FondoLab

Roadbook

How your zones and training plan are worked out

Your zones are shares of your FTP, and your plan follows the rules below. Every number, where it comes from, and which are FondoLab's own judgement.

The numbers and their sources

A number marked FondoLab default is our judgement, not a study's finding; we say so rather than borrow authority.

NumberValueBasis
Power zones 1 < 55% · 2 to 75% · 3 to 90% · 4 to 105% · 5 to 120% · 6 to 150% · 7 above, of FTP Coggan's seven training levels, the most widely used power zones, so they match what other training apps show. They are a convention for describing training, not thresholds measured in you. Each zone runs up to the next one's start, so every watt has a zone. [1]
The FTP the zones use your profile's FTP: from your latest fitness test, else as you entered it FondoLab default. The same FTP pacing uses, so your zones and your race targets never disagree. The plan's watts are shares of it too.
FondoLab default
The phases to your A-race counted back from race day: taper 1–2 weeks, specialty 3–6, build 4–8, base the rest (at least 4, at most 20), up to 3 weeks of prep before it. Under 8 weeks: no base. Under 3: taper only FondoLab default. Planning a season in phases is a coaching convention: it rests on old stress theory rather than on trials that show one model works best, so these lengths are our judgement, and we say so. [2]
The taper 2 weeks (1 when the race is under 8 weeks away): 65% then 45% of the last full week's riding, with the same number of hard sessions, shorter Two meta-analyses agree: a taper of about two weeks, with training volume cut by 41–60% and intensity and frequency kept, gave the largest performance gains. Our two steps cut 35% and 55%, about 45% on average. A taper after a block of overload worked better still, so the last specialty week is the hardest of the season. [3][4]
Recovery weeks every 4th week in prep, base and build, at any age, at 60% of the week before, with one short hard session FondoLab default. A lighter week to absorb the ones before it is a coaching convention, not a finding: no study we know of compares the rhythms. It costs no fitness: the fitness number (CTL) dips because it is an average of your recent load, not a measure of what you can do. A cut in volume with the intensity kept is how a taper works, and two weeks of it raised performance. Fitness is lost when training stops or is too little for weeks, not in one lighter week. Masters cyclists recovered from hard intervals as fast as young ones, so riders over 50 get the same rhythm. [3][5][6]
How fast your riding grows from the hours you ride now, each loading week 10% or 1 h more, whichever is bigger, up to the time you have; fitness (CTL) rises at most 5 a week FondoLab default, to start where you are. The "10% rule" is a cautious convention from running, not a proven safeguard: in novice runners a programme built on it had as many injuries as a faster one, and in another study runners who added 10–30% a week were injured no more often than those under 10%. From few hours a 10% step is minutes, so the plan always allows an hour. Cycling carries no impact, so the running studies are if anything cautious for it. What caps the growth is the fitness ramp, which counts the load, not the hours. [7][8]
Very hard sessions a week 1 in prep's test week and the first half of base, 2 from then on, 3 in build and specialty under 40 with 8 h a week or more; 1 in recovery weeks; always at least 48 h apart, never the day after the long ride or a race Most training of successful endurance athletes is easy, with about one session in five hard. The numbers per week are our default. Masters cyclists recovered as fast as young ones 48 h after hard intervals, but felt recovered later; so from 50 the plan keeps the hard sessions and their rhythm. [9][6]
How hard a workout is easy: endurance, long and recovery rides, openers · moderate: tempo, race-pace climbs, race rehearsal · hard: threshold, low-cadence tempo, threshold after hours · very hard: VO2max intervals, over-unders, the fitness test FondoLab default. Only very hard workouts, with work above threshold, count toward the very hard sessions a week and keep 48 h between them. Hard ones, at threshold or just under it, take the week's one tempo or threshold slot, which the plan keeps off the days next to two heavy ones and off the day before a very hard one. Openers are three single minutes on an easy ride: they wake the legs, not tire them.
FondoLab default
The warm-up 15 min building from 45% to 75% of FTP (10 min when the day is short); before threshold, over-unders and VO2max work, then 3 × 30 s at 105–115% with 90 s easy after each FondoLab default, the way most coaches do it. A few short efforts near interval power before the first interval get breathing and the legs going, so the first interval is not spent finding the pace. They are kept short, with easy riding after each, so they wake you up without costing anything the intervals need.
FondoLab default
Carbs on training rides none needed under 60 min; then an hour: easy 40–60, moderate 60–80, hard 70–90, very hard 80–90 g; a race rehearsal eats the race's own plan FondoLab default. The harder you ride, the more of your energy comes from carbohydrate, so the harder workouts get more. Up to about 60 g an hour one sugar will do; above it you need glucose and fructose together, and the gut learns to take more with practice. Easy rides stay below race intake on purpose: race amounts are practised on the race rehearsal, eating what you plan to eat on the day. [10][11]
Easy and hard riding pyramidal: most riding easy, less at tempo and threshold, least above it Polarized and pyramidal training gave the same VO2max and time-trial gains in a network meta-analysis of each athlete's data; polarized had a small VO2peak edge in highly trained athletes over short programmes. Pyramidal fits a gran fondo's long tempo climbs and riders with little time: that last part is our judgement. [12][13]
What a gran fondo plan trains FTP and long efforts: endurance, tempo, threshold, VO2max, durability and the race's climbs; no sprints, anaerobic or breakaway work FondoLab default. A gran fondo is ridden at a steady effort for hours, so the plan leaves out what races are won with in a sprint. VLamax is shown with your fitness test but not aimed at: no source we read gives a VLamax to aim for.
FondoLab default
Fitness tests in the plan in the first week without a race when your FTP is not known, then in a recovery week without a race once the last test is 8 weeks old FondoLab default, so a test is never older than the 90 days after which the profile marks it out of date, and every target follows your current FTP.
FondoLab default
