 From Las Vegas, expecting the signal from the noise. It's theCUBE, covering Interconnect 2016. Brought to you by IBM. Now your host, John Furrier and Dave Vellante. Okay, welcome back everyone. We are here live in Las Vegas for IBM, Interconnect 2016, special presentation of theCUBE. I'm John Furrier with the course Dave Vellante, Silicon Angles flagship program that theCUBE is on the ground. We are here with Neil Henderson, high performance consulting with USA Cycling. Apex coaching is the hashtag. Great to have you on. We're going to talk about sports data in Las Vegas. You know we love sports data. Welcome to theCUBE. Thank you very much guys. So sports and data is changing the fan experience, the athlete experience and the sport itself, well owner of the team or in this case, cycling, national cycling team and whatnot. You're doing some pretty compelling things. So first share what you're doing, how you got here, what you're currently doing right now. Yeah, so track cycling itself is a really old sport. It started in the early 1900s actually was the biggest spectator sport in the United States back then. It's gone through some evolution and kind of dropped a little bit. What we do with track cycling at the Olympics is there's about five different medal events. I work with our women's team pursuit program. And specifically we have four women who are covering basically 16 laps of a 250 meter track and they're trying to do it as fast as they possibly can. We have a lot of tools available to us in the very beginning. It was simply a stopwatch writing down splits and we see basically the speed that they were going. We've really advanced the technology and tools that we're using now to monitor the athletes during every training session during competition and even assessing recovery aspects outside of competition and training. So share us what you have here. You have some props, it's the old way notebook, stopwatch, evolve the spreadsheet on the BC. Now you have iPads, you've got sensors. What do you have here? Correct. So we don't have everything that we use here on the bicycle itself. There's an instrumented power meter. It has strain gauges and it's measuring the power output every half second during the ride. So we also have cadence, speed being measured there. We have a standard heart rate monitor right here. We have something that's a little bit newer which is a BSX muscle oxygen sensor. It uses near infrared spectroscopy and we are actually measuring the amount of oxygen that is being utilized in a given muscle group. This one would go on the calf and be telling us about what's going on inside the muscle from an oxygen perspective, which again until recently would have been a very invasive procedure. We have then when we are riding we are using the cloud services and the IBM JSTAR team has worked with us in developing app where the information is being collected from all these separate streams and devices going up to the cloud and then coming back down to our pad here, iPad where we can actually see in real time what's going on as well as have a really full breakdown of each rider's data because we do have four riders at a time. Each of their pieces of information we are able to select and see. The next thing we are adding to this is actually the wearable technology with the solos glasses. You gotta put those on. That's good, it on. Oh yeah, there we go. And on with the solos glasses we're able to customize which feeds of information are being presented to the riders in real time while they are riding on the belt. And what are they going to do with that information? What are they going to see? Yeah, so we can customize a few things. So there's a certain amount of energy that they have to expend to go a given speed and if they say miss an exchange which the rider from the front will move to the back because the amount of power required on the front is much higher, as they move to the back if they miss that exchange they have to expend a certain amount of energy to get back just inches behind the rear wheel of the riders ahead of them. They see that they've made an expenditure of energy and on one of our charts here it breaks out. Really, if their energy tank based on all previous rides is running low and they may have to then take only one more pull and get out of the group because we start with four riders though only three have to finish. So one person can burn out. They tap out, literally. Yeah, they're out and they pull out and out. Just go until you blow, right? Yeah, so this gives them an a quantitative measurement of actually I don't have the energy to take one more pull because if they slow the group down. Do the other riders see the other riders results so they can kind of like it's kind of a team approach like look I'm going to be the weak link here I'm going to drop back? Correct, so with the information each rider in this graph is actually overlaid in one graph but we can select one single athlete and just show their information alone and share that with them. But then with the team, we review everyone's data. And the rider has any control over what he or she sees or you guys control that from a central summary? So with the solos we're going to be customizing and again, we have some athletes who may be very visual and they're going to want to see red, yellow, green light in terms of energy level. Somebody else may be much more quantitative in their