There are different options on the market for optimized AI training plans that are based on applying machine learning to your individual data. In this post, we compare the different options and their features.
We are big fans of TrainerRoad (TR) and their training plans which we have personally used extensively in the past. TR has announced adaptive training which "uses machine learning and science-based coaching principles to intelligently adjust your training plan, so you get the right workout, every time".
At this point TR is recommending one-off workouts 'TrainNow'. This feature looks similar to Garmin's daily recommended workout. TR is looking to incorporate more machine learning into their app in the future. For more info, see DC Rainmaker's post.
AI Endurance uses cutting edge machine learning technology (we even developed some our own methods) to create personalized AI training plans. The plans are based on your historical data and are predicted to give you the biggest performance gains for the time you have available for training. The AI is guided by best practices established around endurance training. It adapts your training as you progress through a training plan.
You can get cycling plans, running plans and also triathlon plans that take both your running and cycling data into account simultaneously. We predict your performance on your event date for the most common event types from 5k to Ironman.
You can also easily get your AI Endurance workouts into Zwift for both cycling and running. So if you're using Zwift already you can simply execute your personalized workouts there. You can also do your workouts directly from your Garmin device.
We believe that this is just the beginning of using machine learning to help endurance athletes improve. It is a classic problem where no human being can possibly wrap their head around all the possible ways we could structure endurance training and what the possible outcomes would be. AI however - can.
We recap the results of following AI Endurance’s cycling training plan that saw our FTP grow according to AI Endurance’s predictions, following the instructions on how to improve FTP.
by Stefano Andriolo. Building on previous work, we refine a method to accurately determine the relationship between DFA alpha 1 and power. This method can be used to track fitness and thresholds of an athlete. We find in some cases ramp detection tends to overestimate thresholds, a finding mirrored in recent physiological papers. On the other hand, thresholds based on clustering of DFA alpha 1 values tend to agree well with this new method. We propose a hybrid lab and everyday workout experiment to further study the relationship.
by Grant Paling. It has been a while. I hope you’ve all been ok. The winter months are hard. Darkness, cold, rain. It’s not always easy to get motivated to train. It’s also a time when maybe it’s a bit more appealing to stay on the sofa and watch some good TV - as you might guess from the blog title I’ve been watching Loki on Disney Plus, among other things (and yes I am a bit late to the party there).
We compare polarized training, threshold training and AI optimized endurance training. AI optimized training yields the best results, followed by polarized training with threshold training in third. The results are inline with current exercise physiology research. If the training composition is not optimized to the individual athlete, substantially smaller gains are to be expected.
