KOM Informatics Adaptive Workouts transform the recommendations generated by the Interval Advisor AI into workouts specifically tailored for the athlete. You may be wondering how that works out in real life. In the software industry "dogfooding" (eating your own "dogfood") is the practice of using one's own products or services to test that product or service out in real world conditions. As KOMInformatics.com lead developer, and a competitive cyclist, I'll be using Interval Advisor both with (for indoor workouts) and without (for outdoors workouts) Adaptive Workouts.
My competitive season ends in August and doesn't start up again until the following April. I'll still ride 5 days a week, sticking with mostly Zone 2. I mostly will take a break from hard aerobic interval sessions, although I will throw in a few here and there. I'll do sprint sessions every 7-10 days or so, but reps are down compared to in-season sessions. Starting in late November, early December, I ramp up Zone 4 work, starting to prepare for harder sessions in January. Given this context, on Tuesday, November 2023 I got an Interval Advisor recommendation for Zone 4.:
Interval Advisor recognized that I hadn't done any Zone 4 work in the last 30 days, but still logged some extensive hours. It adjusted the recommendation from the 17:10 from the last session, down to 15:45, to give me an opportunity to reacclimate to this type of work. Interval Advisor provides a context for the recommendation by keeping the athlete apprised of the average and max workloads involving interval sessions of the selected zone. This information will show that the recommendation is not coming from left field, and should inspire confidence that the session can be completed successfully.
After listing the Zone 4 Adaptive Workouts, I selected Traditional Threshold. This workout is a progression which will take you all the way from 1 x 8:00 to 3 x 15:00. Clicking the [Generate] button yields a workout that is much the same as any other workout in the system. You can generate a projection for the workout, and schedule it. You'll want to do this, because on the ride that fullfills the workout, you'll get a Workout Performance report which evaluates how well you lived up to the projection. You can also open the Workout in Workout Builder which shows what you're in for:
Given my Recommendation, Adaptive Workouts is dividing the total into 2 intervals of about 8 minutes.
Next, I saved the Recommendation. On the ride that fullfills the Recommendation, you can provide feedback on whether the session was too hard or too easy. This will help Interval Advisor make better Recommendations in the future.
I wanted to run the Workout on Zwift because I like the visual cues Zwift provides when performing workouts. Workout Builder has a utility to generate a .zwo file, which you then need to save in your Zwift custom workouts folder.
The first interval went well. However, I developed some nausea on interval 2 which I attribute to eating peanuts to close to the start of the workout, and had to cut interval 2 short. I got downgraded on the Workout Performance Report as a result:
Since I never had developed nausea before in an interval session I decided not to mark the session as too hard just yet. The more interesting question is how would Interval Advisor evaluate this session in prescribing the next workout? After this Tuesday interval session, I did Z2 sessions on Wednesday and Thursday, and took a rest day on Friday, before generating another Interval Advisor Recommendation on Saturday, November 25, 2023:
Interval Advisor recognized I had a bad day on the first session, and dropped back to a statistics based Recommendation of 16:19 for the second workout in the intervention, not a whole lot different than the 15:45 prescribed for the first. In general, Interval Advisor will respect your most recent workout, provided you're performing at least an average level for you. If you're below average, it will use a statistics based Recommendation. I find that the latter type of Recommendation is usually a little harder for me.
The second workout in the intervention went well. I completed both intervals. The second interval felt harder towards the end, as it should. Although I missed the projection for average/weighted average watts for the entire workout, I exceeded the projection for the interval portion. As far as I'm concerned the miss of the second interval in session 1 was a fluke. I marked this workout as OK using the Recommendation Feedback utility. Here is the Workout Performance Report for the session:
After this Saturday interval session I did a 2 hours Zone 1 trainer ride on Sunday, and took a rest day on Monday. Then it was back to another Traditional Threshold session on Tuesday, November 28, 2023. Here's the Interval Advisor Recommendation from that day:
Interval Advisor upped the Recommendation from Saturday (16:37) to 17:37, applying the principle of progressive overload.
The third workout in the intervention went well. I completed both intervals. Although I missed the projection for average/weighted average watts for the entire workout, I exceeded the projection for the interval portion. I marked this workout as OK using the Recommendation Feedback utility.
This article is meant to show how well Interval Advisor guided my training over the course of 10 days or so. I didn't do a bunch of "before" tests that exercise physiologists typically do when they are validating an intervention. Also, 3 workouts over 10 days is probably not enough to draw any definitive conclusions about the efficacy of the workouts. That being said, the data thus far supports the idea that fitness was gained over the 3 workouts.:
The images on the left show the second session, the ones on the right are from the third. At the same wattage average Heartrate is down from the second (159, 153) to the third session (158,150). Humon Zone is from the Humon Hex device which measures muscle oxygen. Zones go from 1 to 5. 5 indicates that oxygen supply is dropping in response to demand at a very high rate. The closer the average Humon Zone is to 5, the more it indicates that mitochondrial function is not all it could be. Average Humon Zone dropped from the second (4.95, 4.47) to the third session (3.85, 4.43). The colored graphs show the Humon Hex zones over the course of both intervals for each session. Red shows Zone 5. So, the heart didn't have to work quite as hard, and the mitochondria were more efficient comparing session 3 to session 2. These improvements occurred despite the third session being almost a minute longer per rep than the second.
I always schedule my aerobic interval sessions the day after either a rest day, or a day when I do an "openers" ride. But what would Interval Advisor's response be if I asked it for a Recommendation the day after a tough interval session? Here's what happened on November 29, 2023.
After evaluating my ability to complete back-to-back sessions, Interval Advisor is telling me to rest up! People recover at different rates though. If you've demonstrated an ability to complete back-to-back interval sessions Interval Advisor will provide a recommendation for you in this situation.
Over the course of 10 days the Interval Advisor AI provided solid recommendations for 3 Zone 4 interval sessions. The workflow in KOMInformatics, getting an Interval Advisor recommendation, selecting an Adaptive Workout, and generating/scheduling a workout worked flawlessly. Although the intervention was too short to be definitive, the data thus far supports the idea that the workouts improved fitness.