and update Day B accordingly?

Done. `training-plan.md` updated with: - Full strength split (Day A / Day B) with the correct rotation schedule - Day B now ends with the beach finisher: *push-ups* (not floor press), bent-over row, lateral raise, ab wheel/hollow body - Jump rope as standard 3-min warmup on both days - Note about the Garmin timer format (do reps, rest the remainder) - Treadmill + sauna noted as the rainy-day fasted option The plan is now complete. Ready to use it for weekly scheduling whenever you want.

On it — pulling current fitness + existing calendar for next week now.

Welcome back! Here's a quick status snapshot: *🔵 Inbox (needs your attention)* *36 tasks in Review* — mostly agentd/CLI unification and health dashboard work. P1 highlights: - `t-162` NeedsHelp — Execution details overwritten on retries (P1) - `t-627` NeedsHelp — Task system event/notification layer (P2) - `t-265.8` NeedsHelp — Analytics and funnel tracking (P3) *Top Review items to clear:* - `t-763` Unify agentd CLI - `t-769` Ava: simplify slash commands + provider login - `t-775` Finish agentd migration to Omni/Agent - `t-776` agentd daemon can't send to persistent agents after restart - `t-788` Fix Ava dropping final response in group chats - `t-789` Fix ava.service permission denied writing logs *🏃 Training* Your plan starts Monday Apr 20 (tomorrow). Week targets: Ride 218 load, Run 43 load. Want me to program this week's workouts to intervals.icu? *📬 Email* Nothing flagged unread in the last 10 days (maildir may just be quiet). --- What would you like to tackle first — clearing the Review queue, training plan, or something else?

Here's your week at a glance (workouts at Glenmoor given the cold/rain): *Sun 4/19* - 10am-1pm Church - 5:30-8pm Carpathians Mens Club meeting *Mon 4/20* — busy day - 6am Run + 6:45am Strength A @ Glenmoor - 9am-12pm Chyann deep cleaning - 9:30am Gymnastics (Solo Spouts) - 1pm Engineering Sync - 2pm Parasail All Hands (April) - 6pm Bill Morgan — kitchen walkthrough *Tue 4/21* - 6am Run + 6:45am Strength B @ Glenmoor - 10am-1pm Watch babies 👶 - 11am Kate massage *Wed 4/22* - 6am Fasted walk - 11am Your massage 💆 *Thu 4/23* - 6am Run + 6:45am Strength A @ Glenmoor - 1pm Engineering Sync - 2pm Proj-TTS Launch Team Sync *Fri 4/24* - 6am Fasted walk *Sat 4/25* - Kate's parents arrive (multi-day) --- Monday is the crunch day — back-to-back from 6am through evening. A few things worth prepping: - *All Hands* — anything you need to have ready or present? - *Bill Morgan kitchen visit* — do you have a list of what to show/ask him? - *Glenmoor bag* — packed and ready to grab Monday morning? What do you want to prep first?

Got it. Here's the full prep list organized by area: --- *💼 Work — Mon morning priorities* 1. Review `task inbox` — 36 tasks in Review, some NeedsHelp. Clear the queue before eng sync. 2. Key open work you'll likely report at Thu sync: agentd migration, ava fixes (t-775, t-776, t-788, t-789) 3. Nothing needed for All Hands — CEO is driving, just show up --- *🏃 Training — this week* *Strength A (Mon/Thu):* - Jump rope 3 min - KB swings 3×15 - Single-leg RDL 3×8/side - KB goblet squat 2×10 - Copenhagen plank 3×20s/side - Dead bug 3×8/side - Pallof press 2×10/side *Strength B (Tue):* - Jump rope 3 min - Bulgarian split squat 3×8/side - KB clean & press 3×5/side - Step-up w/ knee drive 2×10/side - Side plank w/ hip lift 3×20s/side - Bird dog 3×8/side - Suitcase carry 2×30m/side - _Beach finisher:_ push-ups 3×10, KB bent-over row 3×8/side, KB lateral raise 2×12, ab wheel 3×10 *Night-before checklist:* lay out gym clothes + pack any extras (headphones, water bottle, towel for sauna). Do this Sun/Mon/Wed/Thu nights. *Sunday:* grocery run — load up on protein (fish, eggs, chicken), veggies, whatever you need for fasted-day OMAD meals Wed/Fri. --- *⛪ Carpathians Fundraiser (Kentucky Derby Dinner)* From the Jan emails, your role is *Silent Auction + Chinese Raffle* — projected $1,450 income target. Per the committee assignments: - Secure donated auction items (each Carpathian supposed to bring ≥1) - Recruit volunteers for setup/auction management - Create bid sheets for silent auction items - Confirm all items are secured and inventoried Most recent email (Feb 18): flyer was being finalized. No update since then, so tonight's meeting (5:30pm) is your chance to get a status check from Ben Stoffer. *Questions to ask tonight:* - What items have been secured so far? Who has them? - Is there a date for the event? (Need to know auction sheet deadline) - Has the Chinese raffle basket approach been confirmed? --- *Tonight (Sun)* Church 10am-1pm → groceries → Carpathian