Thursday, April 22, 2010

IM Bike Execution: Using RPE, Power and HR


Racing to the best of your ability in an Ironman is not about being fast. It’s about being smart. It’s about maximizing knowledge and tools at your disposal to turn out your best performance possible. On the bike portion specifically, it boils down to effective planning. With several years of IM races, including Kona, under my belt, as well as prior triathlon coaching experience, I’ve found success with the following approach to developing an effective IM bike plan. The framework is based on a philosophy of racing against yourself rather than others. It’s about focusing on the variables you can control and ignoring the ones you can’t control. Some people have described it as “racing in a box.” Its basis for execution comes from writing a pre-race plan that is primarily strategic (i.e., static) in nature but tactical (i.e., dynamic) enough to adapt to unforeseen circumstances or changing conditions.

The ideal race-execution plan, in my opinion, leverages three major tools – 1) rate of perceived exertion, 2) power and 3) heart rate. Rate of perceived exertion is an internal gauge that racers must analyze on their own, while the bulk of this article will cover more easily measured – and manipulated -- power considerations. I will also touch briefly on heart rate, which is often overlooked in IM when using power.

One important issue that I will not be addressing is nutrition/hydration. This specific area is highly individualized and merits a separate article. The following information presupposes that your nutrition/hydration needs are being adequately addressed leading up to and during race day.

Rate of Perceived Exertion (RPE)

While power and heart-rate tools provide objective measurements, RPE is harder to quantify. Regardless, it is probably the primary meter for determining proper pacing on the IM bike. RPE is simply that “feeling” of how hard you believe your body is working. There have been several attempts to quantify RPE with the use of scales, but I’m not here to tell you the IM bike should be done at a specific number on Borg’s scale. IM RPE is something that is fine-tuned through hours and hours of long ride training. Leveraging this experience can be the most significant tool in helping you determine what RPE is appropriate for the IM bike. Unfortunately, race-day excitement and ego can have this strange psychological impact on your pace-judgment skills. In addition, even for those who are known to be well-disciplined overall, RPE can still be very deceiving at times. For example, how many of you have felt totally fresh in the first 20 or so miles yet have been guilty of riding well above the effort you could sustain for 112 miles? That type of effort will always come at a cost later in your race.

I believe the most significant lesson to learn regarding RPE is to know when you need to listen to it versus when you need to ignore it. The general rule of thumb is to always ignore your RPE when it’s telling you to go harder and use it as the primary meter when it’s telling you to slow down. In other words, even if your current power is slightly below your target but your RPE is telling you to slow down, you should listen to your RPE in this case.


Four elements define our ability to determine proper pacing using power:

1. Functional Threshold Power (FTP)

The concept of FTP was introduced via the Normalized Power Concept and is the foundation for the framework. If you’re not familiar with the NP Concept, then I highly suggest you take some time now to learn the concept before continuing forward. The best place to start is at the following Web site:

However, in case you want to get through this article first before reading the above, I’ll provide some quick definitions here:

• Normalized Power (NP) – An algorithm that provides a better measure of the true physiological cost or training intensity of a given workout than average power provides.

• Intensity Factor (IF) – A ratio of NP to FTP which allows the differences in fitness between individuals to be taken into consideration.

• Training Stress Score (TSS) – A value that quantifies overall training load or the physiological cost of a ride by taking into consideration the interaction between training intensity and volume/duration.

Establishing your FTP is the pragmatic or functional approach to determining power at lactate threshold (LT). Power at LT is one the most accurate predictors of endurance performance ability. By definition, FTP is the highest power you can maintain in a steady state (or quasi-steady state) for a prolonged period of time (i.e., ~1 hour) without fatiguing. Fatigue is indicative of your inability to sustain target power, which is evident in your power profile. For example, more than a 2-percent drop in power anywhere after the first ~10 minutes is evidence of fatigue.

There are several well-documented ways of establishing your FTP (see: or

You can use any one or multiple ways to determine your FTP for training, but there are additional factors you should consider when fine tuning your FTP for IM execution. Those factors are:

• An emphasis on race specificity (i.e., low-variable power in your race position at race cadence)

• An emphasis on 4 to 6-hour Maximum Power (MP) from training

• An emphasis on 30-minute and 60-minute MP from race data (e.g., Sprints and Olys)

Note: The evaluation of MP should always be based on normalized power, not average power. In addition, by stating MP I’m not suggesting you’re required to do a 4 to 6-hour TT. However, there’s an IM proverb to keep in mind, “The IM bike should be your easiest long ride of the year.” In other words, your long rides should be tough and challenging, e.g., >IF = .77, but not so hard that you can’t run about 30 minutes off the bike.

