TriForward
Racing & Strategy ·

Your Hour-Four FTP Isn't Your FTP: The Problem With Intensity Factor

Race %FTP targets are calculated against a fresh FTP — but that threshold falls 5–15% by hour three, so the same target quietly gets harder to hold as the race goes on.

Your Hour-Four FTP Isn't Your FTP: The Problem With Intensity Factor

Sprint Summary

The short version — read this if you're short on time.

Intensity Factor isn't broken as a concept — it's a useful shorthand for how hard you're working relative to a benchmark. The problem is narrower and more specific: that benchmark is measured fresh, and the physiology of a long bike leg guarantees it won't stay fresh for the length of time you need it to hold.

The measured Ironman fade (239 W to 203 W in strong finishers), the arithmetic illustration of how identical watts can represent a 78% or 88% effort depending only on fuelling, and the 17-percentage-point spread in published coaching targets all point toward the same conclusion: race pacing is a moving-target problem, not a single-number problem, and the honest answer is that nobody has yet run the controlled trial that would tell you the single best way to solve it at 70.3 or Ironman distance.

What the evidence does support is scaling your target down with duration, fuelling to defend it, and treating the back half of the bike as the part of the race where the plan matters most — because that's where the ceiling you calculated your plan against has already started to fall.

Safety & Context

This article covers pacing arithmetic and published coaching guidance, not medical advice; the fuelling and heat comparisons referenced come from controlled lab trials and should not be replicated as fasted or water-only race-day practice.

Anyone with cardiovascular risk factors or a history of exertional heat illness should get individual pacing and fuelling guidance from a physician or qualified coach rather than applying any single %FTP number from this article directly.

Full Distance

The complete research and analysis.

Most age-group triathletes set their bike power target the same way: take FTP, multiply it by a percentage a coach or article recommended, and treat that number as fixed for the whole ride. The problem usually isn't the percentage. It's the FTP.

By the time you're three or four hours into a 70.3 or Ironman bike leg, the FTP you tested in a fresh, rested state on a Tuesday evening has already moved — and not in your favour. Intensity Factor (IF), the ratio coaches use to describe how hard you're riding relative to threshold, is computed against that fresh number for the entire race, even though the ceiling it's measuring against has been quietly falling since somewhere around hour one.

This article is about that specific problem: the target number itself, not how you feel while chasing it, and not how smoothly you pace around it. We've covered feel-versus-data pacing and the cost of power variability separately; here the question is simpler, and in some ways more uncomfortable — what if the number you were pacing to was never really achievable for the length of time you needed it to hold?

For a time-crunched age-grouper training 6–10 hours a week, the FTP test is one of the few pieces of race-week planning that feels genuinely solid. It's a real number, tested on a real day, and it anchors every zone, every workout target and every race plan you build afterwards.

That solidity is exactly why the threshold-drift problem matters: it isn't that your FTP test was wrong, or that your training was wrong. It's that the number was only ever true for the conditions it was tested in — rested, fed, cool, and short. None of those conditions describe hour three of a half-Ironman bike leg, let alone hour five of a full.

The bike is also not a side event you can afford to get wrong. Across a large systematic review of IRONMAN 70.3 racing, cycling is the single strongest predictor of finishing time, more predictive than either the swim or the run. If your race power target quietly becomes harder to hold as the ride goes on, that isn't a minor pacing inconvenience — it's happening during the discipline that decides the race.

What the evidence says

The bike decides the race, so the target matters more here than anywhere else

A PRISMA systematic review of 86 studies on IRONMAN 70.3 racing found cycling makes up 50–60% of total race duration and predicts finishing time with a correlation of 0.85 — higher than running (0.82 for men, 0.75 for women) or swimming (0.63 and 0.46). V̇O₂max alone explained 67% of the variance in overall finishing time. An independent analysis of 16,611 professional race records confirmed cycling as the most predictive split, with r = 0.85 and R² = 0.800. For scale, age-group finishers average 5:46:53 (men) and 6:16:18 (women) over the 70.3 distance across more than 823,000 finisher records.

Evidence grade: strong for the association, though it is explicitly correlational — fitter athletes both ride faster and finish faster, so cycling predicting the result doesn't prove the bike alone causes it.