Planned load TSS per step: hours × IF² × 100, IF the middle of the step's target as a share of FTP; all-out efforts count at 110% Training Stress Score as its makers define it: an hour at FTP is 100. The all-out share is our estimate. Load shapes the plan's progression and taper; it is not used as an injury-risk gauge, because the acute:chronic workload ratio's link to injury has never been shown to be causal. [14][15]
Fitness, fatigue and form fitness (CTL) a 42-day and fatigue (ATL) a 7-day average of daily TSS, form (TSB) fitness minus fatigue; from the fitness you give, else your hours now at 45 TSS an hour; races at 56 TSS an hour (4 h when there is no prediction); fitness rises at most 5 a week (3 under 3 h a week) The impulse-response model as Coggan uses it, with its usual 42 and 7 days, so the numbers match what other training apps show. The curve is the plan's, from planned sessions: it says what the plan asks of you, not how fit you will be. The start from your hours, the race load and the ramp are FondoLab defaults: a ramp of 5 a week is a common coaching guide, not a finding, and the plan trims a week's riding rather than exceed it. [14]
How deep form goes in loading weeks about −10 to −30; a fitness ramp of 5 a week, held, settles form near −25 FondoLab default, from coaching practice, not a study. Coaches using the Performance Manager call −10 to −30 the range where training works, past −30 a short overload, and much lower a reason to back off. A steady ramp sets the depth: fitness trails the load by 42 days and fatigue by 7, so form sits near minus the weekly ramp times five. The plan goes there in loading weeks and comes back to around zero in recovery weeks and for races. [14]
Days off off the bike, or a T-shirt ride: under 1½ h at a really easy pace, never pushing on a climb FondoLab default. A day off is for absorbing the load, not for more of it. Spinning the legs is fine as long as it stays a ride you could do in a T-shirt and talk through the whole way; the plan counts no load for it.
FondoLab default
How a week is filled races and your own rides first; the long ride on one of the two days with the most time, with a day off after it; very hard sessions at least 48 h apart on days with 45 minutes or more, one of them the day after a day off; one tempo or threshold session, never the day before a very hard one; endurance up to the week's hours; at least one day off, the day before a very hard session when you gave time every day; two easy days before an A or B race and an easy day after any race FondoLab default, so hard days never follow each other and the plan fits the time you gave it. The key session goes the day after a day off, as coaches commonly advise, so you ride it fresh; a hard session the day before the long ride is fine. The day after the long ride is a day off, even if you gave time that day: the longest ride of the week is the biggest load to absorb. A Sunday long ride makes Monday that day off. This is coaching practice, not a trial: the studies show how fast you recover from hard intervals, not which day they should fall on. [16][6]
Your own rides counted at 65% of FTP when easy, 75% steady, 85% hard and 90% a race; a hard ride or race is one of the week's very hard sessions, a steady one its tempo session, an easy one as long as the long ride is the long ride; never moved, only flagged when it falls on the easy days around a race, next to another hard day, or is one hard ride more than the week holds FondoLab default. A group ride's load depends on the group and the day, so the plan takes your word for how hard it is and counts the middle of that band (IF 0.65 to 0.90, the intensity factors coaches give for endurance, tempo, a hard group ride and a race). Once Strava has the ride, the week counts what you rode instead. The plan never drops a ride you go to: it plans around it, and tells you when riding it hard costs the week. [14]
Your settings on the Plan page: 1 to 3 hard sessions a week instead of the default by age and time; the long ride on a day you pick; a recovery week after two loading weeks instead of three; polarized weeks without the tempo or threshold session before specialty; low-cadence tempo as every other tempo session of prep and base FondoLab defaults until you change them. Base and recovery weeks keep their own limits whatever you choose, and hard days stay 48 h apart. Polarized and pyramidal training gave the same VO2max and time-trial gains, so either suits a gran fondo; the specialty weeks keep their race climbs either way, as rehearsal. Low-cadence tempo is six-minute blocks at 88–94% of FTP at 50–60 rpm, a muscular effort. [12][17]
Sessions you change a swap keeps the day's purpose: a hard session for a hard one, tempo or threshold for tempo, a long ride for the long ride; a moved session changes days with that day's ride; steps you write count as very hard with 5 minutes or more above threshold or an all-out step, hard with 10 minutes at 88% of FTP or more, moderate with 10 at 75% or more FondoLab default. The plan keeps every day you change, and a week you lock, and plans the rest around your races and settings. It checks the 48 h between hard days and tells you when a change breaks it, but leaves the choice to you. How hard your own steps are follows from the same lines that sort the library's workouts, so they count toward the week's hard sessions as the library's would.
FondoLab default
Endurance ride steady riding at 56–75% of FTP Steady riding you could keep up all day. Most of your training is like this: it builds the engine and teaches your body to burn fat. [9]
Long ride steady riding at 56–75% of FTP The week's longest ride, steady and easy. A gran fondo is long: this is where you get used to the hours. [9]
Low-cadence tempo 4 × 6 min up to 7 × 6 min Six minutes just under threshold in a big gear at 50–60 rpm: a muscular effort. It builds force on the pedals without raising VLamax, as heavy strength work in the gym does. [17]
30/15s 2 × (10 × 30 s / 15 s) up to 3 × (13 × 30 s / 15 s) Thirty seconds hard, fifteen easy, in series. In trained and elite cyclists they raised VO2max and power more than longer intervals did. [18][19]
VO2max intervals 4 × 4 min up to 6 × 5 min Four to six minutes as hard as you can hold evenly. They raise your ceiling, which lifts your threshold too. [18]
Varied VO2max intervals 4 × (3 × 30 s / 70 s) up to 6 × (3 × 30 s / 70 s) Five-minute intervals with three hard surges in each. Cyclists spent longer near VO2max this way than at an even power. [20]
Threshold after hours 2 × 10 min up to 2 × 20 min A long ride with threshold efforts at the end. A fondo is decided late, when you are tired: this trains that. [21]
Race rehearsal 2 × 20 min up to 2 × 40 min A long ride with blocks at the power your race's pacing plan aims for, eating what you plan to eat on race day. Your gut gets used to the carbs. Without a pacing plan: 70–80% of FTP. [11]

