Assessing endurance performance – The value of field testing

We have recently launched a short course entitled, “Cycling Science: The essentials of cycling physiology and coaching”. In this article, we are covering some of the content from the second module which focuses on assessing endurance performance. For further information about the course, please click HERE.

Field testing.jpeg

 

Performance testing is often implemented to assess adaptation to a training programme. We have previously written a more detailed article on laboratory-based performance tests, which you can find here. Laboratory-based assessments require the use of expensive equipment, are time-consuming and can be disruptive to a cyclist’s training programme. The increased affordability of power meters has resulted in an increase in the collection and analysis of field test data. However, when field test data is being collected and analysed, the following points should be considered;

 

1. Is the testing protocol valid?


A test is valid if it correctly measures what it claims to. In other words, a performance test for a cyclist should produce data that can accurately predict cycling performance. For example, a one repetition maximum bench press will have little value for predicting cycling performance. There is some debate as to whether laboratory-based assessments are valid measures of endurance cycling performance, but the strong association between the variables; VO2max, threshold, cycling economy and peak power out, and cycling performance is hard to argue against.

 

2. Is the testing protocol reliable?


Reliability refers to how repeatable the results are. In other words, if we repeated the same testing protocol, on the same athlete, with the same equipment, under the same conditions, we would expect the same results. Every test will have a certain amount of measurement error, and for a test to be considered reliable, the error or measurement must be low. Field testing is less reliable than laboratory-based assessments due to exposure to factors that could affect the outcome of the test (environmental temperature, road surface etc.).

 

3. Is the test sensitive to change?


If the cyclist’s performance has improved, the data collected from the test should reflect that.

 

Field tests can take numerous forms, but the most commonly used field tests are hill climbs, time-trial (TT) efforts of varying duration or distance, and standardised training sessions. The 20 minute time trial is commonly used to determine a cyclist’s functional threshold power (FTP). There is a strong association between the average power output during a 20 minute TT and a cyclist’s threshold. If you are interested in reading more about this link, we recommend you read this article. Performing a TT effort requires a quiet stretch of road, or a track that will ensure that your effort is not interrupted. The effort in the figure below was performed on Franschhoek Pass by a young road cyclist. Following a sufficient warm up, the cyclist hit the lap button and held a relatively constant power output until the timer reached 20 minutes.

 

20 minute time trial output.png
Figure 1: 20-minute time-trial effort up Franschhoek Pass Cadence (Blue), Power output (Purple), Heart rate (Red).

 

training data table.png

 

Pacing during these longer TT efforts is an important consideration and can have a large effect on the result of the test. The importance of pacing is probably best illustrated in the two figures below. Figure 2 shows a well-executed time-trial with a relatively constant power output (green line), compared to Figure 3 where the power output is highly variable. Cyclists should be familiarised with these type of efforts in order to assist them in improving their pacing strategy and ultimately performance.

 

well executed time trial.png
Figure 2: A well-executed time-trial.

 

poorly executed time trial.png
Figure 3: A poorly executed time-trial.

 

Field tests can also be disguised as standardised training sessions, such as a set of 3 x 10 minute intervals. An increase in average power during these intervals could be valid measures of improvements in cycling performance. These field tests can track changes in performance and supplement the laboratory data which is collected less often. The data in the set of graphs below was collected during a training block aimed at improving this cyclist’s sustainable power output, and each session involved a set of 10 minute intervals. Each date represents a single training session, each dot represents a single interval and the green dashed-line represents the average for that particular session. The cyclist was able to increase their 10 minute power output (session average) as they progressed through this training block. A simple training session that is repeated at regular intervals during a training block can serve the role of a field test and provide an indication of the cyclist’s training progression. Tracking the variables of cadence and heart rate provides further insight into how the session was executed and the fatigue levels of the cyclist.

 

data from standard training session.png
Figure 4: Data from a standardised training session that can be used as a field test to track progression.

