By Raji Shah · Published 2026-10-23
Running by feel means setting the effort of a run by how hard it feels instead of by the pace or heart rate on your watch. It isn't a step back from data. Sports scientists treat perceived effort as a measurement in its own right: an effort rating for a whole session, multiplied by its minutes, tracks heart-rate-based training load in endurance athletes, and in self-paced running, it appears to help set the pace. Feel also has blind spots, which is why the research points to using it alongside the watch, not instead of it.
Quick answer: perceived effort is a valid way to measure training load, not a guess. Rate a run from 0 to 10 and multiply by its minutes: in the few small runner studies that have tested it, that score tracked heart-rate-based training load. When heat or hills make the same pace harder, runners who pace by effort slow down on their own, though not always by enough to keep the effort level. Feel is least precise for newer runners, especially on easy runs, and caffeine, music and switching between a treadmill and a track can shift it. The practical approach is to rate every run, do some runs with pace hidden, and compare how each run felt with your pace and heart rate afterward.
What RPE and session RPE mean
Perceived effort, measured as a rating of perceived exertion (RPE), is your own judgment of how hard you're working. Most research uses one of two scales that grew out of Gunnar Borg's work: a 6–20 scale common in lab studies, and a 0–10 category-ratio scale (Borg, 1982). Apple Watch, Polar and Strava all ask for a rating from 1 to 10.
A widely studied version is session RPE: one rating for the whole session, multiplied by its length in minutes, which gives a training-load score (Foster et al., 2001). A 45-minute run you rate 4 scores 180. The 0–10 scale used for it labels 0 as rest, 1 very, very easy, 2 easy, 3 moderate, 4 somewhat hard, 5 hard, 7 very hard and 10 maximal, leaving 6, 8 and 9 unlabeled (Haddad et al., 2017).
The simplest field version is the talk test. In a lab study of 16 healthy volunteers, the stage where they first weren't sure they could speak comfortably sat close to their ventilatory threshold, the intensity where breathing starts to climb steeply, and when speaking was uncomfortable they were consistently above it (Persinger et al., 2004). If you can talk in full, comfortable sentences, you're probably below that threshold, where easy runs belong, though a treadmill study of 15 people found even comfortable speech came above it (Quinn & Coons, 2011).
Is perceived effort accurate?
As a measure of how hard your training is, mostly yes, with limits. A 2017 review found 36 studies that tested session RPE and judged it valid and reliable across sports, ages and levels, though none of the studies in its validation table were of distance runners (Haddad et al., 2017). Most of the runner studies that exist point the same way:
- Recreational marathoners: over five months, seven runners' session-RPE load closely tracked a heart-rate-based load, and the two methods showed the same split of easy, moderate and hard training (Manzi et al., 2015), though in road cyclists RPE counted far more training as moderate or hard than heart rate did (Sanders, Myers & Akubat, 2017).
- Adolescent distance runners: in 15 runners over two weeks, ratings tracked heart-rate-based load most closely when given right at the end of the session (Mann et al., 2019).
- Performance: in seven trained runners over 15 weeks, training models built from session RPE, from heart rate and from pace all tracked their 1,500 m time-trial results to a similar degree (Wallace, Slattery & Coutts, 2014).
The limits shape how you should use it. A single rating is noisy: in a lab study of 10 recreational athletes, session-RPE load varied about 28% between repeats of the same session, against 11–16% for heart-rate-based load (Wallace et al., 2014). Across 2,560 people doing lab exercise tests, RPE rose in step with heart rate and blood lactate, but individuals varied widely around the averages (Scherr et al., 2013). And in 22 high school cross-country runners, the session rating on its own barely matched the average heart rate their GPS watches recorded (Garcia & Bazett-Jones, 2022); the closer matches above multiply both measures by the same minutes, which can inflate the agreement. Treat one rating as a rough reading, and weeks of ratings as the real signal.
Why the same pace can feel harder from one day to the next
Researchers separate the work you do, called external load, from your physical and mental response to it, called internal load. For a runner, pace and distance are external load; heart rate and perceived effort are internal load. The same external load can produce a different internal load in different people, and in the same person at different times (Impellizzeri et al., 2023). Studies have measured several of the reasons:
- Heat: running at a fixed intensity in 35 °C, seven active men saw heart rate climb 19% between minutes 15 and 45, and the run became genuinely harder relative to their maximum (Wingo, Stone & Ng, 2020).
