Evidence Strong
How Olympic weightlifting champions are made - with Dior Anderson
I think we've moved on from the narrative of talent development being just a simple eye test or one off competition performance to say, yes, that's the new future Olympic champion. Unfortunately, we need to begin to dot more perspective, not with the technologies where we look at the progression of the athlete over an extended period of time, rather than, you know, very small, specific windows. So welcome to our evidence strong show. I'm thrilled to have you. And if you could briefly introduce yourself. Great to be here, Alex. Thanks for inviting us on. So quite right. I'm Diora Anderson. So I completed my PhD back in 2020 at Bang University, in which I conduct to the series of investigations to understand the development of performance or end performance development in weightlifting athletes using machine learning, which I guess helps to consider a range of factors in profiling high performance together in one research project. And so very excited to, to delve into those findings and any further information you might want to know. Awesome. So I had Vicki come on and discuss the first part of the big project you mentioned, looking at retrospective data. So asking athletes kind of to go back and think through some things that could have been related to their success, or lack thereof, in terms of weightlifting performance. And today we'll be talking about the study titled capturing the holistic profile of high performance Olympic weightlifting development. And this study was prospective. So could you tell us what it means and why you decided to run this study the way you did? Yes, of course. So most research to date, certainly up until the date of our publication, typically adopted a retrospective inquiry approach. And so by using this perspective approach, we can first of all, of course, progress the literature, because retrospective inquiries tend to have limitations, such as recall bias athletes who tend to recall memories earlier early in their formative years, perhaps tend to be biased by events that happened later on. And so by deploying a prospective research design, this enabled us to monitor those factors that we or the literature determines as being important for our high performance, longitudinally. And so we can, we can get to real snapshot in terms of how these pro these factors develop and contribute to the success of some of these in this case Olympic weightlifting athletes. And so it's again that a step on this was a, this was the third chapter from memory in my PhD thesis. So it was a step, it was a follow up study to do the study that you would have discussed with Vicki. Excellent. So who, who were the participants and how did you go about recruiting them and what, what did you ask them to do? Yeah, so we completed the project as a whole in collaboration with the national governing body for weightlifting in Wales, weightlifting Wales. And so we worked very closely with their performance directors and pathway managers to essentially gain access to too many of the youth and junior lifters that were currently on their pathway at the time. And we used predominantly, as mentioned, youth and junior lifters, which we're in is from an age range between 13 at T1, a baseline up to 18, 19, and we monitor those athletes for a period of around two years on the study. So, so did you, did you have any criteria for who in terms of lifters would qualify for the study, wouldn't or you had everyone join, whoever was registered with weightlifting Wales or how did you go about selecting. Yeah, so we worked very closely with the six performance academies across Wales. And so it was, it was pretty much an election produce service whereby the academy coach would select a subgroup of athletes and lifters that were primarily easy, easy to access, constantly training in the environment. And so we could gain access to them pretty much on a weekly basis to monitor their progression. And so it was, it was, it was more of a selection based on availability, rather than a specific performance metrics. So, okay, how many athletes did you end up having. So we actually monitored across the two year period, we were able to monitor 2069 athletes. We are, we, at baseline, we did capture a few more than that. I think we captured around 50, but to follow, of course, the attrition rate meant that we had to essentially cut the sample size down to 29 over across the two year period. And could we ask why some of them dropped out? Through a combination of factors mainly to drop out of the sport, but then also have factors such as availability on follow up testing. And so, because the geographical locations of the academies were such that we had to plan ahead for testing dates to factor in things like travel, budgeting and so on, not all athletes could attend the follow up sessions, which I guess was a logistical limitation of this, this study, if you like, because profiling athletes on 650 data points roughly meant that you're either going to lose out on some data points or you're going to lose out on the observations, the athletes. So, but I think that 30 across the two year period is still quite a substantial amount, given the effort involved. Okay, so what have you done? Essentially, we developed the framework initially that tied in factors in high performance development into a call it battery of tests, but essentially it's a an informed data capture, which tied in, again, is holistic profile of athlete development, which includes subject areas such as demographics of family sporting history. So what sports did did mom or dad play? Were there any siblings? Were they also involved in sport? The number of places you've lived home place throughout development, so the town sized birthplace effects relative age in