The measured fade — what a top-10% finisher's own power meter recorded

The clearest field measurement of exactly how much a race power target erodes comes from six well-trained male triathletes who finished in the top 10% of an Ironman, wearing SRM power meters on a flat, three-lap, 180 km course. Across the successive laps, their mean power fell from 239 ± 25 W to 203 ± 20 W — a 15% drop (P < 0.05) — while cadence dropped from 89 to 82 rpm and speed fell from 36.5 to 33.1 km/h.

This is a small, dated study — six athletes, one race, 2006-era equipment — but it remains the canonical field measurement of Ironman bike fade, and a 15% drop in athletes finishing in the top tenth of the field is a sobering baseline for anyone racing further back. Pacing shape also differs by sex: among 60 professionals at an Ironman World Championship, men rode with more negative pacing and women with more even pacing.

The arithmetic problem — and why it matters more than the fade itself

Here is the part that turns a known fatigue problem into a target-setting problem. Intensity Factor and every%FTP race target are calculated as race power divided by a fresh, rested FTP, for the whole race, start to finish. But the evidence above says the “rested FTP” ceiling doesn't hold still — it falls as the ride goes on. That means a target that looks conservative when you calculate it in your kitchen the night before the race can become genuinely aggressive by the time you're living inside it.

We can illustrate this using the published means from a controlled crossover trial that measured critical power — a close cousin of FTP — before and after three hours of riding under three fuelling conditions. Fresh critical power averaged 277 W. After three hours on water only, it had fallen to 236 W; after three hours at 120 g/h of carbohydrate, it had only fallen to 266 W.

Take a rider targeting 0.75 of fresh critical power — 208 W, a number most coaches would call conservative:

  • After three hours unfuelled, that same 208 W is 88% of the rider's remaining critical power.
  • After three hours at 120 g/h, the same 208 W is 78% of remaining critical power.

Same watts on the head unit. Same fresh-FTP-based plan. A materially different real effort, depending entirely on how the rider fuelled — not on how hard they decided to ride.

It's important to be precise about what this calculation is: it is our own arithmetic applied to that study's published means. It is not a finding the study itself reported, and it should be read as an illustration of the mechanism rather than a measured result. But the mechanism it illustrates is real — fuelling, heat and pacing all ultimately act on the same variable, which is what your target watts cost you in the third or fourth hour, because they all move the ceiling the target is a fraction of. The same logic applies to heat, where V̇O₂max has been measured to fall 15–19% within 45 minutes of even easy riding in hot conditions, a finding confirmed in a separate trial.

Distance changes the shape of the fade — and even sprint distance isn't exempt

The way power fades also depends on race distance. Eight trained male triathletes raced sprint, Olympic and half-Ironman events within two months, wearing power meters and GPS. At sprint distance they spent 43.8% of the bike more than 10% above their mean power; at half-Ironman distance that fell to 20.9% (p = 0.002 vs sprint), while time spent well below mean power rose from 5.9% to 13.6%. The run-pacing shape flipped too: negatively paced after the sprint bike, positively paced — meaning slower as the run went on — after Olympic and half-Ironman distance.

The practical reading: short-course racing is genuinely intermittent and can tolerate a negatively paced run; long-course racing is smoother, and the run is essentially always positively paced. Fading on the run at long-course distance is the norm, not a personal failure — how you pace the bike decides how much you fade.

Don't assume this is only a long-course problem, either. In eight international and national junior/U23 male triathletes, a simulated stochastic sprint-triathlon cycling bout produced significant declines in fatigued-state running measures — V̇O₂peak, velocity at V̇O₂peak, and time to task failure all fell meaningfully (all p < 0.05). Durability, in other words, isn't a long-course-only concern; it shows up even over a sprint-distance bike leg.

How far apart the practitioner targets actually are

If the threshold that IF is measured against is moving, you'd hope the published guidance on what IF to target would at least agree with itself. It does not. Here is what different coaching sources publicly recommend for the same half-Ironman bike leg:

That's a 17-percentage-point spread — IF 0.70 to IF 0.87 — all offered as guidance for the same distance. None of these targets has been validated against measured race outcomes in a controlled study; every one of them is practitioner opinion, however experienced the practitioner.

The one point of genuine agreement across all of them is directional, not numerical: the target should scale down as expected bike duration goes up — which is exactly what the fade data in this article would predict. There is a small amount of validated training-side context worth adding here too: across an Ironman season in nine triathletes, more training time in the easiest zone correlated strongly with faster racing (r = −0.92) and more time in zone 2 correlated with slower racing (r = 0.94), though a sample of nine makes those correlations unstable, and it describes training distribution rather than the race-day target itself.