- Dehydration: 17 trail runners held a matched moderate pace in the heat after 22 hours of fluid restriction and rated it 16 instead of 13 on the 6–20 scale at the finish, with heart rate about 15 beats per minute higher (Casa et al., 2010).
- Altitude: elite middle-distance runners living at 2,100 m ran their threshold sessions 5.8% slower than at sea level, and the sessions where they held their usual speed felt harder (Sharma et al., 2017).
Heart rate has limits of its own. It drifts upward on long or hot runs even at a steady pace, so a given heart rate stops matching a given pace. It also lags at the start of a hard effort: in a cycling study, it took close to a minute to complete about two-thirds of its rise (Schneider, Wing & Morris, 2002). Our heart-rate zones guide covers using zones, and our optical heart-rate guide covers the sensor side.
How effort steers your pace
One model proposes that the brain combines signals from the body into a sense of effort and uses it to set pace in advance (Tucker, 2009). A competing model proposes that perceived effort itself is what limits endurance (Marcora & Staiano, 2010). Both put perception at the center, but they disagree on how it works, and neither view is settled.
Race data fit that picture. In nine runners who ran both a 7-mile road race and a half marathon, effort rose faster in the shorter race, but plotted against the share of the race completed, the two curves matched (Faulkner, Parfitt & Eston, 2008). Expectations matter too. Well-trained runners held at a fixed treadmill speed rated it harder as soon as they were told to keep going past the 10 minutes they'd expected (Baden et al., 2005). And trained runners who knew a 7% climb was coming at the end of a 3,000 m treadmill trial ran the flat first part 2.4% slower than when they didn't know, without rating it any differently (Brick et al., 2019).
When feel is the better guide
Because perceived effort folds in heat, hills and fatigue, it's most useful when a fixed pace or heart-rate target stops meaning what it usually does:
- Hot days. When 13 recreational runners ran a self-paced 10 km on a treadmill at 35 °C, every one was slower than at 25 °C, by about 14 minutes on average, and it still felt harder; heart rate didn't differ (Rodrigues et al., 2023). A cool-weather pace target doesn't carry over to a hot day; expect a slower pace at a given effort.
- Hills. Eight experienced runners pacing themselves on an outdoor course slowed about 23% on the climbs, yet still worked harder there than on the flat, and those whose pace varied more with the gradient kept their oxygen use steadier (Townshend, Worringham & Stewart, 2010). Our hilly race pacing guide covers how to plan for it.
- The first minute of hard reps. Heart rate is still catching up, so feel or pace is the better live guide for short intervals; our Zone 5 running guide explains the lag.
- Building fatigue. In 19 national-team rowers, session ratings drifted harder during a four-week block that raised low-intensity volume by about half in weeks two and three, while time in each heart-rate zone held steady, and the harder ratings went with rising fatigue scores (Pind et al., 2021). A familiar run that feels unusually hard is worth noting.
When feel can mislead you
- Newer runners, especially on easy runs. Over six weeks of run training in 40 untrained men, those guided by RPE hit their target intensities less precisely than those guided by heart rate, most of all at the easiest target, where the RPE group's error averaged about 16% against 4%. Their fitness gains were similar (Johnson et al., 2017).
- Caffeine. Across 21 placebo-controlled studies, caffeine lowered RPE during exercise by about 5.6% (Doherty & Smith, 2005). Recreational runners asked to hold a fixed effort after caffeine chose slightly faster speeds, and in about a third of trials ran clearly faster with heart rate around 12 beats per minute higher (Green et al., 2016).
- Music. A meta-analysis of 139 studies found that music lowers perceived exertion slightly without a significant change in heart rate (Terry et al., 2020).
- Sleep loss. After 30 hours awake, 11 men covered about 3% less distance in a 30-minute self-paced treadmill run at the same perceived effort and heart rate (Oliver et al., 2009). A run that feels normal after a night without sleep may simply be slower.