aspect and they may see the percentages. They're at 70%, 40%, 30%, 10% rather than color. So we can customize what type of information is being displayed there and even how that is graphically displayed. So I'm struck by the old school notebook and stopwatch and even to this day you see it at like the NFL Combine and do you have, is there friction in terms of adopting this technology? Talk about the sort of that component. Absolutely, I've been involved in track cycling for about eight years. I had an athlete who had two parents who were Olympians. One of them was a world champion and an Olympic gold medalist. The other one was the father was a bronze medalist at the Olympics. So when I started coaching him there was a lot of pressure and expectation in terms of being a phenomenal cyclist which he had the right genes as a starting point. That always helps. If you want to go to the Olympics choose your parents well. So he had the good starting point there. And I came into track cycling with him literally as a novice to track cycling but I'm a scientist. I'm a data scientist. I'm a physiologist and the coach. I was a swim coach first. And in that way we measure and quantify certain things. I approach track cycling in the same way that I approach some of those other sports using data and information and in training in a way that the old school coaches the people who have been doing it for decades for years and passed down through again this is a hundred year old sport. They looked at me like I was crazy that I was doing things that shouldn't be done or couldn't be done when in fact they worked and Taylor Finney is the cyclist I coached he won two world championships and competed at the Olympic Games as an 18 year old and finished seventh. What's the learnings that you could share with folks? We have a lot of people I'd certainly in California where I live a lot of cycling going on out so you're in Boulder. What would you share with the cyclist enthusiasts out there who goes out every week and rides with those friends? What have you learned? Is there anything that you could deduce and just share with the folks out there? Resources they can get involved in and do they have to strap everything together like this? Yeah. So there are a lot of different tools available at this point right now that are sometimes proprietary. There's this piece of software that'll evaluate this this piece of hardware that'll evaluate that choosing something that kind of fits what your needs and goals are is a big part of that. Some people, again, we're gonna have I have a lot of engineers that I've coached over time that they want to see everything. And so, you know, we talk about what those options are and how to place those things and what to look at, what's critical and what's not because we can get a little bit encumbered by too much information. What is critical? So we look at things like your individual response and your power output, speed for a given segment of a climb which there's some software packages that'll show you that automatically. What would be something that you could share with the weekend rider that they might not know about or they're probably reading some stuff on the block for the most part? What would you say that you've uncovered insight-wise that might be new information? Yeah, I think something that we're gonna see in the future is some analytics where we're looking at an individual's response to exercise. I would say as a rule of thumb, most people train at a middle level intensity and they can get a little bit fit but not change their performance. If you want to be truly fast, you have to be doing very hard work but then also actually fairly easy work to recover. You only make change at that upper limit, not in the middle. You're not making change in the middle. So you have to push that but then you have to recover from it and that's really a big takeaway that we have. So think about when you used to train either yourself or you used to train athletes years ago and how that's changed now, especially in the context of what you just said because intuitively you're sort of always new. They used to tell us to do long, slow distance and then go for it, right? Okay, but so you kind of intuitively knew that. Talk about what's changed between how you change a high performance athlete today versus say 10 years ago. Yeah, the individual pieces of information that we can gather from their physiology and their output becomes so much more powerful in that way again, pun intended, cycling power, huh? If we look at say heart rate, if you rewind 20 years ago, even people were still talking about 220 minus your age, that's a horrible, terrible dead in the water formula that has nothing. It's like shoe size. I might as well say, you know, size nine plus or minus like, you know, your favorite color which would represent the scale and that shoe may fit you just cause you're lucky but that's not the bell shaped curve. You have very normal people at the extremes of that and so a high heart rate isn't better than a low heart rate. It is what it is. It's like hair color. You have what you have or don't have it. So the individual response to what you're doing, your heart rate, your power output, your speed, you're able to look at that now and in the future we're gonna be seeing the analytics of where you respond best. What type of things will give you the most benefit that you do in