Training and Racing with Power: My Experience using Stryd
A little over one year ago, I decided to back Stryd on kickstarter. The idea of measuring your output power while running really appealed to the physicist and data nerd in me. I had already been collecting GPS-based pace data for about four years, heart rate data for a couple of years (but not necessarily training using HR), and taking post-run notes on training conditions, feeling, etc. Power seemed like a very interpretable and tangible metric that I could add to my corpus of run data and one that I could really make sense of.
I received my Stryd Pioneer in later November 2015 and immediately began using it. I’ve used it on every run since that time, both in training and racing. I’m using this post to present my experiences in learning to use the power metric in my training, and discuss how it has helped me to reach a new level of knowing myself as a runner.
Data Collection and Logging
For all of my runs, I use a Suunto Ambit3 Peak to collect GPS, HR, and Power data. The Stryd Pioneer is a HR monitor as well, and connects to the Ambit 3 via Bluetooth. Early on in my use of Stryd, I wore two chest straps (Stryd + a Polar H7) since the Stryd-Ambit pairing wasn’t yet capable of streaming/receiving both HR and power. That issue was fixed in December so I’ve only had to wear one strap since then. You must first configure the Stryd unit with your weight before using it. Luckily my weight is pretty consistent at 145 lbs, so after changing it for 1 lb differences here and there I ended up just leaving it alone and never changing unless I wore a pack/vest (more on that later).
All of my run data is imported from the Ambit 3 Peak into my Suunto Movescount profile, then fetched to Training Peaks and also to the Stryd Powercenter. That data flow works pretty well, and each of the three web-based platforms has its affordances and constraints. I’ll show examples from each in my discussion below.
Preparing to Train with Power
In the first month and a half with Stryd, I was “just running” without a training plan, but continuing to maintain and build my base for the season ahead. I ran one 10k race at the beginning of that time period, but otherwise I was just collecting data and getting familiar with the power metric. One thing I noticed right away was how much my “real-time” power fluctuated while running. Stryd uses a 10s averaging for the power it reports to the watch. Even then, slight variations is gradient or effort are very noticeable in terms of the power output reported, which makes sense. But it takes some getting used to and figuring out how to maintain a target average over an interval. The key is to use the Lap function on the watch and to view lap average power.
In mid-January I started on a power-based training plan developed by Mike Ricci at D3 Multisport meant to prepare me for the Leadville Trail Marathon in June (you can view my Leadville race report here). Part of that plan incorporated a B race, which ended up being the Boulder Mountain Marathon in May (you can view that race report here). Mike provided the plan for me in Training Peaks, hence the use of that platform as my central location for workout logging and analysis.
Before beginning training, I needed to define my Critical Power (CP) and critical pace. If you’ve read Jim Vance’s book Run with Power you’ll recognize these values as rFTPw and rFTPa. These are the power and pace outputs you can theoretically hold for a one hour maximum effort. Near the beginning of the plan I conducted the Stryd Critical Power (CP) test, which consists of running 3' and 9' maximum effort intervals separated by a 30' recovery period. I had bootstrapped my CP from my earlier 10k effort as being 221W, and after the 3–9 test I calculated it to be 225W. This is a pretty important reference number, as it is the basis for defining zones and effort/output ranges for workouts in Mike’s plan. I conducted the test again in April and again found my CP to be about 225W. My corresponding critical pace is about 6:30 min/mile.
Mike’s plan had a lot of variation in intensity, which was great. In the past I had just used mileage based plans and was totally guilty of running almost everything at the same intensity. Intensity Factor (IF) is a pretty easy measure to get your head around. It is simply you normalized power (NP; for an entire run or interval) divided by your CP (rFTPw). So and interval of IF = 1 means that you ran at your CP for that interval. Within a single workout, Mike had me doing a mix of low and high IF stuff, especially early on. Another useful measure to characterize that mix is the Variability Index (VI) which is the ratio of normalized power to average power. Workouts with high VI had a lot of variation, while those around VI=1 had little variation in output/intensity. Later in my training, I became particularly interested in my efficiency at different output levels. To examine that, I used the Efficiency Index (EI) which is the ratio of average speed (m/min) to average power (W). More on this later.
So to summarize, the key power-based metrics that I used in training and post workout analysis were: CP/rFTPw, NP, IF, VI, and EI. The one metric that I was most concerned about in day to day workouts was percent of CP. I highly recommend reading Jim Vance’s book Run with Power for descriptions of these metrics and more.