The weather for your training

Where FondoLab reads the forecast for your week, what it sends for an interval route, how it checks a route and picks one for a session, when it says to set off, and what it suggests on a wet day. The plan itself never changes with the weather.

NumberValueBasis
Where you ride your home spot, or a place away on its dates, rounded to 0.1° (about 10 km) FondoLab default. Your browser rounds your location before it sends it, and our server rounds whatever comes in, so no address can be read from it. Riders in the same cell share one forecast. An endurance ride goes where you like, so a finer spot would not make its weather any truer.
FondoLab default
The week's weather today and the next 7 days at that spot, over your hours, or from sunrise to sunset; when it rains is shown from a 20% chance FondoLab default. The day's sky is the one met most often in those hours, the wind its mean as a vector, so a wind that turns does not add up. Under a 20% chance, the hours of rain would only be noise. Sunrise and sunset from NOAA's solar equations. [22]
Rides on the dry days a day is wet when its driest start in your hours, for a ride of your time that day up to 120 minutes, spends 30% of it or more in the rain; from tomorrow to 3 days out, and only in a week the forecast also gives a dry day you have time on; the long ride, the test, the hard and tempo sessions and the endurance rides then go on the dry days first, and the days off on the wet ones FondoLab default: we ride outside when we can. The weather only moves the week's sessions between its days: the 48 h between hard days, the day off after the long ride, the hours and the number of hard sessions stay as they are. When the hours do not fit on the dry days, a wet day still gets a ride. Today keeps what the forecast said yesterday, so its session does not move once the day is on; days past never change. A week it rains through stays as it was: there is no better day to move to.
FondoLab default
What a route sends a point every 10 km, at most 8, each rounded to 0.05° (about 5 km); the first and last 1 km left out; a spot passed twice asked once FondoLab default. Training routes are short and close to home, and rain a few kilometres apart is as fine as the hourly forecast can tell apart. Leaving out the ends means a route that starts at your door never sends it. Race weather asks for up to 40 points, because there the wind on each road matters.
FondoLab default
Route checks read from the GPX once, when a route is saved, for the workouts it is tied to, at the level with the longest work step; a warning only, never a block FondoLab default. You picked the road for reasons a GPX file does not show, so FondoLab trusts you and only says when it sees no part of the route that suits the workout. A dismissed warning stays dismissed until the check's result changes. The work step is the session's longest unbroken run of work, timed at an even speed for its power on the gradient, from your bike, position and FTP; not the pacing engine. Traffic lights cannot be read from a GPX file, so they are not checked.
FondoLab default
Low-cadence tempo a climb averaging 4–8% whose steepest 200 m is at most 10%, as long as the work step FondoLab default. A big gear at tempo wants a long, even climb: too flat and the cadence rises, too steep or with a ramp in it and the effort turns into a fight.
FondoLab default
Race-pace climbs a climb within ±2% of one of your A-race's 3 biggest climbs, as long as the work step; without a course for the race, any climb that long FondoLab default. The race's gradient is what the effort is trained for; its biggest climbs by metres climbed are the ones that decide the day.
FondoLab default
Tempo blocks · Threshold · Over-unders · VO2max intervals · Varied VO2max intervals · Fitness test a stretch with no descent steeper than 2%, each 50 m read over 200 m, as long as the work step FondoLab default. A long, steady effort needs road that keeps you pedalling; a real descent breaks the interval.
FondoLab default
30/15s · 40/20s not checked FondoLab default. Any road holds a repeat this short.
FondoLab default
Which route a session gets the route tied to its workout; of two, the one whose ride best fits the day's time (the longest that fits, else the shortest), then the drier at its best start; a ride timed at the session's mean power as a share of your FTP, on your focus race's bike: its speed on the flat, plus for each metre climbed the time it costs up and down at 5%, downhill no faster than 50 km/h FondoLab default. The route you choose for the day wins over all of this: any of your routes, for any ride, until the plan puts another workout on the day. For choosing a route, how long the ride takes is what matters, so its length and climbing are enough, ridden with the pacing engine's physics but not its plan. Without a route for the workout, the session is ridden where you like, with a prompt to add one.
FondoLab default
The best start the fewest expected wet minutes: the time on each stretch of road times the chance of rain there when you ride it; starts within 1 min of the driest count as equally dry FondoLab default. Rain comes first, because it changes a session more than temperature or wind does: wet roads, cold and spray. A chance of rain cannot be told apart to the minute, so close starts count as equal and the temperature and the ties below decide.
FondoLab default
Then the road of equally dry starts, the fewest minutes on an icy road, then on a road that is not dry, a damp minute counting half; starts within 10 minutes on ice, or 15 on a road that is not dry, count as equal. The road holds the rain up to 2 mm; of what lies on top of the first 1 mm, 50% runs off each hour, and the rest it loses to the air: wet from 0.9 mm, damp from 0.1 mm. It is icy at 2 °C or under when it rained in the 6 hours before The water store and its limits are those of FMI's road weather model RoadSurf. RoadSurf keeps the water on top of the pores until it evaporates; FondoLab lets it run off the road's camber, a FondoLab default, so a road after a downpour does not stay wet all day. How fast the road dries is Penman's evaporation from a wet surface (FAO-56) with the forecast's temperature, humidity, wind and cloud and the sun's height: sun, wind and dry air dry a road in an hour or two, a grey, still, humid day or a night barely at all. The road's own warmth and dew are left out, and so is traffic, which wipes busy roads drier. When the rain stops the road is not dry yet: spray, grit and less grip in corners. Ice is worse than wet, so it comes first; its bounds are a FondoLab default, shared with the kit advice. [23][24]
