Applied Sports Science newsletter – June 23, 2016

Leighton Baines admits he was “anxious” get back into pre-season training.

The Everton left-back has been putting in the hours at the Finch Farm gym, weeks ahead of the squad’s official return on July 7.

Baines, who missed a large chunk of last season after undergoing an ankle operation, is determined to hit the ground running this term under new boss Ronald Koeman.

… their margin for error narrowed — because of fatigue, because of injuries, because of waning focus. Kerr harped on his team to keep things simple instead of trying for the home run all the time. The players did not always listen.

“When you win, it kind of masks a lot of things,” Harrison Barnes said before the start of the playoffs.

At the same time, Kerr chose not to rest any of his stars in the final weeks of the regular season. His players wanted to win 73 games, so he made a pact with them: As long as they were honest about their health, he would let them play. Kerr did so with some trepidation. He would have preferred to find opportunities to sit Curry and others.

It can’t be easy trying to pick a team for a huge football tournament like the Euros, carrying the hopes of an entire nation. Football managers may have great skill and intuition, but it is, after all, not an exact science. But what if their talents could be supported by more precise tools informed by the latest research?

It turns out this is becoming a possibility. In a series of scientific studies, we have shown that simple neuropsychological tests of football players’ executive functions and working memory can help predict how many goals they will score, how many passes they will make and how successful they will be overall.

Football players are typically selected from an early age based on their football skills and fitness through a complex, rather nebulous system. In the English system, players typically earn their first full-time training contract at 16. By 21 the attrition rate is 75% or above.

Planning an extraordinary challenge? Prof Greg Whyte OBE is the man to speak to. Having trained David Walliams to swim the English Channel and Eddie Izzard to run 43 marathons in 50 days, Greg has become the go-to guy for anyone attempting the extreme. A former modern pentathlete, Greg competed in two Olympic Games and has won European bronze and World Championship silver. So when MR was offered the chance to speak to Greg at the Centre for Health and Human Performance (CHHP) ahead of our own extraordinary challenge – the Llyn i Llyn swimrun on Saturday 6 August – we jumped at the chance. Here’s what we learned.

Objectives

The purpose of the study was to examine the independent and interactive influences of athletes’ perceptions of autonomy support from their coaches, fathers, and mothers on the athletes’ self-determined motivation.
Design

Cross-sectional survey.
Method

High school athletes (N = 335; M age = 15.75 years; 62.4% female; 84.2% Caucasian) completed surveys assessing the constructs of interest near the end of their season.
Results

Hierarchical regression analysis results showed that autonomy support from all three social agents significantly and positively predicted self-determined motivation (R2 = 0.32), and the two- and three-way interactions significantly added to the prediction (total R2 = 0.35). Results showed that a relatively high level of self-determined motivation was associated with the perception that at least two of the three social agents provided high levels of autonomy support.
Conclusions

The provision of autonomy support from coaches, mothers, and fathers relate to athletes’ self-determined motivation both independently and interactively.

PURPOSE:

To compare the effects of three different high intensity training (HIT) models, balanced for total load but differing in training plan progression, on endurance adaptations.
METHODS:

Sixty-three cyclists (peak oxygen uptake (V?O2peak) 61.3±5.8 mLkgmin) were randomized to three training groups and instructed to follow a 12-wk training program consisting of 24 interval sessions, a high volume of low intensity training (LIT), and laboratory testing. Increasing HIT (INC) group (n=23) performed interval training as 4×16-min in wk 1-4, 4×8-min in wk 5-8 and 4×4-min in wk 9-12. Decreasing HIT (DEC) group (n=20) performed interval sessions in the opposite mesocycle order as INC, and mixed HIT (MIX) group (n=20) performed the interval prescriptions in a mixed distribution in all mesocycles. Interval sessions were prescribed as maximal session efforts and executed at mean 4.7, 9.2 and 12.7 mMolL blood lactate in 4×16, 4×8 and 4×4-min sessions, respectively (P<0.001). Pre and post intervention, cyclists were tested for mean power during 40-min all-out (Power40min), peak power output during incremental testing to exhaustion (PPO), V?O2peak and power at 4 mMolL lactate (Power4mM).
RESULTS:

All groups improved 5-10% in Power40min, PPO and V?O2peak post intervention (P0.05). Further, an individual response analysis indicated similar likelihood of either large, moderate or non-responses, respectively, in response to each training group (P>0.05).
CONCLUSIONS:

This study suggests that organizing different interval sessions in a specific periodized mesocycle order or in a mixed distribution during a 12-wk training period has little or no effect on training adaptation when the overall training load is the same.This is an open-access article distributed under the terms of the Creative Commons Attribution-Non Commercial-No Derivatives License 4.0 (CCBY-NC-ND), where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially.

With a few clicks, any sports fan can find Steph Curry’s scoring average or field goal percentage for a game. But what if someone wants to visualize every shot Curry took during the regular season, or how his three-point shooting percentage stacks up against the best ever?