- Mental fatigue. After a fatiguing mental task, 20 experienced runners ran 3,000 m slower with no difference in their effort ratings (MacMahon et al., 2014). Whether the effect is real is disputed: analyses that correct for publication bias suggest it may be small or absent (Holgado, Mesquida & Román-Caballero, 2023), while a broader meta-analysis still finds a small one (Brown et al., 2020).
- Treadmill versus track. At the same effort rating, 55 recreational runners went about 20% slower on a treadmill than on an indoor track at 3 out of 10, and about 10% slower at 8 out of 10 (Nguyen et al., 2026). An effort learned on a treadmill may not give the same pace off it.
- Small changes. In a cycling study with simulated wind, riders' effort ratings didn't register riding 5% harder into a headwind (Atkinson & Brunskill, 2000). Feel may miss changes this small, so the researchers recommend a power meter rather than heart rate or feel.
Can you learn to run by feel?
Judging effort and pace looks like a skill that improves with experience, though the direct evidence is small and much of it comes from cyclists:
- Asked to run 400 m laps at set paces with no feedback, 12 collegiate runners missed by about 3 seconds a lap on average, against about 9.5 seconds for 16 recreational runners (Green et al., 2010). Experience and fitness are mixed together in that comparison.
- Among 139 recreational marathoners, those with faster personal bests, more long-distance race experience and higher weekly volume slowed less in the second half (Swain, Biggins & Gordon, 2020).
- Cyclists who rode four 4 km time trials without knowing the distance or seeing any distance feedback were far slower in the first trial than riders who had both, but matched them by the fourth (Mauger, Jones & Williams, 2009).
Hiding information can have a cost. Thirteen men running a self-paced 6 km treadmill time trial with no distance feedback were slower than with accurate feedback (Faulkner, Arnold & Eston, 2011). That study hid distance rather than pace, but it's still a reason to try pace-blind running in training before you rely on it in a race.
Does training by effort work?
The evidence is encouraging but thin, and no trial yet compares training purely by feel with GPS pace targets in trained runners:
- Fitness gains: in the six-week study above, untrained men trained by RPE improved their aerobic fitness and 12-minute run distance about as much as men trained by heart rate (Johnson et al., 2017). Sedentary adults who trained only by holding a "somewhat hard" 13 on the 6–20 scale improved their fitness compared with a no-exercise group, and the work they chose at that same rating rose as they got fitter (Parfitt, Evans & Eston, 2012). With the treadmill display covered and no supervision, a later group gained less and many dropped out (Parfitt, Olds & Eston, 2015).
- Self-paced intervals: nine well-trained runners ran 6 × 4 minutes aiming for the highest average speed they could hold. Whether they rested 1, 2 or 4 minutes, they peaked at a similar effort, around 17 on the 6–20 scale, and with no clock they chose recoveries of about two minutes (Seiler & Hetlelid, 2005).
How to train by feel with your watch
The watch stays useful. What the research supports is rating effort routinely and checking it against what the watch recorded:
- Rate every run from 0 to 10. Do it at the same point each time, such as when you save the activity: in adolescent runners, ratings given 30 minutes later were about a quarter lower than ratings given right away (Mann et al., 2019), though a later review reports ratings holding steady from 1 minute to 14 days after a session (Foster et al., 2021).
- Track the trend, not the single run. Multiply each rating by the run's minutes and add up the week. In 68 recreational runners, weekly changes in this score didn't always match changes in running time alone, and the researchers recommend logging both (Napier et al., 2020).
- Use the talk test on easy days. If you can't talk in full, comfortable sentences, you're probably running harder than an easy run calls for.
- Hide pace on some runs. Set up a data screen with only time, distance or heart rate, run the session by feel, and review your splits afterward.
- Calibrate on a familiar route. If the same loop at the same effort gets faster over weeks in similar conditions, your fitness is probably improving. If it feels harder at your usual pace, check for heat, dehydration or a recent jump in training.
- Keep the effort, not the pace, when conditions change. On hot, hilly or high-altitude days, hold your planned effort and accept a slower pace.
- If you're new to running, keep a check. RPE was least precise for untrained runners at easy intensities, so a heart-rate or pace ceiling helps while you learn.
Four sessions to practice
- Talk-test easy run: 30–50 minutes at an effort where you could hold a conversation the whole way, with pace hidden.