Then the temperature of equally dry starts, the one whose air temperature over the ride is nearest what you ride in best: from 16 °C to 30 °C counts as equally good, under it warmer is better and over it cooler; if you run cold the lower bound is 6 °C higher, if you run warm 6 °C lower; starts within 2 °C count as equal FondoLab default. After rain, the temperature changes a session most: cold stiffens legs and asks for more kit, heat costs power and fluid. A rider who runs cold gets the warmer hours, but above 30 °C cooler is better for everyone. You set whether you run cold or warm where you say where you ride.
FondoLab default
Equally dry starts in turn: rain within 0.2 mm, the road and the temperature (above), riding in the dark within 1 km, headwind within 5% of the ride, strongest gust within 5 km/h; then the earliest FondoLab default. Each step keeps the starts close to the best that remains, so a tie never chains along. Wind only breaks ties: on a training ride it changes how hard the road feels, not whether you get wet.
FondoLab default
The starts tried every 30 minutes on the clock, within your hours for the weekday, or from sunrise to sunset; the ride must be done by the end of them FondoLab default. The forecast is hourly, so a finer step would only pretend to know more. Without your hours, the latest start is the one that still gets you home by sunset. You set your hours on the page where you say where you ride. [22]
How far ahead today and the next 3 days; later days show the outlook only FondoLab default. Further out, the hour a shower comes cannot be forecast well enough to plan a start around; the day's outlook still says whether it looks wet.
FondoLab default
A ride without a route wet minutes at your home spot, or the place away, over the ride's planned time, read 5 minutes at a time; riding in the dark counted at 25 km/h; no headwind FondoLab default. Endurance, recovery and long rides go where you like, so one point is the honest answer. Pick a route on the session for the weather on the road: it is then asked for as an interval route is. Which way is driest would take a ring of points around home, a later step.
FondoLab default
Forecast calls only when you open a page that needs one; the week's weather kept 2 h, a route's 1 h for today and 3 h for later days; past 80% of the month's calls, the best start is worked out at home, not along the route FondoLab default. Nothing is fetched in the background, and This week never asks for a route. The weather models update a few times a day, so a fresher answer would rarely differ. The monthly guard keeps the forecasts going to the end of the month for every rider.
FondoLab default
The temperature on the road the forecast point's temperature moved to each stretch's own height: 0.65 °C colder for every 100 m up from the height Open-Meteo forecast for A route's forecast points send no elevation, so the forecast is for the terrain of a rounded spot up to 5 km off the road, and a summit between two points has none at all. The standard atmosphere's lapse rate moves it to the road, and nothing more is sent. On calm, clear mornings cold air lies in the valleys and the air warms with height instead; a forecast at the ground cannot see that, so the kit says so then. [25]
How cold it feels on the bike the temperature on the road, plus the wind chill of the air past you (your speed and the headwind) against riding the flat at 25 km/h in still air; +3 °C on a climb, -3 °C on a descent, -2 °C in rain (from a 50% chance, or any rain in the hour); -2 °C on a recovery ride, +2 °C on intervals FondoLab default, with the wind chill index of Environment Canada and the US weather service for the wind. Kit charts are written for riding the flat at a steady pace, so that is the zero. A climb is slow and hard work; a descent is fast, with no work and a damp jersey from the climb. The wind chill is only defined at or under 10 °C; above that it is read at 10 °C, so a descent still feels colder on a mild day. Sun and humidity are left out. [26]
Speed on a descent your own speed rolling down at no power, on your bike and position, up to 55 km/h FondoLab default. A training route's timing rides descents at the flat speed, which is fine for when you get home but too slow for how cold the air is. Past the cap, riders brake for corners and traffic.
FondoLab default
Coldest and warmest the part of the ride of at least 5 minutes that feels coldest, and the one that feels warmest, at the best start; closer than 2 °C, one line FondoLab default. A single 200 m of road decides nothing: you would not change for it.
FondoLab default
What to wear by how cold it feels: arm warmers under 16 °C; a base layer under 16 °C; knee warmers under 13 °C; a gilet under 12 °C; light gloves under 12 °C; a thermal jersey under 10 °C; tights under 8 °C; overshoes under 7 °C; a winter jacket under 5 °C; winter gloves under 5 °C; a cap under 4 °C FondoLab default, after the usual cycling kit charts. What cannot come off on the road is chosen for what 75% of the ride feels at or above, so you start slightly cold. What can come off covers the colder parts, read 5 minutes at a time and joined when closer than 10 minutes: worn throughout when needed for 90% of the ride, on at the start and off by a time, or in the pocket for the coldest part. A heavier piece worn does the lighter one's job.
FondoLab default
Rain jacket and ice a rain jacket from a 40% chance of rain on the ride or 5 expected wet minutes, in place of a gilet; ice warned of at 2 °C or under on the road, when it rains or rained in the 6 hours before FondoLab default. The forecast is for the air at 2 m; a road can freeze a little above zero in the air, on a clear night or in the shade, so the margin.
FondoLab default
Running cold or warm, and cold valleys every band moved by 3 °C, and the best start sought warmer or cooler, set where you say where you ride; and a warning for calm (under 10 km/h), clear (under 30% cloud) starts before 10:00 on a route with 100 m or more between its lowest and highest point FondoLab default. Riders differ a lot in how cold they feel; three choices keep it simple until riders ask for more. The cold valleys are a warning, not a correction: a forecast at the ground cannot say how cold they get.
FondoLab default

What reaches Open-Meteo, and what is kept, is in the privacy notice.