Being able to sustain high relative power in a reasonably aggressive aero position for 1 hour is one thing. Being able to sustain high relative power in a reasonably aggressive aero position for 5-plus hours is a completely different challenge. So, if you haven’t spent an appropriate amount of time doing race-specific preparation training, then depending solely on short-interval power testing to establish your FTP might lead to some disappointment on race day. Therefore, the evaluation of your long ride data from training can help you determine if the lack of specificity training might be a concern. There’s a concept known as “Raise the left, fill the right,” which is in reference to your power-duration curve. “Raise the left” refers to an emphasis on raising your FTP, and “fill the right” refers to an emphasis on maintaining high relative power for an extended period of time. As stated earlier, that extended period of time is around 4 to 6 hours, and high relative MP would be considered to be an IF of .75 at a minimum – ideally higher. So if your 5-hour MP is less than .75 then a more conservative estimate of your FTP would be recommended, and possibly a greater emphasis on specificity training during your long rides is required.

Note: How you train will certainly have some influence on your typical long-ride power. For example, athletes who train something like ~20hrs/week on the bike might have relatively lower IFs for their long rides. In addition, those who place emphasis on high-intensity interval training during race prep might tend to ride easier during their long rides. However, in either of those situations I think it's important for the athlete to take at least a couple of weeks to see how comfortable they are doing some 4 to 6hr long rides at IFs > .77 and still not be too shelled to run off the bike for ~30 minutes.

When field testing, some people have a tendency to choose the position in which they can produce or sustain the most power for an interval, and that’s not necessarily their race position. That’s probably fine for establishing power training levels, but your goal for racing is to make sure your FTP is truly reflective of power you can sustain in your race position. All FTP testing should be done in your race position.

What I’ve discovered is that the longer you’ve been training and racing with power, the less dependent you need to be on FTP field testing. Race data can often provide you with a more accurate reflection of your FTP than field testing, since a lack of motivation is rarely a factor when racing. I know many people, including myself, who dearly struggle to get properly motivated for an FTP field test.

An appropriate pace for the entire IM bike is based on a percent range of FTP. You can think of this target as your “power reserve.” Please see: for more detail on the concept of power reserve. More importantly, real-time execution based on a pre-race plan is the best approach to achieving this target. Your pre-race plan should include real-time power targets for the flats, short climbs (includes rollers) and long climbs, which are all defined independently of wind conditions.

Before I define these real-time power targets for different course terrain (to be addressed in the “Execution” section below), I need to discuss an appropriate target for the entire IM bike, since it’s the basis for determining your real-time power target on the flats.

Personal experience, extensive research and work done with other coaches have determined an appropriate target for the IM bike to be between an IF of .67 to .78. The reason why the range is so wide is because we need to consider the amount of time you’re on the bike in addition to the three other factors described below (in sections 2, 3 and 4).

Our research also determined that a TSS <290 is the appropriate amount of stress to allow a well-trained athlete to have a good IM run. In other words, we start to see a steady increase in the number of athletes blowing up or slowing down significantly on the run when they have accrued a TSS >300.

Note: I’m defining greater than 75% T pace (Daniels) as a good IM run. However, my personal belief is that 80% of T pace is probably very close to an IM athlete’s true potential.

Using the chart above we can find the appropriate TSS target zones for a range of projected bike splits (between 4:30 and 6:30). We can now see why an athlete can potentially yield IFs as high as .77-.78, in theory, on the IM bike and still be in a good position to run well. If you’re on the bike for <4:45, an IF >.75 is still quite rare and is much more the exception than the rule. In addition, for those who bike >6:30, time is simply working against you. My experience says that you’ll likely execute similar run times for an IM bike leg yielding an IF of .67 versus .70. Unfortunately our research doesn’t have very good data for athletes who ride >6:30 since the athlete is often doing some amount of walking during the run in either case.

Lastly, for a broader perspective on the usefulness of TSS, please see the following link:

Now that I’ve explained how the use of TSS/IF/NP is ideal for establishing the physiological cost of the ride, I need to explain how you can achieve the fastest bike split for a given physiological cost. The amount of variability in your ride is critical to achieving the fastest bike split. Variability is determined by dividing normalized power (NP) by average power (AP). This is known as the Variability Index (VI). Modeling has easily determined that the highest AP for a given NP on a completely flat course will yield the fastest time. However, it’s a bit trickier when we’re dealing with a rolling or hilly course. Some amount of variability is required -- the question is how much. The following graph demonstrates the optimal power for a given gradient for an athlete who is targeting ~205w in the flats:

Note: This graph is the result of extensive analysis performed by several individuals. Everyone who has performed this modeling arrives at the same result. Keep in mind that the model doesn't understand that the athlete is limited in gearing. For example, it's quite typical to run out of gears around 35mph.