- Progression by feel: 40 minutes, starting at 3 out of 10 and stepping up one point every 10 minutes to finish at 6; check the pace afterward.
- Even-effort intervals: 5 × 4 minutes at the hardest effort you can repeat evenly across all five, with 2 minutes of easy jogging between.
- Fartlek: 30–45 minutes of continuous running with surges of whatever length and effort feel right on the day, easing back to easy running between them.
How Garmin, Apple Watch and other platforms record perceived effort
| Platform | How you rate a run | Used in a training metric? |
|---|---|---|
| Garmin | After you save an activity, the watch asks for a perceived-effort number and how you felt; the activity's Self Evaluation setting controls how often it asks | Not documented; Garmin's acute load is based on excess post-exercise oxygen consumption (EPOC) |
| Apple Watch | Effort rating from 1 to 10 after a workout; running and other popular cardio workouts get an automatic estimate you can change | Yes: Training Load compares the last 7 days with the previous 28, based on effort ratings and workout duration |
| Polar | RPE from 1 to 10 on the watch (Grit X2, Ignite 3, Vantage V3 and others) or later in the Polar Flow app | Per session: Perceived Load is RPE × duration; longer-term load status uses heart-rate-based Cardio Load |
| COROS | RPE and notes in the COROS app, or by voice on the PACE 4, PACE 4 Pro, APEX 4 and NOMAD | Not documented; COROS describes Training Load as coming from activities with heart rate data |
| Suunto | "Feeling" after each recording: Poor, Average, Good, Very good or Excellent | Not listed in training load or recovery; the Race 2 shows it next to your form in a widget |
| Strava | Perceived Exertion from 1 to 10 when you save or edit an activity | Optional: it can stand in for heart rate in Relative Effort and Fitness & Freshness, which need a subscription |
To hide pace, on a Garmin Forerunner 970 or 570, edit the activity's data screens and swap out any pace field; lap alerts and voice alerts can also show or announce pace, so check those settings too. On Apple Watch, choose which workout views appear and edit the metrics they show. On COROS, edit the activity data pages in the COROS app.
Bottom line
Running by feel isn't the opposite of training with data. Perceived effort is data: in the few runner studies that have tested it, its training-load score tracks heart-rate-based load, and runners pacing by it slow down for heat and hills without being told. In one study of rowers, it also flagged fatigue that heart-rate zones missed. It's least precise for newer runners, especially on easy runs, and caffeine, music and treadmills can nudge it. Rate every run, do some runs with pace hidden, and compare how each run felt with what your watch recorded.
FAQ
Is running by feel better than running by heart rate?
Neither wins in every case. In the few runner studies, training load from effort ratings tracked heart-rate-based load and predicted performance about as well. Feel is the better live guide in the first minute of hard reps, when heart rate lags, and in one rowing study effort ratings picked up building fatigue that heart-rate zones missed. Heart rate was more precise for untrained men on easy runs, so it makes sense to use both.
What should an easy run feel like?
Easy enough to talk in full, comfortable sentences. In one lab study, the point where people first weren't sure they could speak comfortably sat near their ventilatory threshold, but another found even comfortable speech above it, so treat the talk test as a rough guide. On the 0–10 scale used for session RPE, 2 is labeled easy and 3 moderate, though studies that sort sessions into easy, moderate and hard zones count ratings up to 4 as easy.
Is perceived effort accurate for beginners?
Less so. In a six-week study of untrained men, those guided by effort ratings missed their target intensities by more than those guided by heart rate, especially on easy runs, though their fitness improved about as much. A heart-rate or pace ceiling helps while you learn.
Does caffeine or music make a run feel easier?
Yes, a little. Across 21 studies, caffeine lowered perceived effort during exercise by about 5.6%, and music lowers it slightly without a significant change in heart rate. If you pace by feel, you may run a little harder than usual.
What is session RPE?
A rating of a whole session's effort, usually from 0 to 10, multiplied by its length in minutes. A 45-minute run rated 4 scores 180. Polar's Perceived Load uses the same calculation. Apple Watch's Training Load is also built from effort ratings and workout duration, but for runs the rating starts as the watch's own estimate from heart rate, VO2 max and personal data, so change it in the workout summary or the iPhone Fitness app if the run felt harder or easier.