calendar year. And so these are all factors that, again, the literature does suggest it contributes in one way or another towards high performance development. We're also tied in with sport participation history and specific development until milestones, so where there any donor sports potentially in the athletes development. So did they come in from other sports? Did they perhaps sample gymnastics, athletics, rugby from an earlier age? And maybe that increased their exposure to Olympic weightlifting movements as well as specific pathway milestones. So when they did it into the pathway, what were the specific performance criteria or maybe they had to compete in specific compositions at different ages. And so all of these data points we captured throughout this two year period. And then lastly, we also captured specific practice activities that some of the athletes were exposed to. So these included approaches and motivation to practice. So how much proportion of practice involved in the practice versus deliberate play? How that changed throughout the ages and stages in the pathway? Could you explain what deliberate play and deliberate practice are? Yes, so early accounts of understanding expertise in the literature suggest that practice is important for development of high performance. These literature was mainly observed in performance musicians, high performing musicians, and Erickson back in '93 essentially found that he could find a linear relationship between the amount of practice that was completed based on performance level. So those that were in the highest level for their field actually accumulated much more deliberate practice than those that were perhaps at lower intermediate levels. And so the sports, the development of sports performance literature has also extended this to suggest that actually it's not just about deliberate practice, which again is it was characterized initially as being very highly focused on the development of one's own performance. These are through these accumulation of hours of practice, you tended to be very focused, very focused on areas in which you can improve your performance. And this wasn't necessarily an enjoyable element in the development of one's own performance. So since then we've now come to understand that there are elements of practice that isn't necessarily being this highly focused, not inherently enjoyable element of performance or development of one's own performance. And we turn this deliberate play in which you're exposed to practice conditions, but there's an enjoyment aspect you're focused on, you're not necessarily focused on development in your own performance, but it's more so participation to enjoy. And development comes as a secondary lead to that, which also is shown to then support, once one does transition into higher degrees of deliberate practice. The other answer to those difficult moments in deliberate practice, you've accumulated a higher degree of play, you can transition easier through the pathway, when it does become time to become dedicated and focused on development of performance. So the difference between play and practice would be that the main goal of play is having fun and learning the skill of where it would be the secondary goal, and in practice the play element is a little bit sacrificed to enhance the learning of the skill. Exactly that. We went through the areas where you were looking for variables that we cover all of them. So there were a few more in terms of microstructural practice, some of which were based on the athletes approach themselves, including deliberate practice, deliberate play, some non physical practice of watching through learning vicarious experiences, we call it, we call that. And then there was also some aspects of coaching involved in that microstructural practice, so how much of your coaching, how much feedback you receive vocally versus via demonstration or via video. Also factors that considered the transition in terms of the proportion of feedback that you were exposed to at different time points in the study. And in fact, some of the factors that came out around feedback from coach came up as predictive of success. And then there were also factors around constraints versus prescriptive coaching. And so how much of your practice throughout the this observation period is constraints based versus prescriptive your coach telling you how to arrive at a solution versus the coach that could maybe set an environment up to allow you to arrive at your own solution regarding discovery. So we consider the whole range of factors in this particular study that potentially could contribute to the success of these lifters. Okay, and I have one more question about feedback, how did you measure how feedback was given and the same question would go for the coaching methods. So we predominantly used survey stroke semi-structured interviews, so for any athlete that was not able to attend face-to-face interviews, we also conducted surveys. But in essence, we asked them to estimate the proportion of time based, you know, how we first of all gave them a definition of each of the feedback at a stroke on coaching aspects and then asked them to sort of give a graph estimate of how much of their overall practice is represented in terms of proportions. And so for constraints based versus prescriptive based coaching, for example, we asked them to list baseline amount of practice that would be constraint based versus prescriptive based, something like 80/20 would be a typical response or 60/40, similarly for the amount of video versus instruction versus demo they received, typically would be maybe a 33/30 or 33/30 trees, or depending on the athlete's experience and development, sometimes it could be more video, less video and so on. So it was mainly in proportions in terms of