Enter StatMuse. The startup, which is less than two years old and still in beta, utilizes a type of artificial intelligence called natural language processing (NLP) to field stats queries from sports fans. That means instead of relying on normal search terminology, it’s able to process searches formed in traditional sentence structure. For instance, a Google search of, “highest three-point field goal percentage in a season by a player with at least 400 attempts,” triggers a multitude of results, none of which lead concisely to the answer. But a StatMuse query worded the same way brings up over a dozen of the highest season-long three-point percentages displayed in a bar chart, led by the correct answer: the 49.2% mark put up by the Atlanta Hawks’ Kyle Korver two seasons ago (Curry’s 45.4% this past season comes in sixth-highest).

… Under Armour spent $710 million to acquire three quantifiable-self fitness apps and then developed its own wearable band, heart-rate monitor, and digital scale. It is committed to learning from the activity of more than 160 million registered users.

Nike, by contrast, has stepped down its digital efforts since discontinuing its FuelBand wearable in 2014. It claims only about one-fifth as many fitness-app users as UA.

… “If an AI were fed videos of a huge number of past NBA games, and were smart enough to understand the events occurring in the games, then it could do a better job at making tactical basketball decisions like choosing starting lineups,” said Ben Goertzel, a prominent futurist and lead researcher in the OpenCog AI lab at Hong Kong’s Polytechnic University.

“As AIs with robust video understanding become widespread, I’d expect that we could see AI sports assistants start to play a serious role,” he said.

Within five years, he expects the video coordinator position Spoelstra had “could be mostly done by AIs.” Goertzel thinks that once an AI can ‘learn from the patterns in the game, and intelligently extrapolate them into the future’, humans won’t be needed to make the types of decisions they are needed for today. In other words, AI is on its way to replacing much of what a coach does.

We believe that AI technologies are likely to be overwhelmingly useful and beneficial for humanity. But part of being a responsible steward of any new technology is thinking through potential challenges and how best to address any associated risks. So today we’re publishing a technical paper, Concrete Problems in AI Safety, a collaboration among scientists at Google, OpenAI, Stanford and Berkeley.

While possible AI safety risks have received a lot of public attention, most previous discussion has been very hypothetical and speculative. We believe it’s essential to ground concerns in real machine learning research, and to start developing practical approaches for engineering AI systems that operate safely and reliably.

We’ve outlined five problems we think will be very important as we apply AI in more general circumstances. These are all forward thinking, long-term research questions — minor issues today, but important to address for future systems

… With so much money being poured into this issue, the NFL has reached the conclusion that something must be done. So, during the 2014 offseason, it was reported that all 32 teams could be using concussion sensors by as early as the 2015 season.

“We’ve done a lot of validation work over the past 18 to 24 months using some of these devices,” Kevin Guskiewicz, a member of the NFL’s Head, Neck and Spine Committee, said back in June 2014. “It’s really important that we know what the information is telling us and how to interpret it and how we can provide meaningful data back to the player, the athletic trainer or the team physician, the strength and conditioning coach, whoever that may be.”

We study the problem of smooth imitation learning for online sequence prediction, where the goal is to train a policy that can smoothly imitate demonstrated behavior in a dynamic and continuous environment in response to online, sequential context input. Since the mapping from context to behavior is often complex, we take a learning reduction approach to reduce smooth imitation learning to a regression problem using complex function classes that are regularized to ensure smoothness. We present a learning meta-algorithm that achieves fast and stable convergence to a good policy. Our approach enjoys several attractive properties, including being fully deterministic, employing an adaptive learning rate that can provably yield larger policy improvements compared to previous approaches, and the ability to ensure stable convergence. Our empirical results demonstrate significant performance gains over previous approaches.

When we see two people meet, we can often predict what happens next: a handshake, a hug, or maybe even a kiss. Our ability to anticipate actions is thanks to intuitions born out of a lifetime of experiences.

Machines, on the other hand, have trouble making use of complex knowledge like that. Computer systems that predict actions would open up new possibilities ranging from robots that can better navigate human environments, to emergency response systems that predict falls, to Google Glass-style headsets that feed you suggestions for what to do in different situations.

This week researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have made an important new breakthrough in predictive vision, developing an algorithm that can anticipate interactions more accurately than ever before.

Convincing the U.S. women’s soccer team that it needs an edge in Rio is like telling 1992 Michael Jordan he should practice his jump shot. With three World Cup trophies and four Olympic gold medals to its name, the Women’s National Team is the most dominant in soccer history—male or female. And they’re favored to win again.

This time, however, the team will rely on advanced technology, in the form of GPS tracking devices, to provide a boost. Thanks to the miniaturization and power of sensors, it captures several metrics on every player on the field, down to speed, lateral movements, and impacts. That specificity empowers the team’s trainers to tailor workouts and recovery programs—both of which are crucial to improving performance—to each individual.

“We’ve always developed very talented players,” says center back Becky Sauerbrunn, the team’s defensive anchor. “But at the global level, other teams are catching up. So we’re trying to raise the bar, and that’s where cutting-edge tech comes in.” It’s women’s soccer 2.0.

The final stages of draft preparation for teams start when top team personnel meet in draft rooms to set the agenda for the week leading up to Thursday’s draft, starting with the most important step: setting the draft board.