Cyclocross Power Analysis

Now that I’ve got a few races under my belt with power data, I wanted to take some time to go over what I’ve learned and how useful I think it is.  Some of the analysis is for power data only, but I’ll also go over what other telemetry (GPS, heart rate) can mean for racers.


The most important data derived from GPS is lap times.  Lap times should be consistent overall for optimal pacing.  Striking a balance between a good start and good pacing is key.  On a technical course with difficult passing, you could easily lose 30 seconds in the first half lap if you’re stuck in traffic.  However, those 30 seconds may only be the difference between a couple places at the end of the race.  A few seconds may be lost waiting to pass each rider as the race progresses, but overall it becomes increasingly easy to pass as the race goes on.

As far as pacing, I’d suggest trying to slot into the place you intend to finish at the start of the race.  Racing too far ahead can lead to poor pacing later in the race due to fatigue.  Cyclocross lap times are surprisingly consistent.  In this from last year, the leader’s times varied less than 10 seconds over each of 9 laps!

Heart Rate:

Heart rate tracing from Boulder CX Series #2

Above is a tracing from a recent race which shows pretty good pacing.  One of this issues I had, which is reflected in the tracing, was death by bobbles.  I was able to stay with the lead group for the first two laps, but I had a few bobbles a near crashes that left me to regain the group.  This lead to my heart rate spiking a few times until I eventually lost contact with a of about eight riders.

I still rode smart enough to maintain my pace, as can be seen in my consisten heart rate at threshold for the majority of the race.  My lap times were consistent, slowing from 7:35 on the first lap to 7:46 on the second to last.  On the last lap I wasn’t within reach of anyone else so I eased off quite a bit as can also be seen in my heart rate.  Throughout the race as riders fell off the lead pace, I passed about one rider per lap and I managed to finish a respectable fifth place.


Boulder Cyclocross Series #2 – Power Tracing – Lap 1

Above is a power tracing from the first lap of the same race.  It looks almost identical to an .  The power numbers for the first lap are a bit higher than the subsequent laps by about 10%, but actually average the same if the initial 30 second sprint is taken out.

Easily the most important metric will be Normalized Power (NP).  NP is calculated to account for the increasing physiologic demand of higher power outputs.  In a hard hour long criterium, I expect to have a NP very close to my threshold power.  So far this cyclocross season, I’ve been seeing numbers about 5% lower than threshold (320w for this race).  I believe this to be due to the time off the bike.  Four, six second running sections per lap will take 24 seconds, or 5% of an 8 minute lap, thus decreasing power by 5%.  Of course, the power meter doesn’t measure anything when not pedaling, and the running sections also have high physiologic demand, so a short but critical part of the race isn’t being measured.

Another important metric is Average Power, and how it is compared to Normalized Power, which is sometimes referred to as VI (Variability Index).  This shows how even the effort was.  Some courses will naturally have higher variability than others, so I’d take a look at VI as the race progresses.  In the opening laps of this race, my VI was highest (1.15) and lowest (1.08) in the middle of the race when I was riding smoothly and setting my own pace.  I’d suggest that riding the corners at maximum speed will limit the power required to get back up to speed, lowering VI, NP and reducing physiologic demand without increasing lap times.

Whats missing

The last sentence above alludes to a major point of racing cyclocross: technical skill is a major part of the equation.  Inefficient body mechanics will unnecessarily increase fatigue, slowing excessively in the corners will cause higher power output, but not faster speed.  Higher power may not always be a sign of a good race, also the case in road racing.

I’m not sure how helpful it is, but I made a quick spreadsheet with a few possible metrics for ‘efficiency’, comparing speed, power, NP, and VI.  I’m certain it won’t be comparable across courses (it can’t account for more technical courses or slower course conditions), but I think the data shows which lap I rode most efficiently.

Xilinx Metrics

I’ve colorized the data at the extremes in each set of data. Clearly, something happened on the fourth lap (Fifth row, above). From memory, I’d say that I’d given up on sprinting to regain the lead group and was riding as fast as I could at my own pace. My average HR was lowest (ever so slightly), but my speed was second fastest despite my second lowest power numbers.  During this lap I likely had the intense yet relaxed focus that is the essential midset of a successful cyclocross race.

The MPH/VI metric seems to select this lap as well, showing the highest value. Interestingly MPH/VI is independent of actual power output and could probably be compared among different riders on the same course and may be an indicator of smoothness and/or carrying speed through the corners.

About Russell

I have been racing bicycles for a decade. This blog will chronicle my efforts as a Category 1 road racer lining up with the pros.
This entry was posted in Race Strategy, Testing, Equipment and tagged , , , , , , , , , , , , . Bookmark the permalink.

4 Responses to Cyclocross Power Analysis

  1. says:

    Hi Russell – Longtime reader of the blog, always appreciate your posts. Was wondering about something over my pre-ride coffee this morning, and am curious what your opinion would be. How do you think Colorado racers would stack up in other states? Meaning, would a CAT4 racer in CO be a CAT3 in say, Texas? Sorry this is unrelated to your above post, just wanted to hear your thoughts.


    • says:

      No worries on the off-topic comment.

      I’d say Colorado is strong, but not a lot stronger than other populous areas. You might find a Cat 3 race here to be a little harder than other locales, but probably not the difference of an entire category.

      Of course some states and regions are at a large disadvantage (Nebraska, Montana, Wyoming, etc), but there are quite a few regions known for their cycling (Colorado, Texas, Southern Cali, Northeast, Northwest, Florida, and Georgia come to mind). I’d say the biggest factor is the size of the overall talent pool – a combination of the size of the general population and interest in cycling.

  2. Steve says:

    Great write up. I was turned onto your blog based on some posts regarding Lookout Hill climb and have enjoyed all of your race reports and training tips. The power, heart rate and pace information is particularly well done. I understand power and heart rate in training, but could never make the leap to analyzing race results. Typically, I only look at the heart rate data during races to make sure I don’t blow up.

    • says:

      In mass start road races, power (and even heart rate) aren’t too helpful during the race. Sure your heart rate or power may indicate that you’re past your redline, but generally in road racing you’ll be best served by doing your best to stay with the peloton and/or the lead group.

      In a cross race I may occasionally check my heart rate to see if I’m overdoing it, but as I saw at CrossVegas this year, sometimes your HR will be much higher or lower than expected. During a race the best gauge to use are your sensations and race experience. (you’ve probably heard me say a time or two that more race experience is always a good thing)

      But, after the race, there can be a lot to learn by taking a close look at the accumulated data.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Connecting to %s