References

  1. Allen H, Coggan AR, McGregor S (2019). Training and Racing with a Power Meter, third edition. VeloPress, Boulder (book). Read it

    Practitioner's book (not peer reviewed) · Written for road cyclists and triathletes training with a power meter; the zones are a coaching convention, not the result of a study. · Read by us: The power zone table as the authors and TrainingPeaks publish it; the book is not read in full for this entry.

    Defines seven training levels as shares of functional threshold power (FTP), each with what it trains and how it feels.

    • Level 1, active recovery: below 55% of FTP.
    • Level 2, endurance: 56–75%.
    • Level 3, tempo: 76–90%.
    • Level 4, lactate threshold: 91–105%.
    • Level 5, VO2max: 106–120%.
    • Level 6, anaerobic capacity: 121–150%.
    • Level 7, neuromuscular power: short maximal efforts, no percentage given.

    The levels are a convention for describing training, not physiological thresholds measured in the rider: two riders at the same share of FTP can be working quite differently.

  2. Kiely J (2018). Periodization theory: confronting an inconvenient truth. Sports Medicine 48(4):753–764. doi:10.1007/s40279-017-0823-y

    Narrative review · Periodization theory in sport generally; no new data. · Read by us: Abstract

    Argues that periodization, the planning of training into phases, rests on Selye's early stress research, and that its assumptions have not kept up with how stress and adaptation are understood today.

    • New ideas have been added to the framework, but its philosophical assumptions remain largely unchallenged.
    • The author calls for training plans to be re-thought in line with current stress theory, rather than taken as settled science.

  3. Bosquet L, Montpetit J, Arvisais D, Mujika I (2007). Effects of tapering on performance: a meta-analysis. Medicine & Science in Sports & Exercise 39(8):1358–1365. doi:10.1249/mss.0b013e31806010e0

    Meta-analysis · 27 of 182 studies of competitive athletes, with performance measured in the field or in competition; endurance athletes and sprinters together, so not specific to cyclists. · Read by us: Abstract

    Pooled the effect of tapering on performance, by how long the taper lasted and how training changed during it.

    • The most effective taper lasted 2 weeks, with training volume cut exponentially by 41–60% and intensity and frequency unchanged.
    • Cutting volume had an effect of 0.72 ± 0.36 (P < 0.001) on performance.

  4. Wang Z, Wang YT, Gao W, Zhong Y (2023). Effects of tapering on performance in endurance athletes: a systematic review and meta-analysis. PLOS ONE 18(5):e0282838. doi:10.1371/journal.pone.0282838

    Systematic review and meta-analysis · 14 studies of endurance athletes across sports; cyclists are part of the pool, not all of it. · Read by us: Abstract

    Pooled the effect of tapering on endurance performance.

    • Time-trial performance (SMD −0.45) and time to exhaustion (SMD 1.28) improved from before to after the taper; VO2max and economy did not change.
    • Tapering after a block of overload training improved time-trial performance more than a conventional taper.
    • The suggested taper lasts 21 days or less and cuts training volume progressively by 41–60%, keeping intensity and frequency.

  5. Mujika I, Padilla S (2000). "Detraining: loss of training-induced physiological and performance adaptations. Part I: short term insufficient training stimulus". Sports Medicine 30(2):79–87. doi:10.2165/00007256-200030020-00002

    Narrative review · Reviews studies of highly trained and recently trained athletes who stopped training or trained too little for under 4 weeks. · Read by us: Abstract

    Detraining is the partial or complete loss of training adaptations when the training stimulus is insufficient. Part I covers under 4 weeks of it.

    • In highly trained athletes, stopping training quickly lowers VO2max and blood volume; maximal cardiac output, ventilatory efficiency and endurance performance fall with them.
    • Metabolically, more reliance on carbohydrate during exercise, with lower lipase activity, GLUT-4, glycogen and lactate threshold; capillary density and oxidative enzymes fall.
    • The changes are milder in recently trained people.
    • The review is about stopping or too little training, not about a lighter week with the intensity kept.

  6. Borges NR, Reaburn PR, Doering TM, Argus CK, Driller MW (2018). Age-related changes in physical and perceptual markers of recovery following high-intensity interval cycle exercise. Experimental Aging Research 44(4):338–349. doi:10.1080/0361073X.2018.1477361

    Controlled comparison (masters and young cyclists) · 9 trained masters cyclists (about 56 years) and 8 trained young cyclists (about 26), after one session of 6 × 30 s at 175% of peak power. · Read by us: Abstract

    Followed recovery for 48 hours after hard intervals, in performance, blood and how the riders felt.

    • Strength, sprint and time-trial performance recovered at the same rate in both groups.
    • The masters cyclists felt less motivated, more fatigued and more sore at 48 hours (moderate effects).
    • The authors suggest that feeling recovered later could hurt older athletes' training consistency even when their bodies recover as fast.

