Trail Zealot

Colorado hiking, Appalachian Trail thru-hiking, and more...

March 27, 2020, 7:22 a.m.

Well, looks like I pulled off a critical velocity test. It was much easier to get it done this week vs last. I had learned a few lessons about how freakin bad I was gonna feel, and I did my best to set my judgments about that aside. When I'm suffering in a workout, I have a tendency to revert to this "god I suck so much" mindset, no matter what my objective performance looks like. It's almost like the act of running really hard makes me feel slow. I suspect I avoid the higher intensities for exactly this reason. I'm riding high on endorphins right now, but I really think speedwork can teach me some stuff, and not just about running.
For this workout, I really knew what I needed to do (unlike last week). I needed to go out as hard as possible, so I could run out of anaerobic energy supplies faster, so I would spend more time running at critical velocity, and then I needed to run at critical velocity as long as possible (suffering my hardest) to get good data. Woof.
I left the house a little after sunrise, and the air was chilly and foggy, but I still found a couple people on the track. Nothing like last week though; I think the progressively worsening coronavirus situation has people staying in more. Ideally I would have worked out on an empty track, both from a coronavirus perspective and a nerves perspective. But I didn't get knocked off. I jogged a mile in a long sleeve, gloves, and pants, visualizing the effort ahead of me (as well as the pain). I felt prepared. Sarah and I touched on this workout during a nice meandering conversation last night, and it clicked for me. I needed to just go out there and go for it, running boldly and maybe blowing up in an unexpected way. I needed to surrender control. And I was ready!! But that was last night. I felt like running right then. I was able to get back into that headspace for the most part.
As far as I could tell, I didn't start out as fast as I did in my previous efforts, focusing on not pulling a muscle or something. I don't run fast often, and the last thing I wanted was to get injured at a time I'm enjoying running so much. I definitely felt the effort, but I stuck with it. Between 600 and 700 meters into the test (on the straightaway headed north), I distinctly felt my body transition. I think my legs told me first - they sort of stiffened up and it felt like they were made of bricks. I was still working hard, but not getting there as fast. By the time I got through the curve and hit the final straightaway (700m - 800m), I knew I just had to hang on. I was running really slow. It was hard to tell if this was my true CV speed, or if it was a little slower because I relaxed after bonking. I tried to be a sport and run through the full 3 minutes. What I remember is not feeling a sense of panic at the end. I felt like I had done it. Was I holding back to preserve my sense of control, or was I just comfortably dealing this high intensity state? Tough to know.
What I DO know is that I could barely jog after finishing. I stopped completely and doubled over, and tried to trot. Not sure what it felt like, but I was like uh no. So I walked over to my extra clothing and took my sweet time putting it on, sweat dripping off my head and pooling near my feet while I tied my shoes. My diaphragm/stomach area felt like it was being punched as I slowly jogged the mild downhill home. At first I felt like I might throw up. That feeling subsided, but I sure wasn't looking to run fast for the rest of the workout. What followed was an enjoyable unpaced jog back home, followed by another enjoyable run with Murphy.

