Bent Up And Motivated: First Look at My Runalyze Data (2016-2021)

I’d like to think I spent the last 2+ years hibernating from serious training, briefly coming out of the cave for some hard, extended training here and there, but eventually finding my way back to the cave for a while.

First, I had to change careers again in late 2019 after deciding to move back to Vegas.

Then, all that COVID mess started in March 2020, and there wasn’t any practical need to train for most of the year.

Then there was no Vancouver Marathon in 2021, so I just ran a lot on work breaks after starting my new job.

Then I actually got to train for a marathon in summer 2021, but then my lower body decided to implode about midway through, and I never quite got back on track before deciding to abandon ship on that in mid-October.


Now, after a couple years of sustained regular cross training, progress in mostly regular strength training, after having to learn a few more things about running to help stay injury-free and avoid past burnout mistakes… here we are at the doorstep to 2022, and just in time for that I discovered a neat run-data-tracking website called Runalyze.

After porting all my Garmin data over and seeing what they showed me, I was suitably impressed and paid for a Premium membership.

In one place, Runalyze’s dashboard shows you an initially overwhelming amount of data. Along with all of this weeks’ completed workouts with all sorts of intricate detail, the right sidebar shows all sorts of summary data and metrics you can take or leave.

A key metric I didn’t previously use is TRIMP, which is a composite stress score for all of your training. I’m familiar with TRIMP but didn’t have a reliable and accessible way to consistently calculate it… until now. As soon as I uploaded my Garmin data, I had fully calculated TRIMP for everything (running, walking, cross training, EVERYTHING) I’ve done since the day I put on my first Garmin watch.

One metric I wasn’t aware of that I now really like for obvious reasons is Marathon Shape, which based on your data shows a percentage score estimating how ready you are to perform The Longest Run. Like TRIMP, Runalyze immediately calculated Marathon Shape for every single day for which I had Garmin data, whether or not I was training for a marathon at that time.

The other key data from this is Estimated VOMax, from which I can get an estimate of what kind of finish I can reasonably expect in races of any popular distance.

What I did not have, however, was all my run data from my Fitbit era, dating from Christmas 2016 until I switched to Garmin in September 2019. While Runalyze only calculates back a few months and doesn’t need the data to give me current, accurate projections… I did a lot of my best running with the Fitbit and I wanted that data.

So I went to export and then upload my Fitbit data. It took a while for the site to update, but once it did I ran into a problem. My ported Fitbit data did not include the heart rate data, which Runalyze needs to calculate Estimated VO2max, from which it calculates Marathon Shape and the other key metrics. The limits of Fitbit’s json data files would not allow that data to port over.

To get this missing HR data, I’d have to manually enter average HR for all those old uploaded 2016-2019 workouts. I initially went to Fitbit to view it but it turns out Fitbit’s dashboard has been converted into steaming trash and will not readily show old day to day data, so I had to go to my Runkeeper data, which I’ve kept updated since I got back into running, and which includes my average heart rate plus accurate distance/time for every single run I’ve done.

This is a long and tedious data entry project, but I’ve literally built an entire career out of long and tedious data entry projects so while this took some time, it’s nothing I could not bear or handle. To their credit, Runalyze’s interface makes it rather easy to go into individual run entries and edit this info, so bit by bit I entered that heart rate data to get EVO2Max data.

I also noticed the EVO2max is tied to weight, that when I updated this number for an older month that the subsequent EVO2max data changed. While I had weight data uploaded for the Garmin era, this data also did not port over for Fitbit.

I still need to update weight info for all dates in 2016-2019 (I’ve always been good about tracking weight daily or close to it). But in the interim I did use a Fitbit graph of my weight to enter rough numbers by month or quarter depending on major changes, so the EVO2max data shown is roughly accurate.


Looking at 2021 data and looking day over day at all my old data, I gradually became disgusted. I mean, I enjoyed doing this, and for reasons I’ll get into I like where I’m at. But I can now see in objective metrics exactly how far my fitness has fallen since what was probably my peak performance level in 2017, and it’s not age-related either.

I’m just not training like I was for all sorts of reasons both circumstantial and controllable, and I’m certainly not racing as much as I did that year (I ran SIXTEEN races that year, none of which were a marathon, and not including any time trials).

Below is a graph showing my rolling EVO2max (dark line) and my relative marathon shape (orange line) from 2017 to today.

Runalyze’s calculated VO2max and Marathon Shape for my 2017-2019 running data

You can see my EVO2 was substantially, consistently higher during 2017-2018 in Chicago than it has been since I returned to Vegas in 2019.