Training with Power
Here is an example 60' workout from early in Mike’s program: 50' run with 8x30" strides. Warm up at 83–88% of Critical Power (CP). Strides should be done at 95–105% of Threshold Power. During the strides, I would occasionally glance at power on the watch to make sure I was in the specified zone, but with efforts that short (30") it would have to be near the end of the interval. I eventually learned to “feel” that output level for short intervals like this. Figure 1 is the Power , HR , and Pace graph from that workout:
You can clearly see that Power and Pace covary tightly, while HR lags appreciably. Also, HR tends to drift up over the intervals. These were very short intervals in the middle of a 60' run. A cleaner example comes from this hill workout (Figure 2): 15' wu. Then run 15x2' up a hill. (or 6x5') up a hill. Power on the hill should be right at CP up to 10% over CP. HR on the uphills can hit low Z4. Keep HR in Zone 1–2 otherwise. 5' cd.
In my experience, it is much easier to precisely manage output (power) when intervals are uphill, and when they are longer. That said, long runs (with very long output intervals) are a different kind of challenge which calls for the “lap averaging” power monitoring that I mentioned earlier. Here is an example workout for a 2:30 long run: 20' wu at 75% of CP. Then you’ll run 2x20' at 75–85% of CP with 5' easy in between. Then 2x10 at 85–90% of CP — 5' easy in between again. Then cool down. Figure 3 shows the graphical data:
The data are not as clean looking graphically, but the actual outputs hit the targets pretty well (see Table 1).
To perform a workout like this, I entered the time intervals and target power ranges in my Ambit 3 Peak as a new “workout” (using the mobile app). The Ambit would then give me an audio cue at the beginning and end of each interval, along with displaying a message which I had customized with the target power value. That system worked well for all of my workouts.
I should note that I *did* adjust my weight in Stryd each time I went on a long run and wore my Ultimate Direction vest loaded with water and nutrition. That added about 6 pounds, which I deemed appreciable compared to my daily 1–2 pound fluctuations in weight. My NP was generally lower when I wore the loaded vest and felt like I had similar efforts (perhaps because those were longer runs), so I felt like entering the updated weight in Stryd was and important factor.
Post-Run Analyses
After each run, I would first look over the data to see how close my actual outputs hit the prescribed targets. Of course, I would already have some idea of this based on how I felt and what I observed on the watch during the workout. I would also consider my normalized power (NP) over the entire run along with my IF (NP/rFTPw) to get an idea of the overall intensity, independent of duration. When considering duration as well, I would examine the Training Stress Score (TSS) calculated by Training Peaks. Over time, I became very good at estimating what IF and TSS would be for a run before I even looked at the data. And that’s something I want to really emphasize in this post: training so analytically using the power metric made me very aware of my efforts and output and how they linked to my feeling or Rating of Perceived Exertion (RPE; the “R” is sometimes referred to as “Relative”). It may seem a bit counter-intuitive, but rather than blindly relying on the data to tell me how I was performing, I found myself more in tune with my performance qualitatively, and the data helped to verify my perceptions.
The other post-run metric I became interested in analyzing was some sort of efficiency measure. I chose to use the one mentioned above: EI. While not a true measure of efficiency (which would be dimensionless- just a percentage), it is the ration of speed to power. Expressing EI in units of m/min/Watt gives values near 1, which seem pretty interpretable. What is interesting is to use that metric to see where (in your power range) one is most “efficient.”
Figure 4 is a plot of EI vs. IF for a limited set of data (some within-run intervals and some full run averages). This plot clearly shows my EI dropping off with increasing IF, and peak EI around 75–80% of rFTPw (CP). That is where I want to be most efficient since I’m racing in that output zone (trail marathons), but it also tells me that I need to work on “efficiency” at higher outputs. But overall, those are pretty good EI values from what I can gather. I haven’t found a lot of normative data to compare to, but based on what others have been discussing online these values over 1 seem pretty solid.
The Takeaway
So what can the data do for you once you know your running self really well? Do you really need it if you are in tune with your effort, output, and form? For me, it comes down to two things: 1) Precision in managing training intensity (which presumes a good plan), and 2) Pinpointing your weaknesses. Especially on long runs, I tend to get a bit cognitively challenged and need help to track numbers or nudge me back into intensity ranges. Having that support really helps to keep me faithful to the prescribed workout. And post-run analyses (when coupled with field notes about weather, feeling, sleep, stress, etc- all metrics that I track daily) help me to identify where I need to work further. For me, that’s keeping cadence up, becoming relatively more “efficient” at higher percentages of my rFTPw, and getting better at keeping pace up and muscle fatigue down on the long downwhills. Stryd has become a key part of my training, perhaps the central part. Running with power has helped me to know myself better as a runner, and to set and train towards appropriate goals.