percentages that we asked and measured throughout this observation period. And how often did you do that in interviews? So we did a baseline, we captured the composition of practice, a baseline, and then we also did it at T1, which was after 10 months, and then at T2, which was then 10 months on from that. So three data point collections at zero, so when you started the study, 10 months and then 22 months at the end of the study, do you have anything about the procedures you would like to add? There's lots of pack-in, and so there naturally is collecting so much data points, Alex, means that I'm going to miss some aspects, but I think we covered a lot of ground and a lot of bases, theoretically, in profiling these athletes, we covered the psychosocial aspect. We also, and that's to mention that, we also collected psychosocial metrics, which included approach to mastery and performance, harmonious passion and obsessive passion. These can all be found in the manuscript and study paper for yourself, and so we covered a whole range of aspects across these five developmental themes. So did you travel to the athletes, or did they have to come to the year? So we mainly had to travel, it was involved a lot of travel, both in the north and south of Wales. If you know Wales at all, you know that the north and south is operated by quite beautiful mountains, so within each testing session it involved travel between those mountains of, I think I know them quite well as fair to say. And so we would set the testing sites up at the academies, and essentially run the testing days spread out through a testing window of, say, one month each time. In essence. Yeah. But have you thought it was your physical testing? Yes, we also included a battery of physiology and metrics, which involve collecting information and bonding the composition, bodily segment ratios, skeletal muscle strength, which included the grip strength and asymmetry between grip strengths. Back squat to bodyweight ratio, as well as front squat to bodyweight ratio. And then we also calculated some additional diagnostic parameters, which included from the counter movement jump stretch shortening cycle neutralization, as well as mobility and trunk stability in the overhead squat. So we considered a body composition and then metrics right away through to dynamic strength, as well as mobility, we came up with the results. So 650 down to how many, what are they? And so what we ended up with is was a model of approximately nine characteristics. So nine features out of the 685 data points that best separated high performing from local forming weighting athletes in the sample. So those nine features spanned four of our developmental themes, four out of five developmental themes, namely demographics, psychosocial characteristics, sport participation histories, as well as specific practice features around specific practice activities. And so within demographics and family, the school being the main place for sport participation came out of separating performance groups very well, as well as psychosocial characteristic, namely perfectionism, specifically doubts about actions. So if you had more doubts about actions, you tended to be in the lower performing cohort, sport participation history and weightlifting specific involvement tended to center around the exposure to flexibility and mobility training at both age 11 and age 14. So if you're exposed to more flexibility or mobility training at each of those ages, you were high performing, you tended to be high performing later on in the study. Sorry, with the flexibility and mobility, did you have a cut off point in terms of how much flexibility training they had to do? Yes, we look to that in the odd ratio calculations, and so what the machine learning model does is it just determines the kind of the average score and look to positive or negative weighting of that icicular attribute. So it was even, in this case, at this stage, we're suggesting more or less is important for high or low performance. We'll get into the specifics, I guess, of those in the odd ratio in the odd ratio calculations. And so, and so there was in that machine learning model, there was the amount of information received by our demonstration, proportion of extrinsic feedback, as well as the volume of flexibility and mobility practice, as well as some specific exposure to snatch and clean and jerk, such whole practice versus being higher in the high performance group, and then the local performance group, which, again, speaks to the liberal practice. More exposure to practice on specific lifts tends to be high performing the characteristic of high performing athletes, as well as interestingly, the last point being the change in the proportion of information received as feed video feedback between T1 and T2, so that was our only characteristic in the model that actually had a longitudinal pattern element to it, suggesting that as athletes transitioned between time points, if they were exposed to more video feedback, suggesting perhaps they were more comfortable video feedback, maybe they were able to pick apart different things in video feedbacks that you probably wouldn't have done at baseline. They tended to be high performing, so that, again, was a longitudinal data point, which suggested some specific implications around when to expose lifters to two different forms of feedback. So, as a whole, there's lots of one pack in that one model that captures a whole range of aspects in the holistic profile. So, were some of them more important than the others? In essence, yes, when you look within the model itself, we reported the importance level, and so we determined importance level by the amount of algorithms that suggested a given factor was important. Now, what I mean by that is we use four different machine