  7. Buist I, Bredeweg SW, van Mechelen W, Lemmink KAPM, Pepping GJ, Diercks RL (2008). No effect of a graded training program on the number of running-related injuries in novice runners: a randomized controlled trial. American Journal of Sports Medicine 36(1):33–39. doi:10.1177/0363546507307505

    Randomised controlled trial · 532 novice runners preparing for a 4-mile (6.7 km) event; runners, not cyclists, and about injuries, not performance. · Read by us: Abstract

    Compared a graded 13-week programme built on the 10% rule (training raised by no more than 10% a week) with a standard 8-week programme.

    • Injuries were as common with the graded programme as with the standard one: 20.8% against 20.3% (P = .90).

  8. Nielsen RØ, Parner ET, Nohr EA, Sørensen H, Lind M, Rasmussen S (2014). "Excessive progression in weekly running distance and risk of running-related injuries: an association which varies according to type of injury". Journal of Orthopaedic and Sports Physical Therapy 44(10):739–747. doi:10.2519/jospt.2014.5164

    Prospective cohort study (exploratory) · 874 healthy novice runners on a self-structured programme, followed for a year with GPS watches; runners, not cyclists, and about injuries, not performance. · Read by us: Abstract

    Grouped each runner's weekly distance progression as under 10% (or less than the week before), 10–30%, or over 30%.

    • 202 runners were injured. Injury rates did not differ significantly across the three groups.
    • For distance-related injuries only (patellofemoral pain, iliotibial band syndrome, shin splints and others), runners progressing over 30% had a higher rate than those under 10%: hazard ratio 1.59, 95% CI 0.96–2.66, P = .07, not significant.
    • The authors call the study exploratory and suggest novices keep progression under 30% a week over two weeks, pending randomised trials.

  9. Seiler S (2010). What is best practice for training intensity and duration distribution in endurance athletes?. International Journal of Sports Physiology and Performance 5(3):276–291. doi:10.1123/ijspp.5.3.276

    Review · Descriptive studies of nationally and internationally competitive endurance athletes training 10 to 13 times a week; elite athletes, not recreational riders. · Read by us: Abstract

    Reviewed how successful endurance athletes distribute their training intensity.

    • About 80% of training sessions are at low intensity (around 2 mmol/L blood lactate), and about 20% are dominated by high-intensity work such as intervals near 90% of VO2max.
    • Adding more interval training for athletes who were already well trained did not convincingly improve long-term performance.
    • A large base of long, low-intensity training with fewer hard sessions appears to balance the signal to adapt against a stress load the athlete can carry.

  10. Jeukendrup A (2014). A step towards personalized sports nutrition: carbohydrate intake during exercise. Sports Medicine 44(Suppl 1):25–33. doi:10.1007/s40279-014-0148-z

    Narrative review with practical guidelines · Reviews studies of trained endurance athletes, mostly cyclists, mostly in the lab. The author was at the Gatorade Sports Science Institute. · Read by us: Full text (open access)

    Glucose is absorbed through a sodium-dependent transporter, SGLT1, which saturates at an intake of around 60 g/h. On a single sugar, exogenous carbohydrate oxidation tops out at about 1 g/min.

    Fructose uses a different transporter. Taking glucose and fructose together, typically in a 2:1 ratio, raised oxidation to about 1.26 g/min in the studies reviewed.

    The proposed guidelines scale with duration:

    • About 1 h: a mouth rinse or small amounts of carbohydrate can help.
    • 2–3 h: up to about 60 g/h, which a single sugar can deliver.
    • Longer, ultra-endurance: about 90 g/h, which needs multiple transportable carbohydrates.
    • The advice does not depend on body weight or training status.

    The gut is trainable: a high-carbohydrate diet for 28 days increased exogenous carbohydrate oxidation, so high intakes should be practised in training.

  11. Costa RJS, Miall A, Khoo A, Rauch C, Snipe R, Camões-Costa V, Gibson P (2017). Gut-training: the impact of two weeks repetitive gut-challenge during exercise on gastrointestinal status, glucose availability, fuel kinetics, and running performance. Applied Physiology, Nutrition, and Metabolism 42(5):547–557. Read it

    Randomised controlled trial · 25 endurance runners; 2 hours of running at 60% of VO2max, eating 30 g of carbohydrate every 20 minutes. · Read by us: Abstract, in part

    After a first gut challenge, runners spent two weeks training with carbohydrate gel-discs, carbohydrate food or a placebo, then repeated the challenge.

    • Gut symptoms fell more in both carbohydrate groups than with the placebo.
    • Performance in the distance test afterwards improved in both carbohydrate groups, not with the placebo.

  12. Rosenblat MA, Watt JA, Arnold JI, Treff G, Sandbakk ØB, et al. (2025). Which training intensity distribution intervention will produce the greatest improvements in maximal oxygen uptake and time-trial performance in endurance athletes? A systematic review and network meta-analysis of individual participant data. Sports Medicine 55:655–673. doi:10.1007/s40279-024-02149-3

    Systematic review and network meta-analysis of individual participant data · 13 studies with 348 recreational and competitive endurance athletes, each comparing a polarized programme with another intensity distribution; literature searched to February 2024. · Read by us: Abstract

    Compared intensity distributions (polarized, pyramidal, threshold and others) on VO2max and time-trial performance, with each athlete's own data.

    • The primary analysis found no significant difference in VO2max or time-trial performance between the distributions.
    • Exploratory subgroups suggested competitive athletes may respond better to polarized training and recreational athletes to pyramidal training; the authors present this as a hypothesis, not a rule.

  13. Silva Oliveira P, Boppre G, Fonseca H (2024). Comparison of polarized versus other types of endurance training intensity distribution on athletes' endurance performance: a systematic review with meta-analysis. Sports Medicine. doi:10.1007/s40279-024-02034-z

    Systematic review with meta-analysis · 17 studies with 437 athletes, interventions of at least 4 weeks comparing polarized training with other distributions; literature searched to October 2022. · Read by us: Abstract

    Pooled polarized training against other intensity distributions.