 

 

 

 

 




10 Comments

Bigjim, May 24 2018 12:16

Thanks for an interesting article S2S. I have a few queries to raise please.

 

Looking at your 10 min interval test data, I noted that the timeframe between the tests was quite short - is it realistic to expect & even check for performance gains within such a short timespace? I also noted that the average HR values do drop in the second test, but increase in the 3rd. To my mind, the 2nd test indicates a familiarization with the test protocol & associated efficiency gain, & the 3rd test that perhaps the athlete was able to deliver a stronger intensity effort for the test -albeit at the expense of probably burning more matches (and consequentially reducing his longer endurance capability). I haven't looked all that closely at the numbers, but surely a better measure of performance progression would be a higher power output at the same HR values - or some form of correlation between the two?

 

Apologies in advance for the amateur analysis - I am just trying to gain a better understanding of the concept you are illustrating here.   

milky4130, May 24 2018 02:56

Thanks for an interesting article S2S. I have a few queries to raise please.

 

Looking at your 10 min interval test data, I noted that the timeframe between the tests was quite short - is it realistic to expect & even check for performance gains within such a short timespace? I also noted that the average HR values do drop in the second test, but increase in the 3rd. To my mind, the 2nd test indicates a familiarization with the test protocol & associated efficiency gain, & the 3rd test that perhaps the athlete was able to deliver a stronger intensity effort for the test -albeit at the expense of probably burning more matches (and consequentially reducing his longer endurance capability). I haven't looked all that closely at the numbers, but surely a better measure of performance progression would be a higher power output at the same HR values - or some form of correlation between the two?

 

Apologies in advance for the amateur analysis - I am just trying to gain a better understanding of the concept you are illustrating here.   

It is not tests per se, its 3 different training sessions that had 10min intervals in them. Higher power in the last interval will = higher heart rate compared to the previous 2 intervals. If the intervals were performed at the same power average over the 3 weeks, then I would expect the HR in the 3rd interval\week to be lower, at least that's according to my understanding.

SciencetoSport, May 24 2018 03:18

Thanks for an interesting article S2S. I have a few queries to raise please.

 

Looking at your 10 min interval test data, I noted that the timeframe between the tests was quite short - is it realistic to expect & even check for performance gains within such a short timespace? I also noted that the average HR values do drop in the second test, but increase in the 3rd. To my mind, the 2nd test indicates a familiarization with the test protocol & associated efficiency gain, & the 3rd test that perhaps the athlete was able to deliver a stronger intensity effort for the test -albeit at the expense of probably burning more matches (and consequentially reducing his longer endurance capability). I haven't looked all that closely at the numbers, but surely a better measure of performance progression would be a higher power output at the same HR values - or some form of correlation between the two?

 

Apologies in advance for the amateur analysis - I am just trying to gain a better understanding of the concept you are illustrating here.   

Hi Jim. 

Thanks for question and no need to apologise for your analysis.

It is important to note that these are training sessions that were prescribed to a cyclist during a particular training block. We analyse the training data as shown in the graph in order to track progression of the cyclist during a block/season. They are not field tests, but can serve the role of a test due to the data that is produced. 

We would like to address your questions individually, so apologies in advance for the repetition of the text above.

 

1. Is it realistic to expect & even check for performance gains withing such a short timespace?

The typical scientist answer of, "It depends." applies here too. After the first training on the 27th of December 2016 we would have analysed the session and determined that the cyclist averaged 385 Watts for all 3 intervals (session average). Our instruction to the cyclist on the 30th of December 2016 would have been to try and increase the session average by 10 Watts (note that the second session had 5 intervals compared to the 3 in the first). This instruction provides a carrot or target that the athlete can aim for during the second session. Despite the extra 2 intervals, the cyclist still increased their session average to ~392 Watts. If there was no increase, we might have looked at the acute and chronic training loads and if they were excessive, an extra easy day or a rest day may have been prescribed. We always expect an improved performance, but not every session can be a personal best, and if the next session isn't as good as the previous, it is important to consider why that may be, adjust the training accordingly, and then proceed. Stay tuned the the Bike Hub, as we will be writing an article on monitoring in a few weeks. 