learning algorithms to essentially reduce the data set from 685 to 9. And we also were interested in how many of those four machine learning algorithms each factor appeared in, and if it's almost synonymous to asking for different experts which different factors are most important, that's how we thought of it. And so, if all four experts agreed that a specific feature separated the groups, then we determined that feature is being very important. And so, our criteria was based on whether for a very important, fairly important, and important, it was based on whether the factors appeared in four, three, or two of the machine learning algorithms, the feature selection algorithms. And so, the very important factors that came out were exposure to sport participation at school, flexibility, mobility training at 14, being the most important. The fairly important factors being perfectionism and doubt actions, as well as the volume of snuggets practiced by T1, as well as the changes in proportion of video feedback between T1 and T2. And then lastly, the important factors that appeared in two of the machine learning algorithms were flexibility, mobility training at age 11, proportion of extrinsic practice by T1, as well as the volume of flexibility, mobility training by T1. So, there's again a range of different features that all of which I guess in the whole overarching picture are important, but we can look at, you know, the relative importance within those as well. Awesome, so we had machine learning, you define nine aspects that were important. Now, how did you go about having the alterations? So, what's ratios will be separating? If this, how likely the athlete is to be performing well in the future? Absolutely, yeah. So, the thinking behind this was the machine learning approach tended to be quite reductionist in the sentence that we're reducing a data set of 600 data points, all of which have their own theoretical relevance. Down to maybe eight online, which could potentially lead to something called overfitting the data. So, over describing just this sub-sample and not, which isn't necessarily best for generalizability beyond the sample of interest. And so, what odds ratios enable you to do is they enable you to tie in some other factors in this framework that you developed that suggest maybe some consideration, some more consideration outside of the machine learning model. And so, what we found with odds ratios essentially was a few more additional characteristics from each of the developmental themes, as well as that many of which already appeared in the model that I had previously discussed. But others had some, enabled us to consider some other aspects within each of those developmental themes, which, again, helps to bolster out the narrative. And so, within specifically within demographic, some family sport participation histories, we, as well as schooling, as I've mentioned, we also found within the odds ratio that the population of the longest residing home place between the age of six and 12 years. So, having a home place of population that has a population above 11,000, roughly, within wells, that would be approximately a small to medium town size. So, anything above that threshold tended to suggest that was more beneficial for higher performing individuals in the manuscript. I discuss potential implications around infrastructure, town infrastructure, and how that can influence affordances for small participation. So, if you were exposed to more affordances for small participation, that would tend to suggest that you participated in more sports between the ages six and 12. And as we suggested, the implications around play to later support transition through the pathway is sort of implicit within that. We go home a town better. Yes, yes, and so, typically across wells, town sizes of medium and above tended to have more public facilities such as CrossFit, CrossFit gyms, and weightlifting clubs that again enabled many of these athletes to be exposed to weightlifting to begin with, if not weightlifting, some other sport that had a positive transfer effect. And so, home place definitely came out as an important home place early in development came out more important. Then we move on to some of the physiological parameters, and there were a few that came out as important as you would expect, weightlifting is, by nature, quite reliant on physical attributes. And so, the most predictive of these features that came out were the front squat to body must ratio, being above the norm for age and body weight, as well as the back squat to body must ratio. So, these are features that, again, are implicit within the training programs of many of the athletes. And so, you could monitor really the progress in those two data points alongside their training, so there was no need for any additional testing. With the ratios, you said above the norm, what was the norm? So, we used linear models to essentially determine, normalize the each athlete's scores based on an expected value. And so, because we had a wide ranging sample based on age and body weight, we had to normalize their performances somehow. And so, we used linear regression models, which plotted age and body weight against the front or back squat to body must ratio in this case, and determined whether or not they were above the regression line or below the regression line. So, if you were above the regression line, you tended to be above the expected value, and that was the criteria in this case for that specific metric. In addition, I think another factor that's worth a discussion in the metrics and physiology was the mobility and trunks stability score, in which we, so the protocol for that was we essentially had athletes perform overhead squat and perpendicular to that. We essentially recorded their limb angles that