    • VO2peak improved a little more with polarized training: SMD 0.24 (95% CI 0.01 to 0.48), 11 studies, 284 athletes, I² 0%. The advantage was largest in shorter interventions and highly trained athletes.
    • On the other measures of endurance performance, time trials among them, polarized training did as well as the other distributions, not better.

  14. TrainingPeaks (2026). Estimating Training Stress Score (TSS). trainingpeaks.com. Read it

    Method description (not peer reviewed) · Any ride with power and an FTP; a definition, not a finding. · Read by us: The formula and its definitions, as quoted from the article

    Training Stress Score (TSS), developed by Andrew Coggan, expresses a session's load as the product of its intensity and its duration.

    • TSS = (seconds × NP × IF) ÷ (FTP × 3600) × 100, where NP is normalized power and IF, the intensity factor, is NP ÷ FTP.
    • An hour at FTP is 100 TSS.

  15. Impellizzeri FM, Tenan MS, Kempton T, Novak A, Coutts AJ (2020). Acute:chronic workload ratio: conceptual issues and fundamental pitfalls. International Journal of Sports Physiology and Performance 15(6):907–913. doi:10.1123/ijspp.2019-0864

    Commentary on methods · Training-load and injury research in team and endurance sports; no new data. · Read by us: Abstract

    Examined the acute:chronic workload ratio (ACWR), the most used training-load metric for injury risk.

    • Recommending changes to the ratio to lower injury risk assumes a causal effect, and no study has properly estimated one; using it that way remains conjecture.
    • Ratio data and unstated assumptions add further problems.

  16. Carmichael C (2025). How to Build a Strong Week of Your Cycling Training Plan. trainright.com. Read it

    Coaching advice (not peer reviewed) · Amateur cyclists planning a training week; practice, not a finding. · Read by us: The article, as quoted from it

    A coach's rules for placing a week's workouts around rest days.

    • "Schedule the highest-priority workouts directly after a rest day": usually the hardest interval session, or the longest endurance ride when there are no hard intervals.
    • Hard days should not be so far apart that the stimulus of one workout fades before the next.
    • For some athletes, hard workouts on back-to-back days have benefits.

  17. Hansen EA, Rønnestad BR (2017). Effects of cycling training at imposed low cadences: a systematic review. International Journal of Sports Physiology and Performance 12(9):1127–1136. Read it

    Systematic review · 7 training studies found in MEDLINE and SPORTDiscus up to April 2016, with quite different riders and training. · Read by us: Abstract

    Reviewed what training at an imposed low cadence does to cycling performance.

    • There is no strong evidence that training at low cadence helps performance; some studies found no clear effect.
    • The authors tentatively suggest including low-cadence bouts at moderate to maximal intensity.
    • At low power, low cadence means below about 80 rpm, e.g. 40–70 rpm.

  18. Rønnestad BR, Hansen J, Vegge G, Tønnessen E, Slettaløkken G (2015). Short intervals induce superior training adaptations compared with long intervals in cyclists – an effort-matched approach. Scandinavian Journal of Medicine & Science in Sports 25(2):143–151. doi:10.1111/sms.12165

    Randomised controlled trial · 16 trained (not professional) cyclists over 10 weeks: 9 in the short-interval group, 7 in the long-interval group. · Read by us: Abstract

    Compared two interval sessions a week, matched for effort, over 10 weeks, with low-intensity training between.

    • Short intervals: 3 series of 13 × 30 s hard, 15 s easy, 3 minutes between series.
    • Long intervals: 4 × 5 minutes with 2.5 minutes between.
    • VO2max rose 8.7 ± 5.0% with short intervals and 2.6 ± 5.2% with long intervals, and the short-interval group improved more across the power measures.

    The groups were small, and the cyclists were trained amateurs.

  19. Rønnestad BR, Hansen J, Nygaard H, Lundby C (2020). Superior performance improvements in elite cyclists following short-interval vs effort-matched long-interval training. Scandinavian Journal of Medicine & Science in Sports 30(5). doi:10.1111/sms.13627

    Randomised controlled trial · 18 elite cyclists over 3 weeks, 9 per group, with the same total volume and intensity distribution. · Read by us: Abstract

    Repeated the effort-matched comparison of short and long intervals in elite cyclists.

    • Short intervals (30 s hard, 15 s easy) raised peak aerobic power by about 3.7%, long intervals (5 min with 2.5 min easy) lowered it by 0.3%.
    • 20-minute mean power rose about 4.7% with short intervals and fell 1.4% with long intervals.
    • VO2max did not change differently between the groups.

  20. Bossi AH, Mesquida C, Passfield L, Rønnestad BR, Hopker JG (2020). Optimizing interval training through power-output variation within the work intervals. International Journal of Sports Physiology and Performance 15(7):982–989. doi:10.1123/ijspp.2019-0260

    Randomised crossover trial · 14 well-trained cyclists (VO2max 69 ml/kg/min), one session of each kind in the lab; acute responses, not a training study. · Read by us: Abstract

    Compared interval sessions whose work intervals varied in power with sessions at constant power, matched for duration and mean power.

    • Cyclists spent longer above 90% of VO2max with varied work intervals: 410 ± 207 s against 286 ± 162 s.
    • Heart-rate and perceived-effort training loads did not differ, and the extra time was not linked to higher blood lactate.

  21. Maunder E, Seiler S, Mildenhall MJ, Kilding AE, Plews DJ (2021). The importance of 'durability' in the physiological profiling of endurance athletes. Sports Medicine 51(8):1619–1628. doi:10.1007/s40279-021-01459-0

    Current opinion · Endurance athletes in general; an argument with no new data. · Read by us: Abstract

    Points out that athletes are tested fresh, while the numbers they are tested on drift during long exercise.

