How Pacewright Works

Pacewright's training engine is a rule-based algorithm. That means it follows a specific set of rules — written by humans, based on published sports science research — to decide what workout to prescribe on any given day.

An algorithm is not artificial intelligence. It doesn't "think," it doesn't "learn" from patterns across users, and it doesn't make decisions that even we can't explain. Given the same inputs (your training history, your goals, your schedule, how you're feeling), it always produces the same output. The rules are deterministic: if X, then Y. Every time.

The science behind these rules comes from decades of published research in exercise physiology, training load management, and sports medicine. We didn't invent any of it — we implemented it. The researchers who did the hard work are cited throughout this page and in every workout explanation inside the app.

Your data is used for one purpose: adapting your training plan. It is never used to train AI or machine learning models, never aggregated into datasets, never shared with third parties, and never used for any purpose other than making your next workout better than a guess. That commitment is in our privacy policy.

Your training plan starts with three things: where you are now (your current fitness, training history, and experience level), where you want to go (your goal — a race, a PT test, general fitness, or just consistent running), and how much time you have (your schedule and available days).

From there, the algorithm builds a weekly structure based on proven training methodologies. The foundation is polarized intensity distribution — research by Stephen Seiler found that elite endurance athletes across sports spend roughly 80% of their training time at low intensity and 20% at moderate-to-high intensity. This distribution consistently outperforms threshold-heavy approaches in controlled studies, and it applies to recreational runners too.

Within that framework, the algorithm distributes different workout types across your week:

• Easy runs — the foundation. These build mitochondria, capillary density, and fat-burning efficiency without the recovery cost of hard running. About 80% of your training volume. Often paired with strides — short 20-second accelerations at the end that maintain leg speed and neuromuscular coordination with virtually no injury risk.
• Long runs — build endurance-specific adaptations like glycogen storage and mental resilience. Capped at a safe percentage of your weekly volume. Long runs can include structural variations: a fast finish (last few miles at marathon pace), pace inserts (blocks of faster running within the long run), or a progressive build (gradually increasing pace throughout).
• Threshold runs — sustained effort at your lactate threshold, roughly the pace you could hold for about 60 minutes in a race. Pushes that threshold higher, extending the pace you can sustain before fatigue sets in. Can be run as a continuous effort or as cruise intervals — threshold-pace segments with brief recovery jogs that allow more total time at threshold.
• Marathon-pace runs — sustained running at your goal race pace. Distinct from threshold work — slower but held for much longer. Trains the exact metabolic and biomechanical demands of race day.
• VO2max intervals — develop your VO2max, the maximum rate of oxygen your body can process. Prescribed in time-based segments (not track distances) so they work anywhere. Can include microintervals — alternating 30 seconds hard with 30 seconds easy, a research-backed format that accumulates more time at VO2max than traditional long intervals.
• Progression runs — start easy and gradually accelerate, finishing the last portion at tempo or faster. Teaches your body to run fast on tired legs and simulates the feeling of a strong race finish.
• Hill repeats — hard efforts running uphill for 60-90 seconds with jog-down recovery. Provides a VO2max-level stimulus with lower impact forces than flat intervals, plus running-specific strength development.
• Hill sprints — very short (8-12 seconds), maximal-effort sprints up a steep hill with full walking recovery. Targets the neuromuscular system — motor unit recruitment and power — rather than the aerobic system. One of the safest forms of speed work because the short duration and steep grade prevent the sustained speeds that cause muscle strains.
• Fartlek — unstructured or semi-structured surges within a continuous run. Reduces training monotony and teaches pace variability. An excellent entry point for speed work.
• Recovery runs — active recovery that promotes blood flow without adding meaningful training stress.
• Time trials — periodic all-out efforts over a fixed distance to assess current fitness and calibrate training paces. These are assessments, not regular workouts.

If you're training for a specific event, the algorithm organizes your training into periodization phases — a base-building phase, a sharpening phase, and a taper leading into race day. The structure follows the 3:1 pattern (three weeks of building load, one recovery week) for most runners, or 2:1 for beginners and older runners who benefit from more frequent recovery.

If you're running without a specific event — just staying fit, building consistency, or improving gradually — the algorithm builds rolling weekly plans with no countdown and no periodization pressure. Every feature works the same way regardless of whether you have a race on the calendar.

The single most important thing the algorithm does is prevent injuries. Research consistently shows that the #1 predictor of running injuries is rapid increases in training load — doing too much, too soon. Pacewright uses several layers of protection.

Training load tracking

Every workout has a training load — a single number calculated from how long you ran and how hard it felt (your RPE, or Rate of Perceived Exertion, on a 1-10 scale). This method, called session RPE, was developed by researcher Carl Foster and has been validated extensively. It works with or without a GPS watch or heart rate monitor.

