half-marathon-pfitzinger
Skill5 starsChanged 4 days agoMITView on GitHub
Plan, review and adjust half marathon training in the style of Pete Pfitzinger — lactate threshold as the cornerstone, medium-long runs midweek, mesocycle structure, disciplined easy running. Use when the user asks for a half marathon plan or session, wants a threshold workout built, or wants their recent training judged against this approach.
In this repositoryOne other file it gives its agents.
- deploy-checkSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 32
Every zone below depends on it.
- Very specific: line 106
Use create_workout, show the structure, wait for confirmation, then schedule_workout.
- Distinctive: line 149
Everything in one narrow band: Check the spread of paces and heart rates across a month.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Half marathon training, Pfitzinger style
Pete Pfitzinger's approach (Faster Road Racing, Advanced Marathoning) is built on one idea: for races from 15K to the marathon, lactate threshold is the single most trainable determinant of performance. Everything else exists to support it or to stop it stagnating.
This skill encodes that methodology. It does not reproduce his published week-by-week schedules — those are his copyrighted work, and anyone serious should buy the book. What follows is the reasoning, applied to this runner's actual data.
Before prescribing anything
Pull the evidence first:
get_profile— PBs and VO2 maxget_activitieswith a 6-8 week window — current volume, the real shape of the weekget_activity_detailson recent quality sessions — did they hold pace, did heart rate drift
Then establish three things, asking only for what the data can't tell you:
- Current half marathon fitness, not their PB. A PB from two years ago sets no paces today. If there's no recent race, estimate from a recent hard effort or a tempo run and say plainly that it's an estimate.
- Max heart rate, and whether it's measured or Garmin's age-based guess. Every zone below depends on it; if it's a guess, lead with pace instead.
- Weeks until the race, and whether there are tune-up races planned.
The paces
Pfitzinger defines effort by percentage of maximum heart rate, cross-checked against race paces. Use both — pace when it's flat and calm, heart rate when it's hilly, hot, or the runner is tired.
| Session | % max HR | Pace reference |
| Recovery | below 76% | Genuinely slow. Slower than feels dignified. |
| General aerobic | 70-81% | Comfortable, conversational |
| Long / medium-long | 74-84% | Steady, finishing stronger than started |
| Lactate threshold | 82-91% | 15K to half marathon race pace |
| VO2 max | 93-95% | 3K-5K race pace |
Lactate threshold pace is roughly what the runner could hold for an hour flat out. For most people that is close to their current 15K pace and slightly faster than half marathon pace.
The sessions
Lactate threshold runs. The cornerstone. Either continuous or in long intervals, always after a proper warm-up.
- Continuous: 20-40 minutes at LT pace, progressing over a block. Start where the runner is, not at 40 minutes.
- Intervals: 2-4 reps of 10-15 minutes at LT pace with 2-3 minutes jogging. Useful early in a block, or for someone who cannot yet hold 30 minutes.
The discipline that matters: LT runs are controlled. Running them faster turns them into a different session and blunts the next one.
Medium-long runs. The Pfitzinger signature, and the thing most self-coached runners miss. A midweek run of 90 minutes to two hours, separate from the weekend long run. This is where aerobic development comes from at a volume most people never reach because they only run long once a week.
Long runs. Up to roughly 2 hours for a half marathon build. No further — beyond that the cost outweighs the benefit at this distance. Keep it under about 30% of weekly volume.
VO2 max intervals. 600m to 1600m at 3K-5K pace, 3-5 minutes per rep, with recovery roughly equal to rep duration. Total volume of the fast running kept to about 4-8% of weekly mileage. Introduced late in a build, not early.
Recovery runs. Short and genuinely easy. Their purpose is to let the next quality session happen. If a recovery run is anywhere near general aerobic pace, it is doing the opposite of its job.
Strides. 8-10 × 100m relaxed accelerations after an easy run, twice a week. Cheap, and they protect economy while the training is otherwise slow.
Structuring a block
Pfitzinger organises training into mesocycles, each with a focus, each ending in a recovery week:
- Endurance — volume, general aerobic and medium-long runs, strides. Little or no threshold work yet.
- Lactate threshold + endurance — the heart of the block. Threshold work every 7-10 days while volume holds.
- Race preparation — VO2 max intervals appear, tune-up races, threshold work maintained, volume eases slightly.
- Taper — volume drops sharply, intensity is retained at reduced volume. Two to three weeks for a half marathon.
Volume rises about 10% a week, with every third or fourth week cut by 20-25%. Hard days need two easy days after them, not one — a week with three quality sessions in it is not a Pfitzinger week.
Building the sessions on the watch
Use create_workout, show the structure, wait for confirmation, then
schedule_workout. Pace targets as a range, since the runner is chasing a zone
rather than a number.
A 30-minute continuous threshold run for someone with 3:55/km LT pace:
[{"type": "warmup", "duration_seconds": 900},
{"type": "interval", "duration_seconds": 1800, "pace": ["3:52", "3:58"]},
{"type": "cooldown", "duration_seconds": 600}]Threshold intervals, 3 × 12 minutes:
[{"type": "warmup", "duration_seconds": 900},
{"type": "repeat", "times": 3, "steps": [
{"type": "interval", "duration_seconds": 720, "pace": ["3:52", "3:58"]},
{"type": "recovery", "duration_seconds": 150}]},
{"type": "cooldown", "duration_seconds": 600}]VO2 max session, 5 × 1000m at 5K pace:
[{"type": "warmup", "duration_seconds": 1200},
{"type": "repeat", "times": 5, "steps": [
{"type": "interval", "distance_meters": 1000, "pace": ["3:28", "3:34"]},
{"type": "recovery", "duration_seconds": 180}]},
{"type": "cooldown", "duration_seconds": 600}]What to flag when reviewing someone's training
Pfitzinger's criticisms of typical self-coached training, in order of how often they apply:
- Easy days too hard. The most common fault. If recovery and general aerobic runs sit at the same pace and heart rate, there are no easy days.
- No medium-long run. One long run a week and everything else 45 minutes is the classic plateau shape.
- Threshold work missing or mislabelled. "Tempo" run at 5K effort is a VO2 max session in disguise, and recovers like one.
- Everything in one narrow band. Check the spread of paces and heart rates across a month. A tight cluster means the training is one stimulus repeated.
- No recovery weeks. Volume that only ever rises is volume that ends in injury or stagnation.
- VO2 max work too early. It sharpens something that isn't built yet.
Say which of these the data actually shows. If three weeks isn't enough to call a pattern, say so rather than producing a diagnosis.
Limits
This is a training methodology, not medical advice. Pain that changes gait, persistent elevated resting heart rate, or anything that sounds like an injury is a conversation for a physio, not a plan adjustment. Say so and stop.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

