Here are some of the questions, and responses from our guest panellists from our ‘Data Science in the News’ Webinar #10 on Friday, 31 July, 2020
Q: Is there a “race effect”? E.g. are there races where everyone tends to go faster/slower?
- A from Mark Osborne: This sometimes depends on the strategy … some individuals like to sit and kick home, whilst others like to go out and try to demoralise the opposition and hope they can hang on, others prefer a more even paced race … depends on how the race tactics unfold and the attributes of the individuals.
- A from Paul Wu: Yes! Olympics tend to be 0.5% faster on average (95% credible interval between 0.3% and 0.7%) than the World Championships; both are faster than other events such as the Commonwealth Games.
Q: What are the best ways to visualise data for coaches?
- A from Mark Osborne: With lots of simple images enabling straightforward comparisons and nice colors …t hey don’t like numbers … you can’t see a number!
Q: Are the sports betting agencies interested in these dimensions of data science?
- A from Mark Osborne: Yes. Very much so.
Q: When predicting the winning time, do you take the individual athletes into account? For instance, if there is one athlete dominating the field, and you know they are absent from the next competition (e.g. injury), are you predicting the winning time based on the recent times for second places?
- A from Paul Wu: Not in this model I presented today; but this is exactly the next phase of development that we have started exploring. We are leveraging data from Swimming Australia’s results database and Bayesian statistics allows us to predict the winning time given the mix of competitors. Very happy to discuss further!
Q: Has there been any analysis of the effect of sandpapering on Australian cricket performance… is there any evidence that cheating helped?
- A from Chris Drovandi: I personally haven’t seen any analysis. However, it’s interesting to note that my understanding is that the players received only a 1 match ban from the International Cricket Council; the 12 month ban was imposed by Cricket Australia. Australia went on to lose that game badly, so it clearly didn’t help in this instance
Q: Data science and e-sports … thoughts?
- A from Mark Osborne: Every sport will have its own unique metrics that will define success/failure and also identify strengths and weaknesses, and therefore opportunities for improvement.
Q:How many players are using the sensors and are contributing to the data?
- A from Julie Vercelloni: The full team (12 players) will wear the sensors.
Q: Why did the swim time for men and women decreased over time?
- A from Mark Osborne: Athletes are training more (full-time professional athletes rather than the old days of part-time amateurs, and also training smarter and therefore getting closer to the physiological ceiling in many sports. Technology is also improving. In swimming this includes the use of materials/manufacturing techniques in swimsuit which has a significant effect on swim times.
- A from Paul Wu: Basically, humans are getting faster due to improvements in training, nutrition, medical science, and more. Early in the 1900s, it was not even common to commit to sport as a full time endeavour! However, as Mark said today, we are likely reaching the limits of human performance and so improvements are getting smaller, generally speaking. Hence, data science can help coaches and athletes find that extra fraction of a second in speed in an ever-more-competitive landscape.
Q: Such an interesting study Julie. Any plans to expand to other underwater sports? Underwater hockey would be an interesting comparison due to the shallower depth and differences in the way the game is played.
- A from Julie Vercelloni: Some UWH already use the sensors in Europe. I agree that the comparison of physiological responses to depth will be very interesting to examine. Let’s do it together!