The Acute:Chronic Workload Ratio (ACWR)

The algorithm continuously compares your recent training (last 7 days) to your longer-term average (last 28 days) using a ratio called the ACWR. This concept, developed by Tim Gabbett and validated across multiple sports, tells us whether your current training load is appropriate for your fitness.

• Ratio between 0.8 and 1.3: Safe zone. Your recent training is proportional to what your body is adapted to. Proceed as planned.
• Ratio above 1.3: Caution. You're doing significantly more than your body is used to. The algorithm reduces upcoming workout intensity and explains why.
• Ratio above 1.5: Intervention. The algorithm actively scales back your plan to bring you into a safe range. This is the zone where injury risk spikes.

ACWR is a guardrail, not a crystal ball. It doesn't predict injuries — it identifies training patterns that research has associated with elevated risk, and it keeps you away from them.

Volume increase caps

Regardless of ACWR, the algorithm limits how much your weekly mileage can increase from week to week. The caps are stricter at lower mileage (where connective tissue is still adapting) and more permissive at higher mileage (where you have a stronger structural base). If you've been stable at your current volume for three or more weeks, you earn additional headroom for a safe increase.

Single-session spike protection

No individual run will ever be prescribed at more than 110% of your longest run in the last 30 days. Even if your weekly load is fine, a single abnormally long or hard session can cause damage. This cap prevents it.

Recovery spacing

Hard workouts (tempo runs, intervals, long runs) require recovery time before the next hard effort. The algorithm enforces minimum recovery periods between these sessions — typically 48-72 hours depending on the workout type. You'll never see two hard days back to back.

These safety systems cannot be overridden by any setting. You can adjust your schedule, your goals, and your preferences — but Pacewright will never prescribe a workout it calculates to be unsafe.

A static training plan marches forward regardless of what's happening in your life. Pacewright doesn't. The algorithm recalculates your plan based on what you actually do — not just what was scheduled.

When you miss a workout

Missing a run changes your training load numbers. The algorithm recalculates your ACWR and adjusts upcoming workouts so you don't spike your load by jumping back to the original plan. There's no "catch up" — the plan adapts forward from where you are now. If you optionally tell us why you missed (tired, busy, sore), the algorithm responds differently: "tired" is a training signal that may indicate fatigue, while "busy" is a life signal that doesn't change anything about your fitness.

When you have a great week

If you're consistently completing workouts and your RPE reports suggest the effort is manageable, the algorithm gradually increases your training load — always within the volume caps and ACWR limits. Your plan gets harder at a rate your body can handle.

When you report an injury or pain

The algorithm tracks injury reports by body part, severity, and frequency. A single mild report may reduce intensity for a few days. Recurring reports in the same area trigger a more conservative response. Pacewright never tells you to run through pain — if something hurts, the plan adjusts, and we recommend seeing a medical professional.

When life gets busy

You can tell Pacewright ahead of time that you have a light week coming up. The algorithm proactively scales your plan down so you can complete what's scheduled instead of falling behind. Daily flags for sleep quality, energy level, and stress also influence workout selection — if you slept poorly and feel low energy, today's tempo run might become an easy run.

When the weather changes

Running in heat costs more effort — research shows performance declines measurably above 55°F dew point. Pacewright automatically adjusts your pace targets based on current weather conditions so that a "7:30 pace" on a cool day and a "7:48 pace" on a hot day represent the same physiological effort. This adjustment is free for all users because it's a safety feature, not a premium one.

Pacewright doesn't rely on a single number to assess fitness. It watches several signals over time:

• Pace at a given RPE: Can you run faster while reporting the same effort level? That's improving running economy.
• Workout completion rate: Are you finishing workouts as prescribed, or consistently cutting them short? Patterns here tell the algorithm whether the load is appropriate.
• Training load trends: Is your chronic (28-day) load trending upward steadily? That indicates safe, sustainable progress.
• Heart rate efficiency (if you use a monitor): Running faster at the same heart rate is a clear sign of improving aerobic fitness. The algorithm tracks this on easy runs over 4-6 week windows.

When these signals consistently point in the same direction, the algorithm adjusts your training paces and race predictions. When they conflict — pace improving but RPE rising, for example — the algorithm holds steady and flags the discrepancy for you to consider.

You choose how you want your workouts communicated. Pacewright supports three training metrics:

• RPE (Rate of Perceived Exertion) — the default. A 1-10 scale of how hard a workout feels. Validated by decades of research, works without any equipment. "Run at RPE 4" means conversational effort. "RPE 8" means you can only speak in short phrases.
• Pace — target times per mile or kilometer, calibrated to your current fitness. Best for runners who know their paces and like concrete numbers.
• Heart rate — target heart rate zones. Good for enforcing easy effort (Zone 2 running prevents the common mistake of running too fast on easy days). Requires a heart rate monitor — chest straps are more accurate than wrist-based optical sensors.