training? And that benefit is measured in ultimate performance, right? But then you must break that down into its different components. I mean, there's a lot of science that, you know, at doing endless cardio versus building up muscle isn't gonna do anything for you. I mean, so the world of training has come a long way, hasn't it? Absolutely. I mean, we look at what the genetic tendencies are for a given person to respond to a type of training. So these type of intervals for a given person because of their genetics, they're gonna respond better. And if we're capturing enough data from these different streams, we can see that change in performance and be able to make rapid adjustments to their training schedule where in the past we would have somebody do a certain type of training for several months and see if there's a change in performance or not. And if not, then we go to the drawing board and we do a few more months. We can now see that in days, in weeks, that responsiveness and adjust and tailor that training to their individual response. Do you take it into diet and sleep patterns? Absolutely. So our women, our team pursued athletes are currently using sleep sensors that are placed under their bed, under their sheets, and it tracks sleep quality, quantity, because that's a big aspect. We have to do the hard training, but we have to recover from that as well. We have nutrition aspects that we are controlling and evaluating and measuring. The next big thing then is also gonna be the psychology because a competitive athlete, there's an aspect that is purely psychological and performance. When we get to the high level of sport, physically, physiologically, everyone is very good at that high level. When we're at the Olympic Games, those who succeed are often those. So what is that? You hear the term locked in, that athletes locked in. Is that a mental psychology trait? And you see that all the time? And can be trained. And right now, we have rudimentary tools in terms of psychometric analysis to evaluate what's going on. And we do things in training like, say, visualization, self-talk. These are, again, pretty low-level things, though they have high-level impact on performance. When we start to actually be looking a little bit at the neuro aspects of what's going on and developing mental strength, not just mental toughness, but strength and skills that impact our performance, I believe that's, again, a future area that we're gonna see exploited as we do now in physiology and mechanics. Neal, this is great stuff. Love this segment. So you mentioned the heart rate kind of formula is kind of old model. What are the myths can you put to bed, if you will, for this modern era? Obviously, the data science, the science is getting behind it. Can you share some other things that you guys have busted through in terms of myths out there that were kind of like old wise tales or old doctor kind of ways to think about things that you guys have actually figured out? Yeah, in cycling, there's old wives tales that exist, just again, because the sport's been around for a while, a lot of it comes out of Central Europe and there's the Belgian logic or biological thinking, we call it. And it's as simple as things like nutrition. Like if you eat dairy the day before, you're gonna have poor performance. If you take a bath or get your legs in water that they're gonna get heavier. There's stuff out there that people are still espousing that in Central Europe. Fortunately, we don't have those crazy biological thought patterns in the US that we don't have to do that too much, though, when we work with riders from other countries or when our riders go to Europe and there's a director telling them this, that, and I get a text, an email from an athlete, my coach said I shouldn't eat this because some crazy reasons, like, nope, that's not okay, that's kind of an old wives tale. Nutrition, smarter than Korea. Big area, and so, again, when we look at macronutrient content, carbohydrate, fat, protein, there's a balance that's important. Super high protein is not helpful. Extremely high carbohydrate is not helpful. Kind of a middle way in most cases. So it's a balance and equation. You guys have some sort of formula. Yeah, for that person in assessing again. And that's on a per person basis. We are actually using a technology where we're looking at basically an ultrasound image of muscle glycogen. So we are testing before and after training sessions. So glycogen is the storage form of carbohydrates. So we are looking on the legs at the quadriceps and calves for our cyclists before and after training sessions to assess how much of that glycogen they've used. And also in successive days of training, are they repleting enough? We can see objectively that their glycogen levels, they started at this, they are now at 50%, they're at 30% starting a training session. We know that we actually can't ask them to do the workout as originally planned. We need to adjust. And all this data sitting in some IBM infrastructure somewhere and there's IBM analytics. Tell us, what are you using that's IBM? Yeah, so the app that's on the iPad here is one that's been developed for us. And if we look at this, there's an overlay of power output versus time. We can change it versus time or distance. And we have multiple different riders. I can go ahead and select down to just a single rider. So the purple graph here is showing the