If you were to apply power as prescribed in this graph on one of the hillier IM NA Sports course out there (e.g., IM Wisconsin, IM Lake Placid), your VI would be no greater than 1.04-1.05. Remember though, this is what modeling has determined. There are other factors to consider. These factors can include wind conditions and how many bikes are around you on the course. VIs in the area of 1.04 on most any course would be considered a solid steady ride, and assuming these rides yield an IF of less than .75, they almost always have the following characteristics:

• Not riding the hills too hard

• Not coasting too much

• Very few power spikes or quick surges

Note: I will quantify “too hard” and “too much” within the “Execution” section below.

In other words, if you were to analyze the power data of a well-executed IM bike performance in small 30-second chunks over similar course terrain, it would look very consistent. However, do not confuse a low VI with being consistent across the entire ride (i.e., even power splits between first and second half). It’s still possible to have a low VI with a declining power profile. We often see this behavior by professional racers at IMH due to a tactical need to keep the race “in touch” or to catch the draft-legal train toward the front of the pack.

2. External Conditions

As I stated above, a fundamental principle behind the framework is that it’s based on conditions we can control and not ones we can’t control. We can’t control the weather and we can’t control other individuals. I don’t recommend basing a pacing decision on things you can’t control because you’ll be dynamically adjusting your pace throughout the race and find yourself in this constant battle of trying to manage your resources effectively. If this was an open road or a multi-stage race, then we’d be having a different conversation, of course, but an IM is a (non-drafting) time trial with a marathon to follow and has a very unique set of requirements.

The subject of how you should adjust power based on wind conditions is often discussed on Tri forums. We know that increasing your power 2% to 3% in a headwind is slightly faster, but it’s only by a very small amount of time. The problem is that you now have to balance that increase in power with a decrease somewhere else on the course. You can’t simply increase your power by 2% to 3% for an hour when you’re riding into a headwind and not expect to pay the price later on. People often assume that if there’s a headwind on part of the course, then there must be a tailwind somewhere else. Of course, we all know this is not necessarily true. In addition, wind is always unpredictable. How many times have you thought you were riding into a headwind only to make a turn and come to realize you’re now riding into a stronger headwind? Also consider the fact that a windy course is almost never faster than a windless course, which means you’ll be on the bike longer than usual. If that’s the case, then you actually need to consider slowing down since training stress (TSS) is climbing the longer you’re on the bike. So, the recommendation is to just sit on target power for a given type of terrain or possibly consider decreasing power slightly to meet the appropriate TSS target as described above. In my mind, trying to adjust power based on wind conditions is an attempt to be smarter than Mother Nature. I believe you’ll lose that battle more often than not.

3. Fitness Level

Your expectations must be set realistically. This is a tough one because so many people overestimate what they’re capable of achieving in the near term (e.g., at their next race), yet they’re constantly underestimating what they’re capable of achieving in the sport long term. Setting realistic expectations is about having a thorough understanding of what we’ve accomplished in training, race rehearsals and testing. Just because you can hang with a guy on an occasional long ride who typically does a sub-5:15 IM bike doesn’t mean you should set your expectations accordingly. Base your expectations on your own results – nobody else’s. For example, I needed to run in the 3:30s in at least two IMs before I set my expectations as a 3:30 IM runner. Use a self-assessment of your fitness level to help determine whether you should target the upper or lower end of the TSS range.

4. Mental Strength

The most difficult factor to quantify above is mental strength. I’m betting we all like to think we’re mentally strong, but the truth is that many IM runs come down to one’s ability to “hold it together” those last 6-8 miles. In other words, the difference between walking and running beyond mile 18 can be dependent on your ability to simply suck up the pain. There are no exceptions – everyone is in pain. Yes, some are certainly in greater pain than others, but you need to take a hard look at your previous performances to see if mental strength is an issue when determining an appropriate power target. Humility is your friend in this case. Use a self- or coach-assessment of your mental strength to help determine whether you should target the upper or lower end of the TSS range.