Regardless of which metric you choose, RPE is always collected after every workout. It's the universal ground truth that works whether you have a $500 watch or no device at all. The algorithm uses session RPE (how hard it felt × how long it lasted) as its primary load tracking method.

You can also choose between duration-based ("run for 45 minutes") or distance-based ("run 5 miles") workout structures. Duration is the default because it normalizes effort across different fitness levels — a 45-minute easy run is appropriate whether you cover 4 miles or 6.

When you have enough training data, Pacewright can estimate finish times for standard race distances. These predictions use an ensemble approach — multiple established models are run independently, and the results are combined to produce a prediction range.

The models include Peter Riegel's distance-time relationship (personalized to your data), Jack Daniels' VDOT equivalence tables, and critical speed modeling for shorter distances. Each has strengths at different distances, which is why we use several rather than relying on one.

Predictions are always shown as ranges, never single numbers. The width of the range reflects our confidence — predictions based on recent race data over a similar distance are narrow; predictions based on old training data extrapolated to a much longer distance are wide. The range widens automatically when data is stale, the distance ratio is large, or data quality is low.

Environmental conditions (heat, altitude) are factored in when available. A sea-level runner racing at 6,000 feet will see adjusted predictions that account for the roughly 9% performance impact of altitude.

Pacewright offers three coaching personalities: Sarge (direct, no-nonsense), Sage (warm, encouraging), and Quinn (data-forward, precise). You can switch between them anytime in Settings.

It's important to understand what these are and what they aren't. The coaching personalities are different ways of presenting the same information. They change the tone, the wording, and the communication style — but they do not change the algorithm. Your workout prescription, your training load calculations, your safety guardrails, and your plan structure are identical regardless of which coach you choose.

If Sarge says "Tempo run: 3 miles at 8:00 pace. Get after it." and Sage says "Today's a tempo run — 3 miles at 8:00 pace. You've earned this one." and Quinn says "Tempo: 3.0 mi @ 8:00/mi. Load: 42 AU." — the workout is the same. The math is the same. The only difference is how it's communicated to you.

We offer this because people respond to different communication styles. Some runners want to be pushed. Some want to be encouraged. Some just want the numbers. The algorithm doesn't have a personality — but the experience of using it can.

Pacewright's algorithm is built on published, peer-reviewed research. Here are the key principles and the researchers whose work makes them possible:

• Polarized training distribution — Stephen Seiler's research showing that ~80% easy / ~20% hard intensity distribution produces better outcomes than threshold-heavy training.
• Acute:Chronic Workload Ratio — Tim Gabbett's work on training load management and injury prevention through monitoring the ratio of recent-to-chronic training load.
• VDOT and training paces — Jack Daniels' running formula for calibrating workout intensities based on current race fitness.
• Impulse-response model — Eric Banister's mathematical model of how fitness and fatigue respond to training over time.
• Session RPE — Carl Foster's method of quantifying training load using perceived exertion multiplied by duration, validated as an effective alternative to heart rate-based monitoring.
• Rate of Perceived Exertion — Gunnar Borg's foundational work on subjective effort measurement, the basis for RPE-guided training.
• Tapering research — Iñigo Mujika and Sabino Padilla's work on pre-competition volume reduction for performance optimization.
• Environmental performance — Research on heat, humidity, altitude, and their measurable effects on running performance and safety.

None of this research is proprietary to Pacewright. It's in textbooks, peer-reviewed journals, and university courses around the world. We didn't discover these principles — we put them together into a tool that makes them accessible to everyday runners and explains them in plain language.

Our FAQ includes specific citations with journal names, authors, and publication years for every scientific claim we make. If you want to verify anything, the sources are there.

Transparency means being honest about limitations, not just capabilities.

• It is not artificial intelligence. The algorithm does not learn, does not find its own patterns, and does not make decisions that can't be explained. It follows rules written by humans based on published research.
• It does not use your data for anything except your plan. No model training, no aggregate analysis, no data sharing. Your data exists to serve you.
• It does not provide medical advice. Pacewright is a training tool, not a healthcare provider. If you're injured, see a medical professional. The algorithm can adjust your training load around an injury, but it cannot diagnose or treat anything.
• It does not guarantee results. The algorithm applies the best available training science to your situation, but running outcomes depend on factors beyond any algorithm's control — genetics, nutrition, sleep, stress, and consistency.
• It is not secret. The principles behind every decision are explained in the app, on this page, and in our FAQ. Nothing outside the specific implementation code is proprietary. Any app claiming to be science-based should be willing to show how it works. We are.