The bike-run trade-off is still an open question

It would be convenient to say “so hold your power dead level and everything downstream sorts itself out.” The evidence doesn't support that conclusion either way. Two studies that directly tested constant versus variable bike power in triathletes and then measured the run point in opposite directions: constant power produced a faster subsequent 5 km run than variable power in one trial, while variable power produced a longer time-to-exhaustion on the run in another.

Both studies used ten or fewer male athletes over short protocols, and there is no study at 70.3 or Ironman distance that has randomised bike pacing and measured the resulting run split. Any confident claim about the single correct way to distribute effort across a long-course bike leg is, at this point, informed opinion rather than demonstrated fact.

Practical application

None of this makes Intensity Factor useless — it means treating a single IF number as a fixed instruction for the whole race asks more precision of it than the physiology can support. Some things you can act on today:

Common mistakes

How to apply this week

You don't need a lab to act on any of this before your next race, and most of it fits inside a normal 6–10 hour training week:

  • On your next long ride (2.5+ hours), pick a target power and compare your heart rate against it in the final 30–45 minutes versus the first 30–45 minutes. A widening gap at the same power is your own field version of the decoupling signal described above — use it to sanity-check whichever%FTP number you've been handed.
  • Practice fuelling at the upper end of your tolerated carbohydrate rate on that same long ride, not just on race day. The critical-power data above only applies if the fuelling behind it is something you can actually execute without GI distress, and that takes gut training over weeks, not a single race-day decision.
  • Write your race target as a range that steps down through the ride — a slightly higher number for the first hour, a slightly lower one for the last — rather than a single flat IF. This costs nothing to plan and matches the direction every practitioner source agrees on.
  • If you're deciding between two published%FTP targets that disagree, default to the more conservative one for your first attempt at a given distance, and treat the more aggressive end of the range as something to earn once you have race data showing you can hold it into the final hour.
  • Re-read your last long-course race's power file, if you have one, and check whether your back-half watts fell more or less than the 15% measured in the Abbiss data. That single comparison tells you more about your own durability than any single coaching percentage will.

References

Frequently asked questions

What is Intensity Factor (IF) and why does it matter for triathlon race pacing?

Intensity Factor is race power divided by your fresh, rested FTP — a shorthand for how hard you're riding relative to threshold. It matters because every %FTP race target you've ever been given is built on that same fresh-FTP denominator, even though the evidence shows that threshold falls during a long bike leg, so the effort the number represents changes as the race goes on.

How much does bike power actually fall during a long triathlon?

In six well-trained male triathletes who finished in the top 10% of an Ironman, measured mean power fell from 239 ± 25 W to 203 ± 20 W across the bike — a 15% drop (P < 0.05) — alongside falls in cadence and speed. It's a small, dated field study, but it's the clearest direct measurement available of how much a race power target can erode.

Why would the exact same power target be harder to hold in hour four than in hour one?

Because the ceiling it's measured against (critical power or FTP) falls with fatigue, and how much it falls depends partly on fuelling. Using published means from a controlled trial, 208 W (0.75 of a 277 W fresh critical power) works out to 88% of remaining critical power after three unfuelled hours, versus 78% after three hours at 120 g/h of carbohydrate. This is our own arithmetic applied to that study's published means, not a finding the study itself reported — it's an illustration of the mechanism, not a measured result.

What Intensity Factor should I actually target for a 70.3 bike leg?

There's no single validated answer. Published coaching guidance ranges from IF 0.70 to IF 0.87 for the same distance, and every one of those numbers is practitioner opinion rather than a value tested against measured outcomes. The one point of general agreement is that the target should scale down as your expected bike duration goes up.

Does the threshold-drift problem apply to sprint-distance racing too, or only long course?

It shows up even at sprint distance. A simulated sprint-triathlon cycling bout in junior/U23 male triathletes produced measurable declines in fatigued-state running measures, so treating durability as a long-course-only concern isn't supported by the evidence.

Should I ride at a dead-level, constant power to protect my run split?

The evidence doesn't clearly support that either way. Two small triathlete studies that directly tested constant versus variable bike power produced opposite results for the subsequent run, and neither was conducted at 70.3 or Ironman distance. Any confident claim about the single best way to distribute bike effort at long-course distance is currently informed opinion, not demonstrated fact.

Get The Forward

Research-led triathlon guidance, straight to your inbox. No spam, no filler — unsubscribe any time.