You test your athletes a few times a year, you log the FTP, and you build the block around it. It is a good number. It is also one number, describing one effort, and it has quietly been asked to describe a whole rider. Two athletes who share an FTP can race nothing alike, and you already know which of yours they are. The gap between what FTP measures and what actually decides their results is the key to a happy rider. Below are the metrics that fill that gap, what each one tells you about a rider, and a way to test every one of them.
FTP measures one effort your athletes will almost never make
FTP describes the power a rider can hold steady, solo, for around an hour. Now count how often any of your athletes actually does that. A club ride is surges and lulls. A kermesse is a sprint out of every corner. A sportive is three hours of rolling terrain and one climb that decides everything. None of them is a flat solo hour. FTP is real, but it sits at the centre of a target most of your athletes' riding never touches. Treat it as one data point about the engine, not a verdict on form.
Two riders with the same FTP are not the same rider
Give two of your athletes the same 280 watt FTP and you still know almost nothing about who beats whom. One holds threshold for forty minutes before the wheels come off. The other holds it for seventy. That gap, time to exhaustion at threshold, is the missing half of the number, and FTP is completely silent on it. To measure it, have the athlete warm up properly, then hold their current FTP until they genuinely cannot, and record the minutes. Do that once per athlete and you learn more about their ceiling than another test of the ceiling itself.
Time to exhaustion is trainable and almost nobody trains it
Most riders chase a higher threshold and ignore how long they can sit at the one they have. That duration responds to training, and it responds fast, because the limiter is often fatigue resistance rather than raw power. Prescribe work just under threshold, in the region of 88 to 92 percent of FTP, in blocks long enough to hurt: two by twenty minutes, building toward three by fifteen at the top of that band. Retest time to exhaustion at threshold roughly once a season. A longer hold at the same watts is progress the FTP number will never show you.
An athlete's threshold after 2,500 kilojoules matters more than their threshold fresh
Everyone is strong in the first hour. Races and hard rides are decided in the fourth. Durability is what a rider's threshold becomes once they have already spent 2,500 kilojoules of work, and for most it sags, sometimes by 10 to 15 percent, sometimes more. The athlete who fades least wins the sprint nobody thought they had, and often it is not the one with the highest fresh FTP. To test it, have them ride two to three easy hours to bank the work, then do a short threshold effort late in the ride and compare the power to a fresh effort. The drop is the real number.
Repeatability decides group racing, not steady power
Steady power wins time trials. Group racing is won by the rider who can go hard, recover, and go hard again without the top of each effort collapsing. That is repeatability, and it is invisible in a single interval, because one clean effort tells you nothing about the fifth one. Set six by three minutes, all-out, with three minutes easy between, and ignore the average. Look at the drop from the first effort to the last. A small fade means the athlete can answer attack after attack. A big fade means they have one good match, and then they are a passenger for whatever the group decides to do next.
Anaerobic work capacity is a separate number from the engine
Above threshold a rider is spending from a small, finite tank, often pictured as a box of matches. Some athletes have a huge aerobic engine and a tiny matchbook; they crush long climbs but get shelled the moment the race turns punchy. Others are the reverse. This capacity, the total work a rider can do above threshold before they stop, is independent of FTP, and an athlete can be strong or weak at it regardless of their engine. Test it with a two to three minute maximal effort from fresh, fully rested and fully committed. That flat-out power over that duration is the size of the box, and it tells you whether to coach the rider to sit in and wait or to force the race hard and early.
Aerobic decoupling tells you whether an athlete's base is real
A rider can fake fitness for an hour. Nobody fakes it for four. Aerobic decoupling is the drift between power and heart rate over a long steady ride: if power holds but heart rate creeps upward, the aerobic base is thinner than the FTP suggests. Have the athlete ride two to three hours at a steady endurance effort, split the ride in half, and compare average power to heart rate in each half. Under about 5 percent drift is a solid base. This number improves before FTP moves, so it is your earliest sign the winter work is landing, well before a retest would show it.
Carbohydrate tolerance is a performance number, not a nutrition footnote
An athlete can have the engine and still lose because they ran out of fuel, or because their gut refused what they fed it in the third hour. The amount of carbohydrate a rider can absorb per hour is trainable, it varies a lot between individuals, and it directly caps how long they can hold high power. Most start around 60 grams per hour and can build toward 90 to 120 grams per hour with deliberate practice. Program it the way you program intervals: fuel on every long ride, push the amount up gradually, and find the ceiling in training rather than discovering it at kilometre 120 of a race.
The test protocol invents part of the number
Two athletes with identical fitness can post different FTPs purely from how they tested. A ramp test tends to flatter riders with a big anaerobic contribution, because a strong finish inflates the estimate. A 20 minute test punishes anyone who paces it badly and goes out too hard. Neither is wrong, but they are not interchangeable, so a jump in an athlete's FTP can be a change in protocol rather than a change in the rider. Standardise one test across your roster, repeat it the same way each time, and only compare like with like. Consistency of method is worth more than the method itself.
Track a curve, not a point
FTP is a single point on a line that describes the whole engine, and the line is what actually decides the season. Track the curve instead: each athlete's best power for 5 seconds, 1 minute, 5 minutes, 20 minutes and 60 minutes, updated as they ride. Add one durability check and one repeatability session per training block. Read together, that handful of numbers tells you where a rider is strong, where they get dropped, and what to train next, which a single threshold figure never will. It also stops you programming the one number up while the parts that actually cost your athletes races quietly stay flat.
Tracking this for one athlete is manageable. Tracking it across a full roster, catching the durability that quietly slipped or the repeatability that stopped improving before it costs a result, is where a coach either has the picture or does not. That gets a lot easier when every athlete's data sits in one place and the trends surface on their own, which is part of what we are building in Wyder for Coaches. Either way, the point stands: stop coaching to one number, and start reading the shape of the whole engine.