 

2. I also noted that the average HR values do drop in the second test, but increase in the 3rd. To my mind, the 2nd test indicates a familiarization with the test protocol & associated efficiency gain, & the 3rd test that perhaps the athlete was able to deliver a stronger intensity effort for the test -albeit at the expense of probably burning more matches (and consequentially reducing his longer endurance capability). 

 

The drop in heart rate values from the first to the second session is interesting. The decrease in both average and maximum heart rate in the second session might be due to the accumulation of some acute fatigue. Heart rate responds slower when fatigue is higher. Clearly the fatigue was fairly acute, because the heart rate indices increased in the third session which was approximately a week later. The variable that is missing from this analysis would be the cyclist's rating of perceived exertion. If the heart rate and power output were lower, but the effort felt harder, that is a clear sign of fatigue and the cyclist would have needed some rest or easy days. 

 

3. I haven't looked all that closely at the numbers, but surely a better measure of performance progression would be a higher power output at the same HR values - or some form of correlation between the two?

This is the basis of a submaximal test that we use to track fatigue and progression, but we include the cyclist's perception of effort. In the example presented in our graphs, the instruction would have been to give your best 10 minute efforts for the training session. We included power output targets, but we did not anchor the heart rate or power output, therefore the analysis you proposed may not have been appropriate. If we had locked the heart rate at a certain intensity, for example 168 bpm, and then monitored the power output at that target heart rate, we would be comparing apples with apples, because we had anchored the intensity (168 bpm). However, as mentioned above, acute fatigue can reduce heart rate, so what if the cyclist can't reach the target of 168 bpm? Perhaps anchoring power output and monitoring heart rate and perception of effort may be more appropriate, provided the target power is not too high. We are currently investigating this. If you are performing a similar analysis (relationship between heart rate and power output), we would encourage you to include a measure of perception of effort (subjective rating of intensity). That will improve the sensitivity of your monitoring. 

 

Apologies for answer your question with another article Jim, but we hope this adequately answers your questions. 

Please let us know if anything requires further clarification. 

Bigjim, May 24 2018 03:30

Hi Jim. 

Thanks for question and no need to apologise for your analysis.

It is important to note that these are training sessions that were prescribed to a cyclist during a particular training block. We analyse the training data as shown in the graph in order to track progression of the cyclist during a block/season. They are not field tests, but can serve the role of a test due to the data that is produced. 

We would like to address your questions individually, so apologies in advance for the repetition of the text above.

 

1. Is it realistic to expect & even check for performance gains withing such a short timespace?

The typical scientist answer of, "It depends." applies here too. After the first training on the 27th of December 2016 we would have analysed the session and determined that the cyclist averaged 385 Watts for all 3 intervals (session average). Our instruction to the cyclist on the 30th of December 2016 would have been to try and increase the session average by 10 Watts (note that the second session had 5 intervals compared to the 3 in the first). This instruction provides a carrot or target that the athlete can aim for during the second session. Despite the extra 2 intervals, the cyclist still increased their session average to ~392 Watts. If there was no increase, we might have looked at the acute and chronic training loads and if they were excessive, an extra easy day or a rest day may have been prescribed. We always expect an improved performance, but not every session can be a personal best, and if the next session isn't as good as the previous, it is important to consider why that may be, adjust the training accordingly, and then proceed. Stay tuned the the Bike Hub, as we will be writing an article on monitoring in a few weeks. 

 

2. I also noted that the average HR values do drop in the second test, but increase in the 3rd. To my mind, the 2nd test indicates a familiarization with the test protocol & associated efficiency gain, & the 3rd test that perhaps the athlete was able to deliver a stronger intensity effort for the test -albeit at the expense of probably burning more matches (and consequentially reducing his longer endurance capability). 

