10 ideas that I am taking home, after 4 days of learning with TrainingPeaks

2015 ECS

The past four days, I was able to attend two different trainings hosted by TrainingPeaks. The first two days was TrainingPeaks University, which was a small group of coaches that gathered to learn about using the TrainingPeaks software to efficiently and effectively train endurance athletes. The following two days, I was able to attend the Endurance Coaching Summit where there were approximately 150 coaches that came together to learn about the business and science of coaching endurance athletics.

The following is not a comprehensive list of things I learned over the four days, nor a regurgitation of notes that I took. The following are 10 ideas that stick out to me as key take-a-ways as I think back on my experience:

One:

I don’t use all the features associated with TrainingPeaks to work efficiently. This includes features that I did not know were available. My ignorance to these features is partly due to the rapid iterations and updates that they have been going through over the past couple years, but it is also a result of my own workflow and getting into a habit of doing things and not looking to change. The other fact is that my primary use of the software today is to support the running group and not coaching individuals. They have made a lot of progress with the software for this purpose recently, so understanding how to effectively manage the software for this use case is still a work in progress for me. There are some features and services that I had previously written off as a poor business decision, but will likely rethink this to see if EBC’s coaches and the running group could use these features.

Two:

It’s easy to get caught up being a technician and forget that you are coaching an individual. Always understand what you are trying to accomplish with the training, then find the best tools for monitoring. Don’t get caught up in making graphs look pretty and all inclusive that you forget that pretty graphs don’t win competitions.

Three:

When it comes to competition, understand that the athlete is a whole being. This means that a coach needs to consider the psychology of the athlete, not just the physiology and data. A great talk on the concept of pain and suffering was given by Carrie Cheadle, which shared some coaching strategies to help athletes understand how to become comfortable with suffering, allowing them to have their best performances.

Four:

There was a lot of focus on device and sensor technology, associated with the collection of the resulting data. However, very little discussion about the potential associated with advancement in biological and genetic sequencing technology. I would have loved to hear a scientist, researcher and/or coach talk about what will happen when doing full blood workups is simple and accessible (I.e. Theranos) or what role they see genetic sequencing playing within exercise prescription or race planning in the near future (and 10 years from now).

To steal a line from Peter Diamandis, I want to know where the exponential technologies are in endurance sports. While I loved to listen to all the discussions about power meter files and analysis, I don’t believe that a power meter represents that exponential technology. I don’t know what it will be, but I do think in 10 years we’ll look back and it will have seemed obvious (The iPhone still isn’t 10 years old today, 10 years from today… it’s hard to envision).

Five:

Individualize the program based on scientific principles and analysis, because generic prescriptions based on standardized zones are not relevant to many athletes. Dr. Inigo San Millan had a session that discussed monitoring the elite athlete. He shared the value of testing athletes, to specifically understand the individual’s metabolic fitness so that the training is providing the training stimulus they need. He also discussed why this is necessary so that the athlete isn’t put into a program that leads to overtraining. At the end of the summit, he lead a tour of the facilities that are being built by UC Boulder, which will provide a comprehensive athletic testing facility. The great feature about the facility is that there is an entire floor, with everything a world class facility needs, that was specifically added for community access.

Six:

Low Carb plans are not good nutritional strategies for high performance. This is something that I consistently hear from athletes. It may come as a question, “What do you think about doing a low carb diet, to help me become better at oxidizing fat?” It may come as a statement, with a question tagged on, “I’m going to do a low carb diet, how can I perform well while maintaining this diet?” It was very comforting to hear Dr. San Millan touch on the topic briefly. It clearly made a few coaches in the room uncomfortable, as it likely imploaded their world view. However, the research is clear. If you want to perform well, then consuming adequate carbohydrates and sugars is necessary.

Seven:

The predictive models and data analysis process provides a good “learning mechanism” for the coach. In different talks, said in different ways – this was a message I took away from Ryan Cooper (Founder, Best Bike Split) and Robby Ketchell (Data Scientist for Team Sky). It was Robby that used the term “learning mechanism”, which I interpreted to mean that these models give us a framework to run our collected data through, providing insight to the meaning and value of the data. However, it’s important for the coach to be able to understand what the variables are that are inputed into the models. The learning happens when the coach can then start to run the models with variations in the variables, observing the different outcomes.

A great example was when Ryan showed how two different variables could influence a rider’s bike split. He showed the difference between a rider losing 10 lbs versus getting a more optimized aerodynamic position. The result was that an athlete would be more successful focusing on getting aero before they worry about losing the weight, because the benefits of an aero position have greater impact.

The other insight I had while listening to all these discussions, which included Hunter Allen reviewing the new WKO4 software, is that as a coach it’s best to keep your modeling as simple as possible. Then only make it more complex as needed. It’s very unlikely that I will ever use models as complex as Robby does when he is deciding the strategy for Team Sky’s Team Time Trial, but that doesn’t mean I shouldn’t use the data at all.

One final take-a-way, that is aligned with this thought was when Robby said on multiple occasions that we are not to a point that we truly have “Big Data” in endurance sports. There definitely is a long way to go and opportunity for advancement.

Eight:

The data out, is only as good as the data in. The models for monitoring training and predicting performance are only as valid as the data that an athlete, coach or data scientist can put into the model. If an athlete decides to skip a week and not upload their training data, it takes away the value of a metric such as Acute Training Load. Therefore, compliance with uploading the data is incredibly important for monitoring and prescribing training. It is not only about compliance, it’s also about collecting bad data from the devices themselves. Power meters that are not calibrated, power spikes in the file, heart rate spikes and uploading files more than once are all examples of bad data going into the models that seem to happen. This means that either the athlete and/or coach should plan to spend time each week cleaning the data that gets uploaded.

The issue of bad data was discussed by most of the individuals on the final discussion panel. They also acknowledged that this is an area that they see a lot of advancement in the near future. I personally know that the ability for files to auto sync into TrainingPeaks from Garmin Connect has really improved compliance for many of the runners in our running group.

Nine:

Be obsessed. Dirk Friel held a fireside chat with Brad Feld, where Brad discussed the value of being obsessed with your product and/or service. Being obsessed is different than discipline or passion. Passion is easy and discipline can be difficult to sustain, but having an obsession with your service will set you apart.

I think that I have this covered. Sunday night, after the first day of TPU, I had a hard time sleeping because I was thinking through all concepts I wanted to fit into my modeling (Heart Rate Variability being the biggest variable I want added in). When I went out to eat with Eric Atnip on Monday night (he’s in town for Ironman Boulder), I mentioned how I was on the first full week of vacation that I’ve taken since joining Retrofit. His response was that I was working. That may be true, especially because I spent a lot of time thinking about how the principles being discussed could be applied within the weight loss world. It sure doesn’t feel like work.

Ten:

Networking. If all the above was not enough, I was able to network with other coaches, some individuals focused on data analysis within endurance performance, along with individuals that are working on the device and hardware piece. It was also valuable to meet the folks at TrainingPeaks.

This was the first year that the Endurance Coaching Summit was held and I hope that it is put on again next year, because I will return just for the ability to meet interesting people that are trying to accomplish the same things as I am. The future is very exciting when we can spend a little time to discuss how to use technology to further health, fitness and performance

Leave a Reply