Analysis time.
GPS: Accel started around 19:45. Up to speed by 19:54. Seemed to pull up around 22:48. Total time could be from 2:54 to 3:03, depending on how you calculate. I don't think exact time is important.
Device-reported cadence and speed responses lagged a bit. Cadence made a jump around 19:57 (12 seconds after starting to accelerate), and speed matched that timeline. Cadence and speed drop off after 22:45. I think that was real. I kind of pulled up and looked at my watch and hesitated a couple seconds so I would hit it right at 3 minutes. I'll throw that data out when I compute my average critical velocity.
Experienced a brief slowdown 21:22 - 21:26. Not exactly sure what happened. Might have been that I hesitated because I wasn't sure what was on the track ahead of me - turns out it was a bird eating another runner's puke.
I analyzed the 30sec of data from 22:15 - 22:45. I tried 3 different methods to calculate total distance over that interval: summing the device-reported speed values (and multiplying by 1 second each), summing the point-to-point GPS distances, and taking the difference from the end and the start distance as reported by the device in its cumulative distance series. There was no significant difference between the distance-based methods (148.3 m), but the speed-based method estimated a slightly higher value (152.6 m).
To understand the difference between the speed and distance-based methods, I took the numerical derivative of the GPS-based distance series and compared it to the device speed series. I noticed that both time series exhibit changes in speed throughout the interval, but the device speed (informed by the watch's accelerometer) showed a more steady decline over the interval. That trend is likely true, leaving me to wonder why the GPS speed didn't pick up on the trend as much. The difference in the shapes of the time series is larger than the difference in the last-30-second distances. Why?
To get to the bottom of this, I looked at the GPS trace overlayed on satellite imagery of the track. A few things became clear. First, the trace had me swinging outside of the track consistently. If it's a 6-lane track, I spent most of my time in lanes 8-10. This effect seems worst on the east and west sides of the track, when I'm heading N-S. It's harder to eyeball how far off the curved sections are, but the general trend is that GPS always had me running a larger radius than I really was - falsely increasing my reported speed on the curves. I also noticed that my GPS trace leaving the track is more accurate than when I arrived - perhaps the GPS signal grew stronger and more accurate with additional time. That's a good thing.
So now I'm sitting here trying to eyeball where I was on the track at the beginning and end of the last 30 seconds, as a sketchy way of estimating the distance I covered. I have low faith in my ability to pinpoint my location on the track, but I don't have a better option. If only I had recorded video of the entire interval. I'm only half joking. If only I had managed to push through the final straightaway! I would know for sure how much distance I had covered.
As a way of calibrating my distance to the distorted world of the GPS, I tried to eyeball the time associated with crossing the starting line on each lap. Then I found the GPS distance associated with each lap. Lap 1 came to 421 m, and lap 2 came to 431 m. The GPS traces of those two laps look really similar, so I probably just didn't nail down the correct start and stop times for each lap. Tweaking the lap 1 / lap 2 split by 1 second brings the distance values to 427 and 425 m, respectively.
Next, I tried the same approach but eyeballing the 50 yard line. I came up with 417/436 or 422/431 depending on how you split it up. Still can't figure out why that second lap appears so much longer. They should both be 400m because I got into lane 1 around the 50 yard line on lap 1. There's that weird hiccup starting around 20:25, early in lap 2. The device-reported speed freaks out and then resumes its normal course. It seems to follow this point where I checked my watch, making me suspect it is accelerometer related, not GPS.
Finally, I converted the GPS traces to X-Y space and plotted them overtop a regulation track. While it is clear that the GPS trace is always outside of the oval, it's still unclear where distortions are taking place, or how. The last 30 seconds of data includes one curve and most of one straightaway, so I'll just multiply the last 30 seconds' distance by the ratio of the "GPS track size" to the real track size. The lap 1/2 distances are 434.33 and 431.34, respectively (now that I can visually line up the lap start and end points, I got closer to an even split between laps). Also, I can correct for misalignments of the exact start and end points by zooming in. Lap 2's end y-coord is 152.852. Lap 1's beginning y-coord is 152.33. The lap1/lap2 split's y-coord is 150.765. Meaning the exact lap 1 gps distance = 434.33 - (152.33 - 150.765) = 432.75 m. Exact lap 2 gps distance = 431.34 + (152.852 - 150.765) = 433.427 m. We will call it 433 m! That means, if GPS distances scale proportionally to ground-truth distances equally at all points around the track, the last 30 second distance = (148.3 m) * ( 400 m / 433 m) = 137.0 m. Quite a difference. As a sanity check, if ALL the distorted distance were added along the segment of track I was running in the last 30 seconds (highly unlikely), the last 30 second distance = (148.3 m) * ((148.3m - 33m) / 148.3m) = 115.3 meters! This should be treated like an absolutely ridiculous lower bound.
Now to calculate my actual critical velocity!!! My (corrected, likely) CV = 137.0m / 30s = 4.567 m/s. My fastest possible CV = 148.3m / 30s = 4.943 m/s. The slowest possible CV = 115.3m / 30s = 3.843 m/s. The true answer could lie anywhere along the spectrum between the slowest and fastest speeds, but it likely is near the corrected CV. That's a good thing, too, because I don't believe my critical velocity is as fast as 5:26 per mile, which the unedited data suggests. And I don't think my CV pace is actually 6:59 either, as my conservative calculation might suggest. My actual CV pace as measured today comes out to…5:52!
This makes intuitive sense. I've been using 6:30 as functional threshold pace, and that's based on vague understanding of my fitness from race results and the occasional workout. Now, how does TrainingPeaks' FTP relate to CP? Let's see. Well, CP is a physiologically-defined number. The power or pace that can be maintained indefinitely, until the user needs to stop for a different reason. The power or pace that defines the lower end of the severe domain, where steady-state lactate or VO2 values are not attainable - they just max out. FTP is more of a fuzzy concept. It is the power or pace that could be maintained for 60 minutes before fatigue. Jim Vance has a test protocol to establish FTP for running. And it involves a 3-minute maximal test as well as a 9-minute maximal test, where you average your power over both intervals. It screams of cycling ideology. It's apples-to-oranges to compare (90% of the power of) a paced 3-minute interval to an all-out 3-minute interval with a measurement taken over the last 30 seconds. For cycling, it's usually established using rules of thumb, ie 95% of 20-minute power or best power from 40km time trial on a bike. I'm also reading that there's a small drop-off in power from 20 mins (kind of the CP territory) and an hour (FTP territory). In there, Coggan also has a comment that CP and FTP will be similar if the CP test is a long enough duration. For shorter tests (like the one I just did?), estimated CP will be higher than FTP. It will take some work to quantify the relationship between the two paces/powers, but I suspect there is an easy mathematical relationship. Where does FTP live (what domain), if the time to fatigue is 60 minutes? Is that in the upper-heavy domain? This is a question for another day. I've been at this way longer than I expected, even if it was deeply satisfying.