 

The drop in heart rate values from the first to the second session is interesting. The decrease in both average and maximum heart rate in the second session might be due to the accumulation of some acute fatigue. Heart rate responds slower when fatigue is higher. Clearly the fatigue was fairly acute, because the heart rate indices increased in the third session which was approximately a week later. The variable that is missing from this analysis would be the cyclist's rating of perceived exertion. If the heart rate and power output were lower, but the effort felt harder, that is a clear sign of fatigue and the cyclist would have needed some rest or easy days. 

 

3. I haven't looked all that closely at the numbers, but surely a better measure of performance progression would be a higher power output at the same HR values - or some form of correlation between the two?

This is the basis of a submaximal test that we use to track fatigue and progression, but we include the cyclist's perception of effort. In the example presented in our graphs, the instruction would have been to give your best 10 minute efforts for the training session. We included power output targets, but we did not anchor the heart rate or power output, therefore the analysis you proposed may not have been appropriate. If we had locked the heart rate at a certain intensity, for example 168 bpm, and then monitored the power output at that target heart rate, we would be comparing apples with apples, because we had anchored the intensity (168 bpm). However, as mentioned above, acute fatigue can reduce heart rate, so what if the cyclist can't reach the target of 168 bpm? Perhaps anchoring power output and monitoring heart rate and perception of effort may be more appropriate, provided the target power is not too high. We are currently investigating this. If you are performing a similar analysis (relationship between heart rate and power output), we would encourage you to include a measure of perception of effort (subjective rating of intensity). That will improve the sensitivity of your monitoring. 

 

Apologies for answer your question with another article Jim, but we hope this adequately answers your questions. 

Please let us know if anything requires further clarification. 

Bigjim, May 24 2018 03:44

 

 

Thanks very much for the detailed reply S2S - I was not taking the fatigue component into consideration. I definitely now get how the athlete RPE inputs are an important addition - even though they are probably somewhat subjective. At the facility I train at, some tests are carried out with power output anchored & HR change tracked, so I found your article and comments very informative - looking forward to the next one 

cadenceblur, May 29 2018 09:31

Did my latest FTP and showed some improvement, but need some guidance on some of the metrics...
Decoupling came it 4.96% and VI at 1.00 EF at 1.65,
I maintained a really flat line in terms of power throughout the effort and cranked it up over the last three minutes or so. Despite the improvement, what do I make of these numbers? Could I perhaps have gone harder? 

SciencetoSport, May 29 2018 12:42

Did my latest FTP and showed some improvement, but need some guidance on some of the metrics...
Decoupling came it 4.96% and VI at 1.00 EF at 1.65,
I maintained a really flat line in terms of power throughout the effort and cranked it up over the last three minutes or so. Despite the improvement, what do I make of these numbers? Could I perhaps have gone harder? 

Thanks for the question CadenceBlur. 

Before we discuss your Training Peaks metrics, could you provide some more detail on how you tested your Functional Threshold Power (FTP)?

Was it a 20 minute time trial?

Indoors or outdoors?

Apologies for answering with more questions, but some more information on the test would help explain your metrics better.

Cheers

S2S Team

cadenceblur, May 29 2018 01:08

Thanks for the question CadenceBlur. 

Before we discuss your Training Peaks metrics, could you provide some more detail on how you tested your Functional Threshold Power (FTP)?

Was it a 20 minute time trial?

Indoors or outdoors?

Apologies for answering with more questions, but some more information on the test would help explain your metrics better.

Cheers

S2S Team

Hi

 

Sure, no problem...

 

It was a 20 minute time trial conducted indoors.

SciencetoSport, May 29 2018 02:16

Hi

 

Sure, no problem...

 

It was a 20 minute time trial conducted indoors.

Great! 