The big picture is disappointing enough, but it really bothered me when I saw day over day, week over week, what kind of workouts I was knocking out and what stress/benefit it was providing me over time, knowing that today I’m just cross training, doing short break runs, and I haven’t done a run beyond 13 miles in forever… not to mention having gained a few pounds in the couple years since coming back to Vegas.

Now, that said, Runalyze also showed me that quite a few of those runs were objectively poor VO2max runs relative to my heart rate effort and to the eventual average pace. In training for quantity over quality, I frequently didn’t get a good return on my effort.

Granted, as the first picture above shows, that’s still the case today, and probably more so than back in Chicago. And granted there is a caveat that many of my Chicago runs were “haul runs”, where I was carrying my backpack and thus an extra 10-20 pounds in weight. That will slow you down and bring an EVO2max score down for a run. Still, other runs without a backpack had similar scores, and overall I could see that my rolling VO2max was still higher than it is today.

Now I have a full list of workouts where green arrows show a positive VO2max score, a full range of easy runs, quality workouts and long runs. I can see what workouts were more rewarding and what my pace/effort etc here, as well as the maps of those runs and other surrounding circumstances.


The other key data, the Marathon Shape, now gives me a true picture of how well my training prepared me for my marathons. I can now see in hindsight that I was about 70-80% fit to run a marathon for most of my 2018 Chicago and 2019 Vancouver training, reaching 100% shortly before Chicago and almost but never quite getting there for Vancouver (about 94%).

Right now, just beginning to train for Vancouver 2022, I’m obviously nowhere close. My EVO2max is a could-be-better 32, and I’m only in 25% marathon shape. Let’s see how that looks after a couple months of 14-18+ milers.

Hidden in the Sidebar is a helpful calculator that can estimate your performances, as well as show your documented PBs:

From here, I plugged in what my EVO2max roughly was for Chicago 2018 as well as my Marathon Shape at the time, and got a curious result:

According to this, the best marathon I had any hope of running was a tick under 4:47. Before the hiccups derailed that marathon halfway through, I was otherwise comfortably on pace for a 4:10-4:15 effort. This indicates I never really had a chance at 4:10 or 4:15, that I probably would have crashed and burned well before the finish. Not to say the hiccups were a blessing in disguise, but it at least gave my failure a more circumstantial alibi.

Instead of plugging in my Vancouver 2019 estimates, which mostly would have produced the same results (and because I was barely ready to just finish the race, I had no real time goal there)… I’ll skip ahead to a current projection for my next goal marathon, Vancouver 2022.

I entered an EVO2max that’s in line with my most recent 10K race performance (the current EVO2max calculation is obviously a bit low, probably because my recent workouts have been a bit light), and calculate it with the 100% Marathon Shape I’m expecting to go to Vancouver 2022 with.

The expected marathon time is a LOT better at 4:26. If I simply maintain my current fitness while building the necessary fitness and endurance to race the marathon distance, I’d have a clear documented PB in the marathon. Accounting for Vancouver’s course being a bit long at 26.35mi, my likely projected time is around 4:27:30.

This also gives me an accurate marathon pace to train with in workouts… and not a moment too soon given I actually plan to run a pace workout tomorrow.

This of course is not a final product. I have a lot of other research to do, which will help me see the focus I need to have in future workouts.

Tagged , , , , , , , , , , ,

6 thoughts on “Bent Up And Motivated: First Look at My Runalyze Data (2016-2021)

  1. […] been reviewing and referring to Runalyze‘s Marathon Shape metric to not just gauge the shape I was in during prior training cycles, […]

  2. […] had aimed for 10 miles and the 100 minute workouts because Runalyze metrics noted you experience a long run specific training benefit at 9+ miles (marathon […]

  3. […] Runalyze has my current Marathon Shape at a relatively disappointing 62%. Further training that will supercede the almost-zero mileage I had logged six months ago should buoy my shape to about 63-64% by race day. I had been hoping for 70%, and if I had acquired a Runalyze account back when I started training I would have gotten it to 100%. But that’s a lesson learned for next time, and water under the bridge. […]

  4. […] can measure your volume by mileage or rate of perceived exertion. But because Runalyze provides it to me for every kind of fitness activity I do, I’ve been using TRIMP, short for […]

  5. […] Runalyze tells me the TRIMP (stress) on these workouts is now about 20% lower. What this means is I can probably handle a large volume of these workouts without wearing myself out as much. I certainly don’t feel sore or worse than generally tired the day after these workout days. […]

  6. […] 90 minute threshold. (Side note: If I run farther than 8.1 miles on this or any other non-long run, Runalyze will classify this as a long run for Marathon Shape calculations, whether or not I label it as […]

Leave a Reply

%d bloggers like this: