How to read cyclocross race power data

I would like to optimize my use of the TR analysis tools in examining my cyclocross race power data. What are some things I should be looking for? Maybe some highlights on how you all examine your power data?

I am just trying to figure out a structured approach in race data analysis in a way that it will be useful in comparing with future data as I move forward. Thanks!


I was never able to glean anything useful from my power data since every course is different and you burn matches in different scenarios depending on the race. Throw in significant running and I’m at a total loss. I pulled my PM and put on a lighter crankset and just use HR. If I avg over threshold for 45 mins with some bouts near my max I know i didnt hold back. Would be interesting to see how others go about it

1 Like

I’m not really sure there would be much to be gleaned from CX race power data TBH. It’ll be a spiky mess, and you’ll either have the power on the day or not!!

1 Like

Because it’s going to so spiky, the things to look for would be:

  1. AWC/FRC – is it getting bigger?

  2. Repeatability – are you stretching out your ability to make the big efforts?

  3. Fatigue resistance – not of the time trial variety, but closely related to #2: how powerful are your late-race high power efforts?

  4. high VI – good crossers use handling skills and coasting in a manner similar to crit racers. It varies by course, or course, but hypothetically speaking, a NP of 330 for 45 min with an AP of 225 could be indicative that you were better at conserving energy through good bike handling and dropping the hammer harder on the hard parts.

1 Like

I put my powermeter on my cx bike for a race recently just to see. What I learned is that my longest effort of about 30s was the start, then its a spiky mess of either anaerobic spikes or no power, each spike less than 5s. It was a dry, fast course with lots of corners and steep banks.

I do think to learn anything from it, you’ll need a good memory of the race, combined with the gps data. Remember where on the course you got hold up, or put in extra effort to pass someone.
I’d split the data into laps and separate the start too. Then you can compare power on similar part of the course, even though changing course conditions and traffic might account for differences too - for example, if you did progressively less power on a climb, this might mean you faded, or that it got muddier and you had to reduce power to keep traction.

I’m not sure if you can learn that much more from the power data than from your lap times. I’ll probably collect some data again on a different course, but not sure if it’s there is that much to see.

I’d interested in other replies to this though!


I think it’s really interesting to look at power data from a cross race. I’m not a coach or an expert in this by any means, but it’s cool. Here’s my thoughts, please take them with an entire shaker of salt:

Just to get it out of the way, I don’t display power on my head unit during a race. In fact, I put my head unit in my pocket. It isn’t telling me anything I want to know, and it I have to change bikes I don’t want to worry about it.

Variability is a huge one for me. Obviously this varies a lot by course, but I want that variability to be high. I used to try to time-trial my way through races, being smooth, pedaling through everything, minimizing spikes. It worked out acceptably, and I had some very good results where my normalized and average powers for the race were within a watt or two. Down the line I learned that riding this way is just leaving speed on the table. Coasting more, carrying speed, and then hitting it hard when it counts is both easier and faster. So now I look to see how well I’m doing on that since I still want to revert to TT mode. In a good race I usually end up with a VI of somewhere around 1.15.

I also look at the number of hard efforts I do around different time scales. How many 30s efforts did I do? How about at 1m? What’s the average power on those? How does it decay? I figure I can assess where I’m weak and how I can get stronger. It can also help with pacing. Maybe I hammered a straight and couldn’t kick as hard up a hill, did I go faster than when I took it easy on the straight and punched it up the hill? Which hurt more?

It isn’t power, but it’s useful to look at cadence too. Trying to figure out where on a course I’m bogging down or spinning myself out can help me be more efficient.

Other things that aren’t power related but are somewhat relevant:
Make sure you aren’t using a wheel speed sensor or you won’t pick up running sections.
Make sure auto-pause is off.
If you do look at your head unit during a race, I suggest having elapsed time and speed as your data fields. The clock helps with pacing. If your speed is lower than running speed it’s a reminder to swallow your pride, get off and run (I struggled with that a lot, especially in sand).
I’d be more than happy to share some race data files if it helps anyone by illustrating any of this too.

I’d also be super curious to hear what other people look for in a file from a cyclocross race. What takeaways can you come up with?

1 Like

Wow, really great replies so far! Lots of great insights here and enough to keep me looking at my power file for an hour or so. The racing strategies are also helpful. Thanks, much appreciated!

I don’t look at anything but the short power numbers, 10s, 30s, 1m, maybe, sometimes a bit longer if the course has a decent climb in it. I also look at power ave/normalised over each lap, conditions and fatigue are a factor, but depending on how I placed I want to know if my power was dropping off but I was holding my (hopefully podium) place or was I dropping power and spots in the pecking order were slipping away. My point is, lots of useful data, having the autolap turned on will help a lot too.


I am a coach, looking after top Youth riders who don’t have power meters. I look at the HR and can tell an awful lot about the ride an individual has had from that. Look for a peak into Z5 at the start and the bulk of the race is like a hacksaw blade, high Z4 and regularly peaking into the red. I expect to see another peak at the end, certainly if sprinting for a win or podium. A drop in HR can show a fall or hold up but otherwise it’s the sawtooth that I’m looking for to show a consistently hard effort.


We use a lot of the same courses year to year, so comparing my performance year over year is interesting.
I use power to look at AP and NP (and together as VI), average power per lap to see how I am doing relative to last year (is my average up, is my VI down indicating I am smoothing out my power delivery and cornering better, etc.). I also like to be able to track TSS. I haven’t found the tracking of power PRs to be that useful.

1 Like

Here’s an example that build off what @Raythebike said. It’s pretty hard to tell what’s going on with this race by the power data, but the heart rate shows a pretty clear trend - bury it for the first two laps to establish position (HR high and staying there), then settle into a rhythm (that sawtooth HT plot).


Here is what one of my races with power looks like. Both in Strava and Xert. The gaps with zero power are the running sections where I burn a match. I’m working way under my threshold power on avg but physiologically i’m over my LTHR for pretty much the whole race and hit near my max a few times.