Athletica results and impressions after three months, what comes next for off-season?

[NorthK]

Hi all, I wanted to give a summary of my Athletica results as a beta user since starting in May, then ask 1) how is Athletica working for you, and how do you measure that, and 2) how are you all using Athletica during the off-season?

So, overall – my goal was to beat my all-time Strava PR’s in local climbs of 9-25 minutes, which would keep me in the all-time or at least the current year top 10-20 riders in my age group for those segments. I did not achieve that goal. But, I’ve come really close: 15 seconds slower than my best time on a 9 minute climb, set back in 2018, and my 5th best time ever on a 25 minute MTB climb. It seems that my performance has improved more for shorter efforts (10-12 minutes), then longer efforts (25-30 minutes). I attribute that to a whole lot of VO2 max and short threshold interval sessions from the training plan.

I can’t say how my FTP has been affected if at all, as testing FTP is no longer part of the program and I haven’t tried it on my own. Though I really would like to know so I might do that anyway.

The training has been tough but doable – I generally have two interval rides per week on the Athletica low volume plan, when I used to do one per week on my own training program. It’s been hard to fit those two interval rides in with two strength training sessions and longer zone 2 rides. Barely enough recovery if I do everything right – eat correctly, sleep enough, meditate and otherwise manage stress from work and life. Some days I needed to do an easier ride than what was on the plan.

For the app itself, I had some difficulty getting started but then it’s mostly been pretty smooth except when the A/I engine wants to schedule two intervals in one day, or suddenly give me three interval sessions in one week (which I didn’t do), or when it randomly wants to change my heart rate or power zones (which it seems to get mostly wrong). Outside of that it seems to be getting the progression right. I have to say I don’t understand most of the graphs and charts well at all – the Performance Potential, Power Profile, Time In Zone (I almost never get a close match for this), and Session Chart. I’ve been staring at them for months and I’m still not sure how to apply what I see in the charts.

Some improvements I would like to see are:

• Use of hours slept (from a wearable such as Whoop) and resting pulse (probably needs something more accurate than Whoop) every morning to gauge recovery, and use it to automatically modify the training plan. I’m convinced HRV is not good, but resting pulse could be.
• Allow for some free-form riding built into the program every week, for example a Saturday group ride. And allow for automatically recalculating the rides for the rest of the week if I want to do something that’s off-plan such as make an attempt at a PR.
• More understandable and cleaner charts and graphs.
• Continued refinement of the ride scheduler so it knows not to schedule 3 interval rides in one week (I’m too old for that!) and so overall it sticks to the schedule I gave it in the beginning better.
• Simplified U/I – there’s so much in there and most of it I don’t know how to interpret and I probably don’t need to. I’d rather have the A/I engine interpret it all and give me really good recommendations and modifications to the plan. Some people will want to see all the numbers, and I do look at them after every ride. But I can’t say that looking at them makes them actionable or helpful to me.
• Some way I can gauge my progress, built into the plan. Since there’s no FTP tests (for good reasons), I don’t have a way to see if I’m on track, getting stronger, getting weaker, or staying the same for specific durations and effort levels. Unless I go off-plan and try for a PR. Seeing progress is a really important motivational tool when I’m doing hard intervals every week! This kind of progress in a chart could be very motivating.
• Lap data so I can compare my average power for each interval with previous rides to gauge progress (see above bullet point on gauging progress).
• A way for the A/I engine to run experimental training blocks and customize the plan more for my physiology. Would 2x15 or 3x8 FTP threshold intervals work better for me than 7x6 minutes? Would 4x4 VO2 Max intervals work better than 30-30’s? Would two FTP interval sessions per week work better than one FTP and one VO2 max session at certain times in the training plan? And even, would alternating blocks of Sweet Spot intervals followed by Polarized blocks work best for my particular body? The plan seems to stick to the same types of intervals each week but it would be so cool if it could run intelligent experiments on me and optimize the rest of my plan based on my specific training response.
• Built in transition to off-season training, and back to in-season training.

All in all, using the app and following the plan does seem to have made a big difference in shorter (less than 10 minute) efforts, and also in my average speed for zone 2 rides. I may yet be able to break my PR for the shorter hill climb this year. But, it’s been a grind with all the intervals and no “fun rides”, group rides, motivational comparisons, or free-form rides built into the plan. At this point in the season I feel burned out following the plan every ride, and I want to go out for some rides and just do whatever I feel like some days. Truthfully, I’m asking myself if I want to follow an automated training plan at all. This is a problem with any automated plan (not just Athletica), but perhaps there could be a way to build in free-form fun rides, auto re-calculation, and still generate a progression.

