After signing up to run a race, amateur runners usually search for a training plan to help them prepare. Bearing in mind the number of weeks until the race, the frequency with which they want to train and their current mileage, runners are faced with a plethora of race training programmes and usually go with what looks the most manageable as well as what their friends or family have used in the past. All in all, a lot of guesswork is involved.
However Dr Sean Radford wanted to replace that guesswork with science and data, and so created TrainAsONE. It was described by the chief technology officer as “an award-winning, AI-powered running coach."
"Everytime you run, the system adjusts. It is very efficient."
Based on the user’s running history and availability and the date and distance of their upcoming race, the system generates an individual running plan. David Brownlee, the CTO said that there are several benefits to running with TrainAsONE. “It is entirely adaptable. Every time you go out for a run, or if you miss a run, the system adjusts so it is always on track. We massively reduce injury risk. About 56% of runner get injured every year. We target 3%. And because the individual distances, durations and the mixture of workouts are targeted at the specific runner, it is very efficient training, so people need much less exercising in order to get fit and healthy.”
“We ask someone to do a nice, gentle run. That data doesn’t lie.”
The system works by getting the user to submit information about themselves such as age, gender, ideally height and weight, as well as their running history. If the user has been recording their runs with an app such as RunKeeper or Strava, the system can connect to the app and draw out that historical data. Everyone starts by having to do an assessment or “perceived effort” run of about eight minutes at a pace at which they can hold a conversation. That run data is the baseline for the system which is used to start building the plan. The premium service even adapts the training runs to the weather on the day and the undulation of the terrain. “We ask someone to do a nice, gentle run. That data doesn’t lie,” said Brownlee.
He pointed out that while TrainAsONE takes away the need for some aspects of a personal trainer’s service, it could be used as a complementary tool to make the personal trainers more effective. The nature of an AI system means that it cannot offer any advice on nutrition or running technique. “Nothing can match and provide that personal impetus of someone working with you,” Brownlee explained.
“By default we don’t share any data.”
To protect the privacy of the runners, users that want to share their running calendar or their next run have to request a link, and then state whether they want it to be shared on their iCal or with friends. However, this does not include the time or location, and users can delete all of their data if they choose to no longer use the system. “By default, we don’t share any data,” Brownlee stated.
"If you run with the system, we use your data to generate the plan."
That is not to say that the power of collective data is not harnessed. There are now 15 million kilometres of running data in the system. “If you run with the system, we use your data to generate the plan, and then we analyse that data across people in order to enhance the plans for everyone,” said Brownlee.
The first data sets in the system came from published medical research on the physiological effects of running. The original plans were built into the system to start with, and from that the AI and the machine learning has evolved beyond those plans. The TrainAsONE system was developed by Radford who is a medically trained doctor and a mountain marathoner who has been training athletes for a decade. Dissatisfied with the running research available that was principally based on data from elite marathon runners in the 1980s, Radford started creating training plans himself. With that open data as a basis, TrainAsONE now has a proprietary system to generate running plans.
"We could prescribe running workout to reduce blood pressure."
TrainAsONE is taking part in the OpenActive Accelerator, and is looking at ways to combine running training with training for other sports such as cycling and swimming to promote general mental and health wellbeing. Brownlee said: “We already know right now we could prescribe a combination of running workouts that would increase insulin sensitivity or reduce blood pressure by 10 points. We’re looking towards delivering something above mental wellbeing and general fitness – actual improvements in health.”