Does Garmin use perceived effort in its training metrics?
Garmin doesn't document any use. Its watches ask for a perceived-effort number and how you felt after you save an activity, and you can view the ratings in Garmin Connect, but its acute load is based on excess post-exercise oxygen consumption (EPOC).
Should I hide pace on my watch?
On some runs, it's worth trying. Comparing how a run felt with your splits afterward is a simple way to check your sense of effort against the numbers. Try it in training before a race: in a lab study, men ran a 6 km time trial slower with no distance feedback than with accurate feedback, though that study hid distance, not pace.
Research cited
- Atkinson G, Brunskill A (2000). Pacing strategies during a cycling time trial with simulated headwinds and tailwinds. Ergonomics.
- Baden DA, McLean TL, Tucker R, et al. (2005). Effect of anticipation during unknown or unexpected exercise duration on rating of perceived exertion, affect, and physiological function. British Journal of Sports Medicine.
- Borg GA (1982). Psychophysical bases of perceived exertion. Medicine and Science in Sports and Exercise.
- Brick NE, Fitzpatrick BL, Turkington R, Mallett JC (2019). Anticipated task difficulty provokes pace conservation and slower running performance. Medicine and Science in Sports and Exercise.
- Brown DMY, Graham JD, Innes KI, et al. (2020). Effects of prior cognitive exertion on physical performance: a systematic review and meta-analysis. Sports Medicine.
- Casa DJ, Stearns RL, Lopez RM, et al. (2010). Influence of hydration on physiological function and performance during trail running in the heat. Journal of Athletic Training.
- Doherty M, Smith PM (2005). Effects of caffeine ingestion on rating of perceived exertion during and after exercise: a meta-analysis. Scandinavian Journal of Medicine and Science in Sports.
- Faulkner J, Arnold T, Eston R (2011). Effect of accurate and inaccurate distance feedback on performance markers and pacing strategies during running. Scandinavian Journal of Medicine and Science in Sports.
- Faulkner J, Parfitt G, Eston R (2008). The rating of perceived exertion during competitive running scales with time. Psychophysiology.
- Foster C, Boullosa D, McGuigan M, et al. (2021). 25 years of session rating of perceived exertion: historical perspective and development. International Journal of Sports Physiology and Performance.
- Foster C, Florhaug JA, Franklin J, et al. (2001). A new approach to monitoring exercise training. Journal of Strength and Conditioning Research.
- Garcia MC, Bazett-Jones DM (2022). Global positioning system watches and electronic journals: are training-load measures similar in high school cross-country runners? International Journal of Sports Physiology and Performance.
- Green JM, Olenick A, Eastep C, Winchester L (2016). Caffeine effects on velocity selection and physiological responses during RPE production. Applied Physiology, Nutrition, and Metabolism.
- Green JM, Sapp AL, Pritchett RC, Bishop PA (2010). Pacing accuracy in collegiate and recreational runners. European Journal of Applied Physiology.
- Haddad M, Stylianides G, Djaoui L, et al. (2017). Session-RPE method for training load monitoring: validity, ecological usefulness, and influencing factors. Frontiers in Neuroscience.
- Holgado D, Mesquida C, Román-Caballero R (2023). Assessing the evidential value of mental fatigue and exercise research. Sports Medicine.
- Impellizzeri FM, Shrier I, McLaren SJ, et al. (2023). Understanding training load as exposure and dose. Sports Medicine.
- Johnson EC, Pryor RR, Casa DJ, et al. (2017). Precision, accuracy, and performance outcomes of perceived exertion vs. heart rate guided run-training. Journal of Strength and Conditioning Research.
- MacMahon C, Schücker L, Hagemann N, Strauss B (2014). Cognitive fatigue effects on physical performance during running. Journal of Sport and Exercise Psychology.
- Mann RH, Williams CA, Clift BC, Barker AR (2019). The validation of session rating of perceived exertion for quantifying internal training load in adolescent distance runners. International Journal of Sports Physiology and Performance.
- Manzi V, Bovenzi A, Castagna C, et al. (2015). Training-load distribution in endurance runners: objective versus subjective assessment. International Journal of Sports Physiology and Performance.
