Blog Homepage

Assessing the quality of a training session

Previously I have posted an article which listed my four key training principles.  The first one of these was ‘Correct Intensity’.   Whilst it appears quite simple on the surface, this training principle can in fact be the hardest to get right.   Just knowing what the intensity of your ride is supposed to be isn’t enough to ensure that it meets it’s goal and even then doing the ride at what you think is the right pace doesn’t necessarily mean that it will be.

To illustrate this I have prepared a little case study of three ride files, all with the same purpose – Aerobic development through time spent in the E1 and E2 zones.

There are two charts provided (from Golden Cheetah) for each ride.  The first chart shows the power profile for the ride, smoothed to 30 seconds.  The second is a histogram or distribution chart, which shows the percentage of time spent at various power outputs.  The ‘bucket’ size has been set to 20 as it provides good approximations with the training zones.

The 30 second smoothing is an important modification to the normal view (which shows actual data for each second) for this analysis.  The reason for this ‘smoothing’ is to more accurately represent the physiological demand of the session and can help provide a better approximation of what the average intensity was throughout the session.   Golden Cheetah has a nice function which enables this chart to display the training zones behind the power line.  Each zone is a different colour and labelled from Z1-Z7.  Z1 is Recovery, Z2 = Endurance, Z3 is tempo and so on.  For the purposes of this analysis we are primarily interested in the time spent in Z2 and Z3 (and to a lesser extend Z1 and Z4).


Rider 1

The intended duration of this ride was to be 1hr 45min and it was completed as a solo ride.   The terrain used was primarily flat to undulating with only a few short climbs.   The first chart (below) shows the power profile for the ride, smoothed to 30 seconds.


As you can see there were a lot of fluctuation in the power over this ride but in the majority is was completed within the Z2 and Z3 bands (blue and purple background).  You can clearly see that there was a large number of instances throughout the ride where the power dropped off (approx 40 times – see the downward spikes into the pink section at the bottom).  These were most likely from short downhills and stopping at traffic lights.  At the other end of the spectrum, the time spent at or over threshold is a lot lower, with only 11 spikes above the blue Z3 section.  Now the higher these spikes go the more anaerobic the effort (remembering that the power has been smoothed for 30 seconds).  As you can see there were only 2 times where the power went into anaerobic (yellow band) territory throughout the ride.  The first one was on a short steep hill and the second was starting from a traffic light. 

Moving on to look at the histogram.


This chart clearly shows that the majority of time was spent in the correct training zones, with a total of 60.1% of time spent at either E1 or E2.  Given the ride time of 1:46, this equates to approx 64 minutes effective riding time.  Although it’s worth pointing out that time spent just above and below these zones is not totally wasted as it still has a training impact.  Also worth noting is that the time spent ‘not riding’ (20watts or less) was only 6.9% (about 7 minutes).

All up quite a good ride, which clearly met its purpose.


Rider 2

The intended duration of this ride was to be 4hrs and it was completed as a group ride.  The terrain covered would be regarding as undulating to moderately hilly.


The first, and most obvious thing to note about this ride is the massive amount of time spent below the target power (in recovery or the pink zone), a staggering 71.3% of the total ride time.  With so much time spent below the target power, this effectively became a long ‘recovery’ ride that only really stressed the rider towards the end of the session. 

The trouble here comes from two distinct areas -  adhering to the ‘bunch pace’ (which was obviously too slow for this rider) and secondly, having to sit in and draft, which not only reduces the power but also introduces a lot more instances of coasting (even momentarily).    Looking at the time spent in or above the desired training zones, there were some discreet periods within the ride that were fine, most likely when the rider was on the front or perhaps climbing a hill.  

It is worth noting that due to the hillier nature of the course, there was more time spent coasting downhill and this can be seen by the longer periods with low or no power.  There is also a distinct flat section late in the ride, this was from a stoppage but where the Garmin kept the data in the single ride file.   It is important to factor this into the overall assessment of the ride statistics, which we can do via the histogram.