Heart Rate (HR)

There’s a common belief that HR is not required when you’re racing with power. I do believe there’s a fair amount of redundancy when using an HRM along with a power meter. However, ignoring HR in a four-plus-hour event can be a potential mistake. HR is an indirect indicator of cardiovascular (CV) stress. Under most conditions, CV stress levels have a high degree of correlation with well-established power levels. These levels vary from individual to individual, but how they correlate is an important process to understand during your training. The argument often made against the use of HR refers to the times when CV stress is not indicative of your typical corresponding power. These cases usually involve loss of fitness, heat, dehydration and fatigue. However, going into your IM, some of these situations aren’t a concern, given that you should be well-rested and at peak fitness. So, if you think about it, the remaining situations are also conditions in a long-endurance event where you would normally slow down or react accordingly (e.g., drink more fluids). Completely ignoring your HR isn’t usually a problem in events lasting less than three to four hours, but I do believe that high (relative) CV stress can be a huge problem in long-endurance events such as IM. High CV stress is indicative of an issue within the human body, which it can probably handle for up to a certain amount of time. So, allowing the body to become dehydrated, for example, for an hour well into a four-hour event is one thing. However, allowing the body to become dehydrated four hours into the IM bike with a marathon to follow is a completely different story. Lastly, high HR levels tend to lead to digestive problems due to blood being shunted away from the intestines. This leads to slow absorption.

Below I will address defining meaningful HR caps and what needs to be done when your HR is high or above a cap, even when your RPE and power are well within your guidelines.


The goal is to write a race plan which addresses all of the components I described above. This detail will be the basis for execution during the race. Since RPE is based on your ability to interpret how your body is feeling, I don’t put specific guidelines or descriptors for RPE into my race plan. I trust my ability to pace by RPE for much of the race, but there are specific sections of the course or times during the day where I watch my power and HR like a hawk (with a stronger emphasis on power). Every athlete will have to assess their confidence in their own RPE, which might lead to a greater emphasis on power and HR. So much depends on our level of experience, but this issue is just one good reason why so many of us hire coaches.

I define my race plan specific to the IM bike as follows (using myself at IMC this year as an example):

FTP = ~270 watts (w); Race weight = ~145 pounds

Overall Goal

Projected bike time: ~5:10; NP = ~205w; AP = ~195w; ~75% of FTP or IF = ~.75; TSS = ~290

Build power slightly throughout the ride. Target first half at slightly lower power and then evaluate conditions and how I’m feeling to see if I can confidently build in the second half

RPE provides primary guidance; power is secondary; HR is tertiary

- If RPE is high, then decrease power even if power and HR are fine

- If RPE is fine but power is high, then decrease power even if your HR is fine

- If RPE and power are fine but HR is high, then assess accordingly (as described below in “HR Guidelines”)

Note: I base projected bike time from past IM bike experience and long ride data. This assumes good bike course conditions. Others will have to leverage experience from race rehearsals in addition to their long ride data if they have no past IM experience. They should also consider using a wider range of projected bike times.

One reason why I like to target slightly lower power in the first half and build in the second half of the ride is because it allows more flexibility when conditions change, which they often do. In addition, in situations where you are unsure of an appropriate TSS/IF target, then this approach allows you the ability to start more conservatively and just maintain power if you feel like you possibly over-estimated your original target during the race. Lastly, I find that your digestion system works more effectively when you start “slow” and build. Some people might consider this approach to be more advanced. An alternative approach for the average age grouper would be to maintain, not build, power throughout the ride.

Power Targets by Terrain

• Flats: Real-time power @ 205w (same as overall NP target); Target 195 – 200w (.73-.74 IF) for the first ~56 miles

• Short Climbs: Real-time power @ 230 – 245w (85% to 90% of FTP)

• Long climbs: Real-time power @ 215 – 230w (80% to 85% of FTP)


• Increase power slightly when cresting the hill; maintain power as speed increases up to ~30 mph, then start decreasing power and move to a soft pedal between 32-37 mph before coasting at greater than ~37 mph (or running out of gears)

Note: Handling the descents on a consecutive set of rollers can be a bit trickier and involves much more “art,” but the general recommendation is to continue to follow the above approach. Often speed will not exceed 35 mph, so you should do the best job at maintaining power between 60% of FTP on the downhill sections and 90% on the uphill sections.
HR Guidelines

• HR caps for long climbs: ~10bpm over your overall average or equivalent to your Half Ironman average HR (whichever is lower)

• HR cap for the flats:

- First 20 minutes – Same as your cap for long climbs. RPE will likely be deceiving so you want to get your HR down to your target as soon as possible. Consider making HR a higher priority than RPE or power during this section

- Remainder of the ride – Between your race rehearsal (RR) average heart rate (AHR) and RR AHR+ 5bpm. This assumes your RR was done at a time when you were well-rested and was preceded by a long open water swim

Note: The assessment of what to do when HR is high can be difficult at times. One of the challenges with HR is that most people tend to have different ratios between training power/HR and racing power/HR. This delta is often referred to as a phenomenon known as “race-day excitement.” Two other things that could impact your HR are:

• Quick transition from the swim

• The effects from tapering – You’ve shed all or most of the fatigue from your previous race-prep training, and changes in blood plasma volume may have taken place

Those who have a couple of IMs under their belt can gain some insight into those differences, but those performing their first IM will have to leverage experience from RRs/training and play it by feel a bit. I prefer an HR cap for the flats as opposed to an HR target since RPE and power should be your primary and secondary drivers.