Big thanks to @Prof and the developers for working on Athletica, being patient with me, answering all my questions, and to everyone who responded to my posts and questions so far! If you read this far, good on you and please share whatever thoughts you have, and how Athletica has been working for you. Last but not least, I’m wondering how I can transition into an off-season training program in September. I don’t think I have it in my to keep doing two intervals every week year-round. I’m thinking of going onto a Sweet Spot/zone 2 program but I’m still debating.

Thanks for reading.

[Andrea]

Dear @NorthK, this was truly an interesting read. Thank you for taking the time to write such a detailed review, and congrats to you for sticking with the Athletica plan for so long! We couldn’t be prouder.

Many of the points you raised resonate with us, and they are reflected in our roadmap and pipeline. Others, while equally interesting, might take a bit more time to develop. Let me go through your message point-by-point to better address your feedback:

• Use of sleep data: We agree this is a valuable area to explore. However, I see a couple of challenges: 1) integrating with third-party devices such as Whoop, and 2) the reliability of sleep data itself. We’re very interested in adding more integrations, particularly with Whoop, but it’s difficult to say when this might happen. Are sleep hours accurate across all devices? In my experience, only to some extent. The bigger issue is determining how to effectively use that data. We could include a notification like “Your sleep was poor, consider adjusting your training load,” but that brings us back to the core question: shouldn’t the AI coach be able to make those adjustments for you? Unfortunately, if the guidelines on how to use this information are unclear, even the AI won’t have a perfect solution.
• Free rides: You’re right—free rides are important. You can always go off-plan for a day if you feel like it, and the future load will be automatically adjusted. The AI logic can adjust for unplanned rides, but there’s only so much it can do to prepare for rides with unknown duration or intensity. Adding an option in the Workout Wizard for a “PR attempt” is a great idea and something we’ll consider.
• Graphs: We’re working on improving the explanations for the power/speed profile, recovery profile, and performance management graphs. These updates should be available soon. We recognize this is an area that needs improvement, and we’re actively working on making these visuals more user-friendly and actionable.
• Simplified UI: Absolutely, 100%. We started with a more crowded UI, but we’re in the process of simplifying it down to the essential features and functionalities based on user preferences. Your point here is well taken, and we’re working on it.
• Gauging progress: We agree that this is crucial. With FTP testing being phased out, we’re relying on other metrics to monitor progress. Displaying progress and the evolution of these metrics is something we’re aiming to introduce on the platform soon.
• Short intervals: These tend to dominate many training plans, and I’ll let the Prof chime in with more insights on this. From a data processing standpoint, it’s extremely challenging to automatically determine whether a long-HIIT session was better received than a short-interval session. Our plans are based on the latest evidence, which suggests that short 30":30" intervals are among the most effective for enhancing VO2max while minimizing the impact on anaerobic and neuromuscular components. However, you raise a good point—there should be a feedback mechanism to allow users to express or select their preferences. We’re discussing ways to introduce more variety, particularly through the Workout Wizard, where you’ll soon find a broader range of HIIT sessions.
• “Train to maintain”: This feature was designed for handling transitions and the off-season. Personally, I often use “Train to maintain” mode and, when I don’t feel like doing intervals, I use the training library to select more steady-state exercises (which happens quite often!). That said, I agree with you that this option could be better tailored to actual transition and off-season periods, with more sweet spot/Zone 2 sessions aimed at building aerobic capacity and fat-burning efficiency. Currently, it pulls from the existing plans with the goal of maintaining fitness, but there’s definitely room for improvement.

Thank you again for raising these points. We’re committed to continuing our work to meet your expectations and help you get closer and closer (and eventually reach) your goals.