- Marcora SM, Staiano W (2010). The limit to exercise tolerance in humans: mind over muscle? European Journal of Applied Physiology.
- Mauger AR, Jones AM, Williams CA (2009). Influence of feedback and prior experience on pacing during a 4-km cycle time trial. Medicine and Science in Sports and Exercise.
- Napier C, Ryan M, Menon C, Paquette MR (2020). Session rating of perceived exertion combined with training volume for estimating training responses in runners. Journal of Athletic Training.
- Nguyen AP, Kisita V, Van Cant J, et al. (2026). Reproducibility of rate of perceived exertion-based self-selected running speeds on indoor track and treadmill conditions in recreational runners. Journal of Strength and Conditioning Research.
- Oliver SJ, Costa RJ, Laing SJ, et al. (2009). One night of sleep deprivation decreases treadmill endurance performance. European Journal of Applied Physiology.
- Parfitt G, Evans H, Eston R (2012). Perceptually regulated training at RPE13 is pleasant and improves physical health. Medicine and Science in Sports and Exercise.
- Parfitt G, Olds T, Eston R (2015). A hard/heavy intensity is too much: the physiological, affective, and motivational effects (immediately and 6 months post-training) of unsupervised perceptually regulated training. Journal of Exercise Science and Fitness.
- Persinger R, Foster C, Gibson M, et al. (2004). Consistency of the talk test for exercise prescription. Medicine and Science in Sports and Exercise.
- Pind R, Hofmann P, Mäestu E, et al. (2021). Increases in RPE rating predict fatigue accumulation without changes in heart rate zone distribution after 4-week low-intensity high-volume training period in high-level rowers. Frontiers in Physiology.
- Quinn TJ, Coons BA (2011). The talk test and its relationship with the ventilatory and lactate thresholds. Journal of Sports Sciences.
- Rodrigues Júnior JFC, Mendes TT, Gomes PF, et al. (2023). Reduced running performance and greater perceived exertion, but similar post-exercise neuromuscular fatigue in tropical natives subjected to a 10 km self-paced run in a hot compared to a temperate environment. PLOS ONE.
- Sanders D, Myers T, Akubat I (2017). Training-intensity distribution in road cyclists: objective versus subjective measures. International Journal of Sports Physiology and Performance.
- Scherr J, Wolfarth B, Christle JW, et al. (2013). Associations between Borg's rating of perceived exertion and physiological measures of exercise intensity. European Journal of Applied Physiology.
- Schneider DA, Wing AN, Morris NR (2002). Oxygen uptake and heart rate kinetics during heavy exercise: a comparison between arm cranking and leg cycling. European Journal of Applied Physiology.
- Seiler S, Hetlelid KJ (2005). The impact of rest duration on work intensity and RPE during interval training. Medicine and Science in Sports and Exercise.
- Sharma AP, Saunders PU, Garvican-Lewis LA, et al. (2017). The effect of training at 2100-m altitude on running speed and session rating of perceived exertion at different intensities in elite middle-distance runners. International Journal of Sports Physiology and Performance.
- Swain P, Biggins J, Gordon D (2020). Marathon pacing ability: training characteristics and previous experience. European Journal of Sport Science.
- Terry PC, Karageorghis CI, Curran ML, et al. (2020). Effects of music in exercise and sport: a meta-analytic review. Psychological Bulletin.
- Townshend AD, Worringham CJ, Stewart IB (2010). Spontaneous pacing during overground hill running. Medicine and Science in Sports and Exercise.
- Tucker R (2009). The anticipatory regulation of performance: the physiological basis for pacing strategies and the development of a perception-based model for exercise performance. British Journal of Sports Medicine.
- Wallace LK, Slattery KM, Coutts AJ (2014). A comparison of methods for quantifying training load: relationships between modelled and actual training responses. European Journal of Applied Physiology.
- Wallace LK, Slattery KM, Impellizzeri FM, Coutts AJ (2014). Establishing the criterion validity and reliability of common methods for quantifying training load. Journal of Strength and Conditioning Research.
- Wingo JE, Stone T, Ng J (2020). Cardiovascular drift and maximal oxygen uptake during running and cycling in the heat. Medicine and Science in Sports and Exercise.