If it’s very early in the ride and HR is high, it’s most likely due to the swim and swim-bike transition, so I’d give it some time to settle. However, unless your power is on the conservative side, it certainly wouldn’t hurt to back it down a bit either. If heat is not a factor and HR is still high, then the onset of dehydration should be considered. It’s probably best to put some heavy emphasis on getting extra fluids (mostly water) into the body over the next hour. If HR doesn’t come down to a more reasonable level, then consider decreasing power slightly. Only advanced IM athletes should consider ignoring their HR at this point. If heat is a factor and HR is high, then consider the strong possibility you might be out on the course for a longer period of time. Success in the race will be solely based on the ability to run well in adverse conditions. Making a decision to play it conservative on the bike because of the heat so that you can save some (physical or mental) energy for the run is usually the difference between a solid IM run and a very disappointing one.

Mental Strength Reminders

• There’s a solution to every problem or challenge that faces you during the day – never, ever give up on finding a solution

• You need to associate the challenge of finishing a race with pleasure vs. pain. It’s simply a focus on potential pleasure over current pain

• The pain is temporary and the faster you get to the finish line the sooner that pain will go away – so keep moving!!

• Focus on the things in your life that are most important to you. If you think it would help, tape a picture of your loved ones to your aerobars


The following people provided significant data, editing or contributed content to this article: Scott McMillan, Shawn Burke, Rick Ashburn, Mark Van Akkeren, Jason Digman, Kurt Perham, Rich Strauss and Beppo Hilfiker ( Oh, of course, Dr. Andrew Coggan and Alex Simmons too.


Buddha Belly said...

Wow! Great informative article. I just bookmarked it and will definitely be referencing it in the future.
I just started training with power 2 weeks ago so I've got lots to learn. Hope to use it to its full potential as I train for IM Moo in 2011.

Unknown said...

Hi Chris,

This is an excellent post! One thing really stood out for me was the TSS for the ride being <290. Our stats were pretty similar going into IMCA-2010, with me being about 10-lb heavier, but similar FTP (270). I just reviewed my TSS for that ride and it was 373 (IF=0.83); AVG/NP = 206/222. My run was a PR (as was the day), so I am pretty sure I did not over-do it. Perhaps I am an anomaly, but I am curious as to your thoughts on this.

Chris Whyte said...

Thanks for the encouraging words!

I'll admit that I've looked at a ton of IM data over the past 8 years and an IF = .83 is definitely a number I have never/rarely seen. Pros can often pull off IFs around .80 but even they do it less often than you might think. The one problem with using TSS < 290 as a guideline is that it starts to break down for those who are on the bike for well over 6hrs. Although, that wasn't the case for you.

I would take a deeper look into how you were determining your FTP at the time. First question would be: what was your NP/AP for your typical long ride during that period? Any '10 HIM or short course data you could share too? Could you also share your T pace at the time and your actual run pace for the race. I don't want to assume that a PR run = you ran to your full potential.


Bob McRae said...

Sorry for the delayed response Chris (been training).

I have an extremely flat mean maximal power/pace curve. While I am not fast, I do not slow down a whole lot. As an example, my HIM run PR is ~1:32 & IM (IMCA-10) was ~3:24, an 11% drop in pace. Now, had I executed better, it would have been a 3:17, which represents a 7% drop in pace. Compare that to the pro's who run HIM @ ~1:12 vs. IM @ 2:45, which is a 15% drop in pace.

As to address your query about a 2010 HIM, I have just one decent power file (Boise), which was actually a pretty low NP of 225W. It was a train-through race and I rode 50+ miles the day before. (Oddly enough, it was by best placed position at any major race.) I normally ride NP = ~235-240W in HIM's.

My T-pace was about 6:30/mile at that time and my IMCA-10 pace was 7:47/mile (with 2x potty breaks and miles 20-25 @ 8:00+ due to 7:20 miles from 13-20). So, as you can see I ran at about 84% of T-pace. Had I paced better, it would have been closer to 89% of T-pace.

So, as you can see, my power/pace profile is really flat, which means IM is for me:)