Not directly (yet), but don’t let that put you off.
The first aspect to consider is that whilst the current TrainAsONE AI model does not incorporate the anticipated elevation changes (ascent and descent figures) of your target race, it is probably factoring it for you anyway. All trail races and ultramarathons are generally hilly and as the knowledge of the system is learned directly from other people training and racing such events, this elevation bias will be present.
Understandably, people wish to be prescribed / guided on elevation metrics for their training. Commonly a desire for a statement of ‘x metres of ascent / descent’ for a workout or per kilometre is sought – in line with fairly common coaching practices. The issue here is that such rules are made-up, with no scientific data to evidence them (if you know of anything, we’d love to see it). The use of such ‘rules’ is something we avoid within TrainAsONE, deferring to machine learning from real-world data.
Consequently, our recommendation for people using TrainAsONE when training for ‘hilly / very hilly’ trail races is to simply run their easy runs (and long runs) on as near a terrain as their target race as possible. Often before / around the peak of training, users take a weekend away to an appropriate location to put in some descent hill running. The results from this have been very positive, and our users have performed amazingly well in a number of the biggest trail races around the world.
Directly incorporating ascent / descent figures into the AI model is on the roadmap.
Comments are closed.