Let's start with the easiest variable and work our way through. Apologies for including the definitions of the variables, but we have done so for the benefit of other Hubbers who may not be familiar with these metrics.

 

Variability index (VI) is a measure of how constant your effort was and is calculated by dividing your normalized power by your average power. A well-paced TT will have a VI of 1.05 or less, which means that the power output did not vary too much. Consider Figures 2 and 3 in the article above. Figure 2 would have a VI closer to 1.00 (<1.05), while Figure 3 will have a larger VI (>1.05). 

Your VI of 1.00 indicated that you paced your effort really well. Twenty minutes is a relatively short effort, although I am sure it felt like a lifetime, so a VI of 1.00 is to be expected. I would also guess that you are familiar with these type of efforts, and no doubt that has allowed you to create a pacing template to make sure you don't go out too hard and pay the price later. Performing the session indoors will also improve your pacing as there are no undulations or traffic furniture which could interfere with your effort. 

 

Effeciency Factor (EF) is calculated by dividing your power output by your heart rate. Like all other heart rate related metrics, there are some important considerations:

  • A single EF will tell you nothing. The proposed value is in tracking the improvements in this number over time. Was your heart rate lower at the same power output? In other words, was riding at that intensity (power output) less of a stress compared to previously. This could be interpreted as an improved performance. 
  • Don't compare to other riders. Individual athletes will have very different heart rate kinetics, so only compare your EF to previous values. 
  • Heart rate can be affected by numerous factors. Heat is a big one. When we get warmer, we open the blood vessels in our skin (that is why we looked flushed). The increased blood flow to the skin allows us to cool our blood in order to prevent us from overheating. Your muscles and skin are now competing for blood flow and your heart rate has to increase to match the requirement. As I am sure you know after a few efforts indoors, even with a fan, you still get very warm, very quickly. So when comparing your EF to previous efforts, think if the conditions were the same. Was the temperature in the room the same? Was the fan on for all efforts etc. Make sure you try and keep as many factors constant for repeated testing efforts. Failure to do so will affect the reliability of the test.
  • Caffeine. Similar to heat, caffeine will affect your heart rate, so ensure you keep your pre-test caffeine intake constant.
  • Fatigue level. During periods of fatigue due to high training loads, heart rate may be suppressed or sluggish. This could produce a desirable EF or decoupling value, but may not be an accurate representation.

Decoupling is the ratio of power output to heart rate. The concept behind decoupling is that as a rider fatigues, the power output will decrease, but the heart rate remains 'elevated'. Training Peaks state that a Decoupling of 5 percent or less indicates a good level of endurance capacity. 

However, as with EF, this is a heart rate related metric, so the bullet points above apply here too. Also, the effort was relatively short, so a longer effort might be a more valid measure of the decoupling metric. An important consideration, is that heart rate metrics may be suppressed during high training loads. That may produce a low (good) decoupling value, but this may not be accurate. 

 

If the 20 minute time-trial is a test that you will be performing at regular intervals, it is important that you standardise your testing protocol as best as possible. Standardising your testing protocol will go a long way to increasing the reliability to the test.

Pay attention to the following:

  • Caffeine intake - Try and keep this constant. If you have a double espresso before each test, make sure you do the same each time, and about 30 minutes before you expect to start your effort (not warm up).
  • Temperature - If the room is airconditioned make sure the temperature is the same, and if you use a fan, make sure you always use a fan.
  • Standardise your warm up - Make sure your warm up the same before each test. Don't do 10 minutes one time, and 45 minutes the next time.

 

In summary, you paced the effort well and it appears that your endurance capacity is good, but if you really want to see value in these metrics, I would recommend that you track them similar to how we showed in Figure 4 above. 

 

Hope this answers your question CadenceBlur. 

Thanks for getting in touch.

cadenceblur, May 29 2018 02:34

Thanks very much for the input. Will definitely start using those tracking elements. I have quite a bit of data already (power profile etc.) it's just about setting it up correctly to make the analysis easier.