    • Durability is how soon, and how far, those numbers deteriorate during prolonged exercise.
    • Ignoring it can distort how intensity is prescribed, how training load is monitored and how performance is predicted.

  22. NOAA Global Monitoring Laboratory (2026). Solar calculation details. NOAA Solar Calculator, gml.noaa.gov. Read it

    Algorithm documentation, based on Meeus, Astronomical Algorithms · Any place on Earth. Accurate to about a minute between 72° north and south, within ten minutes beyond; the approximations are very good for 1800–2100. · Read by us: Full text

    NOAA's sunrise, sunset and solar position calculators use equations from Jean Meeus's Astronomical Algorithms. We use the same equations for the sun's height at each point of a course and for sunrise and sunset at the start.

    • Sunrise and sunset are taken with 0.833° of atmospheric refraction: the sun's centre is that far below the horizon when its top edge appears.
    • The results are theoretical. Real sunrise and sunset can differ because the atmosphere varies, and errors grow further from the equator.

  23. Karsisto V, Horttanainen M, Kangas M, Nurmi P, Ylhäisi J, Partio M (2024). "RoadSurf 1.1: open-source road weather model library". Geoscientific Model Development 17:4837–4853. Read it

    Model description, with its open-source code (github.com/fmidev/RoadSurf) · The Finnish Meteorological Institute's operational road weather model, in use on Finnish roads since the 2000s; first described by Kangas et al. 2015 (Meteorological Applications 22:544–553) · Read by us: Source code and user manual

    RoadSurf works out the road's temperature and what lies on it from the weather forecast. The water on the road is a store in millimetres: rain adds to it, evaporation takes from it, and traffic wears it off.

    • By default the asphalt's pores hold 1.0 mm and another 1.0 mm can lie on top; more runs off.
    • The road is damp from a tenth of the pores' water (0.1 mm) and wet from nine tenths (0.9 mm); under that it is dry.
    • Evaporation is the bulk formula from the road surface's vapour pressure to the air's, over an aerodynamic resistance from the wind; the road's temperature comes from a full energy balance of the road layers. With no water left, nothing evaporates.
    • Traffic wipes water off busy roads. We leave that out: training rides go on quiet roads.

    We use the store, its sizes and the dry, damp and wet limits. We do not have the road's own temperature, so evaporation comes from the FAO-56 equation for a wet surface instead (allen-1998-fao56).

  24. Allen RG, Pereira LS, Raes D, Smith M (1998). "Crop evapotranspiration: guidelines for computing crop water requirements (FAO Irrigation and Drainage Paper 56)". Food and Agriculture Organization of the United Nations, Rome. Read it

    Method guidelines; the standard Penman–Monteith equation and how to get its inputs from weather data · Worldwide; hourly and daily weather data · Read by us: Equations, from the guidelines and their standard restatements

    The FAO-56 Penman–Monteith equation is the international standard for how much water a surface gives to the air, from temperature, humidity, wind and radiation.

    • The hourly form (eq. 53) has a radiation term and an air term: sun and warmth drive evaporation, and so do wind and dry air. Heat into the ground is a tenth of the net radiation by day and half at night (eq. 45, 46). Wind at 10 m is brought to 2 m (eq. 47); net longwave is eq. 39.
    • The reference crop has a surface resistance; a wet road has none, so we drop it: the open-water case of Penman's equation.
    • The sunshine comes from the sun's height and the cloud cover, by Kasten and Czeplak (1980, Solar Energy 24:177–190): 910·sin(h) − 30 W/m² under a clear sky, times 1 − 0.75·(cloud)^3.4. Asphalt reflects a tenth of it (RoadSurf's dry albedo).
    • It leaves out the road's own warmth, which dries a road faster after a sunny spell, and dew, which can wet a road on a clear night. The road is read as the air sees it.

  25. NOAA, NASA and US Air Force (1976). U.S. Standard Atmosphere, 1976. NASA-TM-X-74335, US Government Printing Office. Read it

    Standard reference atmosphere · Mid-latitude, yearly average conditions; identical to the ICAO standard atmosphere up to 32 km · Read by us: Abstract

    The standard atmosphere is an idealised, steady air column: 15 °C and 1013.25 hPa at sea level, the temperature falling 6.5 °C for every kilometre up to 11 km.

    • We use its lapse rate to move a forecast temperature from the height it was made for to the height of the road.
    • Real air departs from it every day. On calm, clear nights and mornings cold air drains into valleys and the temperature rises with height for a while (an inversion); in sunshine the air near the ground falls faster than 6.5 °C per km.

  26. Osczevski R, Bluestein M (2005). The new wind chill equivalent temperature chart. Bulletin of the American Meteorological Society 86(10):1453–1458. Read it

    Model and its derivation · A model of heat loss from the face of an adult walking into the wind, tested on volunteers in a wind tunnel · Read by us: Abstract

    The wind chill index that Environment Canada and the US National Weather Service adopted in 2001: how cold the air feels on bare skin for a temperature and a wind speed, as the temperature that would take the same heat in still air.

    • The formula is defined for air at or under 10 °C and wind from 4.8 km/h. Its wind is the forecast's, at 10 m; it reduces that to face height itself.
    • We use it for the air past a rider: riding speed plus headwind, at the rider's height. That is more air than the formula expects at a given number, so we use only the difference between two speeds, riding the flat and the stretch in question, not the index itself.
    • It is for bare skin, not for a clothed rider, and it leaves out sun and humidity.

← Back to your training