included the torso and included the shank and included the thigh, and the most predictive angle in this case was the torso. And so, we found that having a torso above the 65 degree mark meant you were much more of an upright stable position in terms of the snatch. And so, that came out as a factor as being highly predictive of performance two years later. Yeah, it's when you consider the holistic picture it, you know, if you tie in the fact that as well as the torso angle, they needed to have a back squat and front squat above the norm, as well as that they needed to have been exposed to more sports throughout early in their development, as well as that they needed to have exposure to sport at school. You know, we're starting to build a really nice narrative here in terms of how to develop high performance and understand the developmental high performance. Will you be talking about deliberate practice and play? Yes, yes, I can do. So, as we progress through and mentally of these features are on the manuscript, so you can implement the attributes that come out as important or separated by themes. So, if you can explore them individually within your area of interest, if you like. So, if you're more interested in some of the physiological aspects, the manuscript will, of course, cover a whole range of those physiological aspects, but let's move on to practice. So within the yod ratio analysis, what we found was by looking at the exposure to different practice types across different ages, we can gauge an idea of when a significant jump in practice volumes tended to occur. And so, by looking specifically in table seven in the manuscript, what we found was that by exploring specific ages, and we could see that from age 10, from age 11, age 12, being exposed to roughly one hour a week of flexibility and mobility training, tended to be more beneficial for high performance, but interestingly, from age 14, that number tended to jump from one hour to one hour and a half, two hours. And from age 14 onwards, there has to be some form of increase in investment in practice volumes, specifically in flexibility and mobility training. In age 14, a similar story can be the set of a strength conditioning training at age 15, and the recommended parameter jumped from 15 minutes to two hours and 15 minutes at age 15. And so then that suggested that from age 15 onwards, there has to be that significant increase in jump from in terms of practice volume, similar story with the weightlifting, the technical elements of practice as well. So, from age 15 onwards, nine and a half hours a week of weightlifting specific technical practice tended to occur in high performing individuals, but how that may be shifted from, so that gives you an idea of practice volumes and the volume of practice that you would have been exposed to, but how perhaps is that represented in volumes of deliberate practice and deliberate play, we also explore it in the study. So what we found was that it's not actually in the manuscript, but it is in my, let's just bear with me, it's in my PhD manuscript. Yes. Let's just bring that up. Some beg and don't break out. Yes, absolutely, absolutely. Okay, so let's just bear with me a second. No, probably. So what we found was that in the T1, in terms of deliberate play and deliberate practice, we found that there was a 40% emphasis on deliberate play versus the practice. So 40% of practice tended to be playful activities, but at IT2, that shifted to 5%, 95%, so 95% of practice was deliberate practice and less emphasis on play. And so that what that suggested was there is a shift away from playful activities early in development to more of an element of focus on the development of one's own performance. That's probably suggested to also coincide with exposure to specific competitions. And so we don't, I guess the athletes themselves aren't necessarily told to move away from playful activities, but by being exposed to specific competitions, that almost encourages them to then develop their performance for the next upcoming competition. And so, you know, whilst deliberate play, a transition from away from deliberate plays important, how we go about navigating that probably happens through exposure to specific competitions. That's interesting. So the high level athletes would transition from 40% play, deliberate play to 5% deliberate play to increase the amount of practice. Yeah, exactly that. The emphasis on the development of their own performance. So how, how you think in terms of ages, are the specific ages or years of experience, when would it be the time to. So I, I'm from the data. And as I mentioned, I know that I'm asking you to kind of. Yeah, yeah, of course. So again, what we're doing here is, is we're, is we're creating from this, from these data points, we're creating the narrative that again, many, but in many instances, you have to connect the dots, you have to have the theoretical background in order to be able to connect the dots. So what was, what I'm suggesting from this is that when you consider that increase in practice volume happened between 14 and 15 years, and that was in flexibility, mobility training, strength, conditioning, as well as technical practice. I would suggest, I would, it would imply that perhaps the emphasis in away from play into practice would also typically happen at that age and stage. So 14 to 15, that's also, that also marks the jump from, from youth, I think into junior. So you, you're now competing at a different age group of 15 years old as well. And so considering those two factors, I think 14 and 50 intended to be the shift is difficult to say based on specifically based on the evidence, because we have such a wide age range. And so monitoring, we didn't necessarily control for age at baseline. Yeah, then you would have to have, again, bigger