An important consideration is that TrainAsONE is not just scheduling a short run, but rather it is determining a combination of runs, the result of each affecting the benefits and risks for all following runs. The whole plan is a balancing act of the calculated benefits (of running) vs risk of injury. The former is based upon patterns of training that have demonstrated improved performance and better race times, whereas the latter is derived from user injury records and research papers.
In general, when a short run is scheduled at a time when you REALLY REALLY feel you can do more, the injury prevention aspect is probably winning – the system is calculating that potential performance gain is outweighed by the increase in injury risk.
An ‘issue’ here is that even with increasing the duration of the run considerably the performance gain is probably marginal, yet the injury risk may rapidly rise. However… even doubling the injury risk might be acceptable to an individual. Our ideal target is ~ 3% seven day risk during active training, and doubling to 6% still means that 94% of the time ‘you will be fine’, and so as an individual I’m happy: ‘I do more running, which I think must be doing me a whole lot of good, and I’m not getting injured’. Unless, of course, you’re one of the 6%. And as a population of runners, if all users were to do this then we would have doubled the number of people getting injured.
At TrainAsONE we see the needed system enhancement is information feedback and enabling informed decision making – the system should provide information on performance gains and risks, doing likewise for any run / plan variation that you or the system wishes to investigate (a complex piece of development with little to no existing science to fallback on, but we’re working on it).