The data behind the pace

Open the stride and quicken the cadence

Today's run (street): 5.5 miles

Conditions were chilly this morning and I had to dress like I was going out for a winter run. I feel like it's payback for all those weeks of unseasonably warm weather that we had in March. Somehow I picked the right combination of gear and set out for a run in a nearby neighborhood. I started out with a loop around a local business park that has one hill that I always dread. I did another half loop before moving on to run in what I call neighborhood #3.

Once around the park

My run was fine and, although I thought I was doing a good job introducing speed at times, I ended up running fairly slowly. When I looked at the data from today's run, I noticed that my current pace hasn't degraded that much since mid-2014. It was interesting to see that my training paces were often 2 minutes per mile slower than my race paces. That gave me hope that I could get back into race shape if I wanted to start focusing on performance.

I decided to do an analysis that compared three data points from my run history: pace, stride length and cadence. I randomly selected thirteen runs between 2013 and today that had data captured via my Garmin foot pod. Pace is measured in xx:xx time format, stride length is typically between .8 and 1.1 meters and cadence usually falls (for me) between 160-180 SPM. Those disparities required me to index the metrics so they could all be displayed on the same scale.

My first reaction when the data was visualized was that faster paces are clearly correlated to longer stride length and faster cadence. Not a surprise. I know that 13 data points doesn't yield statistically significant findings, but it's enough information to be directional. It shows that if I want to get back to 9:30 training paces, I'll need to average between 172-176 SPM and stride lengths between .98 to 1.0 meters.

I have work to do to get to those numbers but at least it's a baseline target. I need to decide whether to focus on cadence and let my stride adapt as needed, or if I should try to open my stride before taking on the tougher metric. I'd prefer the latter, but messing with stride length is tricky because over-striding is the gateway to injury.

The history of my running speed

Directional declines

Today's run (street): 3.6 miles

I decided to do some data mining on Garmin Connect to compare my historical averages with my current performance. In order to keep the information consistent, I only used data captured from one source, my Garmin 210 that I bought in 2010. I know I've lost a lot of speed over the past year and my interest was in seeing whether my recent history is an aberration, or if it merely reflects a long term decline.

Charting the trends reveals a changing relationship between race speed and overall speed. My average pace has followed a linear decline, but my race paces have dropped measurably since 2012. Up to 2012, I generally paced 7.5% better in races compared to my overall average. After 2012, that gap has closed and is now almost equal to my training run times.

As I often say when working with business data, these findings are only directional. The Garmin data, acquired by GPS, has a variable margin of error. I tried to correct for that as much as I could, but the numbers do have some skew. I only selected runs I'd tagged as "street running" to filter out slower trail paces and faster track paces. It's also important to note that the 2014 data is only through May 25, not a full year.

In terms of these findings, I'm not happy to see declines, but at least the drop-off has not been as sharp as I'd suspected. I did today's run as a tempo, taking it easy through the majority of the distance and picking up the pace more at the end. The last mile was a minute faster than the prior few, and I finished feeling great. I wish I could tap into that speed more often, but based on my recent race performances, it's a little more complicated than just trying a little harder.

I swear I didn't run through my neighbor's houses

I love data visualization

Today's run (street): 4.2 miles

The sky is as white as paper and the temperature is dropping. We're supposed to get two inches of ice and snow by late afternoon. Nothing so far, but I can tell it's coming. For that reason, I made sure I got outside early to get in a few recovery miles before conditions got worse.

I had no intention of running as hard as yesterday and had to throttle my speed a couple of times. A recovery run is supposed to be done well under anaerobic threshold to help flush lactic acid from leg muscles. I used my heart rate monitor to guide my pace, averaging 75% of HR max throughout the run. It was a nice relaxing workout, although my (gloved) hands got surprisingly cold.

No actual yards or living rooms were entered

I've written a lot about the variability of GPS as a measurement tool and today's margin of error was particularly egregious. Not only did the Garmin show me starting the run three blocks from my actual beginning point, the accuracy was laughably bad throughout the entire route. The Garmin route map (see above) makes it look like I ran through many people's yards and houses.

Distribution of pace times through the year

Reflecting back on 2013's racing season, I charted my race paces to see if there were any obvious patterns. The data doesn't show any trends that would explain my performance, as times were all over the map. I'm hoping that next year will yield faster times and more consistency. At least I finished the season in a good place.

No regrets

Yesterday afternoon's workout (cycling): 4 miles

When I look back on a race I often think about the things I wish I'd done differently. Sometimes I lament the way I'd paced my run and other times I'll regret that I didn't do enough focused training. I can usually identify at least one problem that I wished I'd corrected prior to race day. But, for yesterday's 8K, I can happily say that I wouldn't have changed a thing.

The important point I took from this race is the knowledge that I'm capable of running a mid-8:00 pace for five miles. Even at the finish I felt strong. I had been concerned that I was losing some of my edge because my average training pace has degraded some over the last couple of years. I realized yesterday that, for many months, I'd backed off too much on performance in favor of having more comfortable runs. Only recently have I persistently focused on performance, including speed work on the track and extra workouts from cycling.

If yesterday's performance is an indicator of  how I trained then there's not much I'm going to change. Even so, although I was happy with the way things went at the race, I can't ignore that I finished solidly mid-pack and toward the back of my age division. There were a number of speedy middle-aged club runners on the course yesterday so I understand why I placed where I did. Maybe next time I can be a little more competitive.