power output in watts versus distance or time. If you hold that up, you'll be able to get, I don't know if you can get that or not. So the power output in watts in the start, they are starting on a fixed gear bike. There is no coasting, it is a direct drive. So they start in the same gear that they finish. They are going from zero to 35 miles an hour within 12 seconds and then holding that speed for the remaining four minutes. So the power output is a very huge demand to get that acceleration. And then with the exchange, you'd see an area where there's lower power while they're sheltered behind the riders. And then when the riders ahead of them pull off, when they're on the front, there's a much higher power output. You guys have this down to the on the minutes and seconds. Down to every half second, this data is being collected. And then we overlay each of the other riders on the team here. And then we have a more complete picture, exactly, and up here in the white then we have the speed of the group at the front, not one individual rider, but net, it's the team aspect. That's what happens. The amoeba. The amoeba, exactly. We are working together to optimize that. So, okay, so that's an app that IBM, one of the hundreds Swift based apps that IBM wrote. What else? Is that that's running in an IBM cloud somewhere? Are you using IBM analytics? Yeah, this is cloud and then coming down for the analytics. The other aspect then is the connection to the solos, to the glasses. That's an IBM architecture of, again, pulling this information, customizing what we want then sent to the Bluetooth glasses, customizing that. So, another app that has been developed with this data that's being collected from the multiple streams and we can select which things are going in there. And you get the oxygen intake. Yep, the muscle oxygen level is there. The heart rate, all of those we can select which is most critical. Well, I got to ask you the hard question then because this is going to be very, probably maybe easy for you, but you just so many cool things that you're working on. What is the coolest thing that you've done that you could be crazy wild, but you're in the data, you're doing the science. What's the coolest thing? The probably the coolest thing is literally this app and the information that we have in the past, it would literally take a four hour process to break down one ride of four riders to get some useful information. We were, the guys from the JSTAR team were with us in Los Angeles this past week. We were doing training sessions, the girls finish and they said, do you want to see the data? A second, literally after they finish, the data is done and compressed. Do I want to keep the selfie? Exactly, and for me it's mind boggling. That was four hours of my life that they just saved that now I can show the athletes and give an immediate feedback loop and they're seeing this for the first time ever again in that real time environment. You're coaching on the spot, so you get the data instantly available, real time and you're giving real insight, real coaching on the spot. Yeah, we're tightening up that feedback loop instead of having to be hours or days to seconds. Huge. Has this data transparency either directly or culturally sort of changed attitudes toward say doping? I mean, I know that there's been a cycle there, no pun intended, but where's that fit in there? Absolutely, the thing is, so we're starting to see some basically sharing of an individual's riders data. We share as our team, everyone, every team member is accountable for what they need to do. We know the power that they should be delivering on the front to hold the team at a given speed based on their aerodynamic drag and all the ambient conditions, so the weather plays a factor in that, the barometric pressure on the day, the temperature and humidity affects their density, so we know whether somebody's either doing or not. And we also would know if somebody all of a sudden started producing 10% more power. That does happen. With high-end athletics, we get to diminishing returns. We are already very good. We seek a half a percent gain here and there and the accumulative effect maybe one or 2% over time, but nobody changes their performance by 5% in a week or a month or even six months at this level. So it is in some ways an actual deterrent that it's not being published as such, but I do see the future that that's being tied to some of the parameters that they do in the blood passport. Our athletes are part of the biological passport where they're having tests over time and tying this type of data to that, again, will flag outliers, question marks that need to be evaluated and so we have an ethos at this point that we do everything the right way. We use technology, not the other things that people have used in the past. Yeah, you get the edge for the data. Get a competitive advantage. Exactly, we can win with that. Neil, thanks so much. What an exciting project. Congratulations for getting that four hours of time just on that one piece. I'm sure this hour is saved on the other side too. Absolutely. That's awesome. Remember, go to Cube Madness here. Starts on March 15th for the folks watching. Every year we do Cube Madness. Go to SiliconANGLE.tv and check it out. This is theCUBE live in Las Vegas for IBM Interconnect. We'll be right back with more high performance content here on theCUBE. We'll be right back.