[NorthK]

@Andrea thanks for your detailed response, and I’m glad to see the Athletica team is working on many of these features. I’ll give a few more thoughts:

• Free rides scheduled into the plan – I think it’s really important for the A/I tool to schedule free rides right into the plan, rather than having to go off-plan for them. I think when the tool “gives permission” to do a free ride, then it feels better like I’m staying On Plan. That might sound silly, but I think sports psychologists will tell you that there is great satisfaction in doing something challenging, completing each workout, and following a plan to get there.
• Long intervals – I think that one of the things missing for me on the Athletica plan was long intervals. I did get some good benefits from the short intervals but I think longer intervals at or slightly above threshold can give benefits for me that the shorter intervals don’t. So, I would say that both are important but the A/I plan only gave me short intervals. This made me fast for short climbs (9 minutes) but didn’t help much for longer climbs (30 minutes).
• Train to maintain – I was on that plan the entire three months I used Athletica. But it still gave me a progressive plan with a lot of intervals – 2x or even 3x per week. That was too many intervals after three months and I reached burnout stage. I think it would have worked a lot better for me to do some steady state rides instead, and I think it should be built into the plan. Additionally, a lot of coaches do 3 weeks on/1 week off to get better rest, which I think would have been good for me. Now that I’ve done it, I see that doing 3 months of intervals is just…too long. I would like to see the plan change modalities with different phases over time. And that is for the on-season. For the off-season, I think what you outlined could work really well (a mix of sweet spot and zone 2 sessions).

[ZHVelo]

@NorthK

• Use of hours slept (from a wearable such as Whoop) and resting pulse (probably needs something more accurate than Whoop) every morning to gauge recovery, and use it to automatically modify the training plan. I’m convinced HRV is not good, but resting pulse could be.
• Allow for some free-form riding built into the program every week, for example a Saturday group ride. And allow for automatically recalculating the rides for the rest of the week if I want to do something that’s off-plan such as make an attempt at a PR.

I agree, I don’t think HRV overnight is useful. I have a Fitbit and it doesn’t tell me anything. Sometimes it does align (as in HRV is way off the normal range) with how I felt I slept, but mostly, I don’t see any use in it.

For number two there is a workaround I already found: you can just add a manual workout. What the plan gave me for weekends basically is Saturday some form of zone 2 and Sunday some longer ride but with intensity. So I just entered for Saturday my ride manually (I know roughly the load it will be, because I am a nerd and plan all my long rides including power estimates and entering them in calculators to see how long each segment will take, and other things like finding water sources along the route and knowing how much sugar to bring) and moved the zone 2 ride to Sunday altering it so that the overload weekend load is the same as it was under the original plan. It’s some effort, yes, but it works. It would go a lot quicker if entering manual workouts could be done by free, shorthand text rather than how it’s done now (I mentioned that in my post).

@Andrea

• Short intervals: These tend to dominate many training plans, and I’ll let the Prof chime in with more insights on this. From a data processing standpoint, it’s extremely challenging to automatically determine whether a long-HIIT session was better received than a short-interval session. Our plans are based on the latest evidence, which suggests that short 30":30" intervals are among the most effective for enhancing VO2max while minimizing the impact on anaerobic and neuromuscular components. However, you raise a good point—there should be a feedback mechanism to allow users to express or select their preferences. We’re discussing ways to introduce more variety, particularly through the Workout Wizard, where you’ll soon find a broader range of HIIT sessions.

My suggestion would be for flexibility in plan choice from user. So it does the whole plan for months and months right, if you give it your A race in say 7 months time. Partition those 7 months in meso and micro cycles and allow the user to choose different options for each meso cycle (say 4 weeks). Then the user could let the plan be as is for 3 months, but then decide for 2 meso cycles they want to use a plan that has longer intervals, so it clicks that option and the tool updates those two cycles with long intervals. That gives flexibility to the user and gives the tool data to compare over a longer time frame how the athlete reacts to different interval types. You could even allow the tool to analyse after the first cycle and tell the athlete “hey, looking at your data, it seems you respond better to shorter intervals, we suggest you go back to them”.

Which also is another thing that would be nice, you don’t see what plans are out there (at least I couldn’t). I gave the plan my general aim and A Race, and it did everything. But if you look at the weeks you can tell it has different plans because every week has a specific name. Cycling All-Rounder / Single Day Event / Base / 1 mid. So somewhere there is a database with all of these and how these may be put together (1, 2, 3, recovery for example are the last 4 months of build for me). Would add a lot of flexibility if users could try some of them. And yes I get that the tool also is supposed to plan your training, and I suppose that is what most people want (and since it’s optional wouldn’t be affected because they can just ignore this option), but personally I want the feedback, not the planning, that I can do myself (in fact, the plan I got is surprisingly similar to what I had planned for myself, which both gave me confidence in it and made me feel validated in how I’ve been self coaching) so letting me pick some different options that I feel would be better or that I just want to try for variety, would be pretty cool. And yes you can manually already change whatever you want, but that’s a lot of work.