The beauty of the histogram is that it can be used in conjunction with the intervals throughout the ride.  In this manner you can check what contribution each segment of the ride made to the overall power distribution.  This is an extremely useful function later in the season where you are completing set threshold or VO2 intervals as you can check the power distribution for individual efforts and thereby assess the quality of each individual effort rather than the ride as a whole, which will have lots of recovery time to skew the overall data.

For this ride, the time spent below E1 (Z2 on the chart) was 71.3% of the total ride time.  Of this 20.2% was spent below 20watts (approx 18% if you exclude the stoppages).  The cumulative time in E1 and E2 was just 18.8%.  Based on the total riding time of 3:20, this equates to approx 38 minutes effective riding time.  When compared to the first ride, this is only just over half of the effective riding time but from a ride that was about twice the duration!


Rider 3

Rider 3 opted for a solo ride and their planned duration was 2hrs.  Terrain was flat to undulating with one longer climb.  Here is the power chart – again smoothed to 30 seconds.


The first thing to note about this chart is the absence of the anaerobic zone markings.  The reason for this is was there was no effective time spent at that intensity.  The chosen course was relatively free of traffic interruptions, which also meant little time lost to short stops.  As with the chart for rider 1, this chart shows that, even though the power still fluctuated (as it does) that lots of time was spent at the right power.  Overall there were still upward of 30 instances where power dropped into recovery, however unlike rider 2’s file, these were brief drops and power promptly went back to the correct intensity.

Moving on to the histogram we can assess the amount of time spent in each zone.



The first thing to note is that the time spent below 20 watts was extremely low, in fact just 3.2% of the total ride time – or just under 4 minutes.   Cumulative time spent in E1 and E2 was 54.8% for an effective time of approx 64 minutes.  Overall this ride is of good quality and is on par with rider 1 for effective time spent in the correct training zones – albeit from a slightly longer overall ride time (2hrs vs 1:45).



First up, it is worth noting that rider 2 acknowledged that this wasn’t the best quality of ride and there was actually some data missing from the start of the file (20-30mins of quality work).  Nevertheless, this type of analysis helps provide more clarity on the effectiveness of a ride and can be used to help fine tune routes and group options for future rides.

It is important to note that while both Rider 1 and 3 chose solo rides, this doesn't guarantee a more effective ride, nor does choosing a group lead to an ineffective ride.   The key is to understand the purpose of the ride and then execute it during the training session, much like understanding the tactics required to win a race and then executing them on race day!  Getting effective training in through group rides is definitely an art form and in the majority the secret is to stick with small groups of riders with similar ability and goals to your own.  By all means go and try various groups out but check the data to see whether it met your goals.  Additionally, it is important to bear in mind the positive social aspect of group rides and losing a little bit of training quality for these positive impacts is fine, just limit the number of bunch rides each week and make sure your other sessions are hitting the mark.

It’s worth noting that while this analysis was done using power data it can quite easily be completed using heart rate.  However, you shouldn’t smooth the heart rate data.  Leave the smoothing to zero or perhaps 5 seconds to help remove some of the ‘noise’. 

Software – while this analysis has been completed using Golden Cheetah it can also be done using other software and riding systems.  In WKO+ you can add horizontal lines to mark out the training zones and then smooth the data.  The histogram functions in WKO are also similar to Golden Cheetah.    Training Peaks (online) also has lots of features akin to WKO (built by the same people).

Finally, while the purpose of this article is to help demonstrate techniques to assess the effectiveness of a training session it is still important to remember the bigger picture.  Not everyone is or needs to be super focussed or dedicated on assessing every single training session and you can easily go overboard analysing ride data.  It is important to remember that we all ride because we love it and training is just part of the process to getting fitter and improving results.  Don’t let this type of analysis get in the way of your enjoyment but Do bear it in mind when you ask yourself whether improving the quality of your training will lead to better results – after all doing 1hr of effective training in under 1.5 – 2 hrs rather than 38min from a 4hr ride leaves a lot of extra time to enjoy the other things in life!


Jason Mahoney Wednesday 25 September 2013 at 10:42 am | | Tips

No comments

(optional field)
(optional field)

Comment moderation is enabled on this site. This means that your comment will not be visible until it has been approved by an editor.

Remember personal info?
Small print: All html tags except <b> and <i> will be removed from your comment. You can make links by just typing the url or mail-address.