sample, which is very costly. Yeah, yeah, of course. Yeah, exactly that. But, but, and that, I mean, the, the literature also does suggest that investment into sport tends to happen between 13 and 15 years of age to the developmental model of sports participation. And why co-teying colleagues suggested that actually a transition into specialization tended to happen at 13 anyway. So it kind of coincides with, with existing models of talent development. So saying specialisation, do you mean specifically weightlifting? Or do you mean just athlete, a person dedicating themselves to sports? So they decide I'll be, I will be an athlete, I will invest my time into that. And as opposed to investing into arts or academics or. Exactly that, exactly both. Okay. You tend to invest, you tend to specialise by saying you invest in into sports that tends to be into a single sport. Right. So you transition away from a sport, solvantly, which happens between six and 12 years, into specific specifics or by 13 and 15. And that, that tends to coincide with your identity as a person. You tend to tell others at that age that you are a weightlifting athlete. Whereas before perhaps you didn't quite have that sense of identity. That's also fostered as mentioned with by family values in sport. And so by having often having a parent involved in sport or a sibling involved in sport that tends to foster that identity as you transition into those specialization years. Excellent. Other aspect. And you say, oh, yeah, you mentioned volume of snatch as a whole, how, how it will, how it came up with what's ratios. Yeah, that's so, so interestingly, the, the volume of practice in the snatch as a whole. So whole purpose is part practice in the literature just is the degree to which you, or it character arises, the degree to which you make a complex skill simpler. And so in many instances, such as a golf swing, you would need to break down the skill into parts before combining those parts to a whole movement. What the data suggests is that in the snatch, you actually need to, at some stage, buy, I think it was, let's, the other ratio suggested 15 by 15, those movements needed to, in well organized. And so that you increase the amount of practice warning that you complete in the snatch by 15. So again, that needs to be quite a stronger emphasis of practice in the snatch as a whole movement. I'm going to start from the first phase while the way to the, to the catch, mainly because the complexity of the movement is such that a degree of high organization needs to happen at the moment level in order for you to be able to add load and increase performance. And that wasn't, that wasn't represented in the, in the clean and jerk in the same way. So the clean and jerk, in essence, there was more emphasis on parts, part practice, specifically, because the clean and jerk is, in essence, separate lifts in itself. And so there needed to be some emphasis on the clean, as well as some emphasis on the jerk separately. And so we didn't, we had a difference in the relationship between all and part practice in the snatch versus the clean and jerk. So did you have like proportion of training? So let's see. Again, we didn't report without any manuscript. So let's see, we have it in, there were so many, there's so many, there's so many points in findings that we had to cut down quite a lot in order to publish. So let's see at T1, the proportion of part practice for the clean and jerk in all 23 local forming athletes was at least 50% of practice at T2 practice in the clean and jerk 10 to be at least 70%. So part practice actually increased in the clean and jerk throughout this particular study. So 70% of the lower performing cohort, 70% of practice tended to occur more or so in the lower performing cohort, suggesting that there is an emphasis on access of the clean and jerk being a whole movement in higher performing athletes. So actually the lower performing cohort tended to break down the clean and jerk into sub sub movements or parts, whereas this wasn't the case in the higher performing cover wall. So the higher performing cohort tended to practice is the clean and jerk as a whole movement and more so than the lower performing. So that was in the clean and jerk in the snatch. Okay, so we pick up part of the most important findings so we don't if it didn't come out as a predictive factor in terms of percentages. We don't actually report the specific percentages because they hadn't come out as predictive in the manuscript. So there was in other words, there was no differences in the percentages in predicting high performance. What we had found though in the snatch was that there was a higher accumulation of volume for the snatch in higher performing individuals of whole practice. So a higher accumulation of whole practice in the snatch for higher performing individuals by T2. So specifically by T2, higher performing athletes had accumulated at least 313 hours snatch part practice and at least 327 hours of snatch whole practice. And so roughly, I guess 50/50 proportion split, maybe a little bit more emphasis in the whole practice with the higher performing individuals. Whereas this wasn't the case in there was only one in 23 of the lower performing athletes that accumulated this volume practice. So it's suggesting that overall the volumes in both needed to be higher but there was no emphasis of the snatch in terms of proportions. Hence why we hadn't reported. Okay, so following up on this question, do you think that the difference, part of the difference was how much the high performance group was practicing in general, more flexibility training, more strength and conditioning, more skill training, more whole movement practice? Yeah, yes, in short, the data suggests that higher volumes of practice tended to be best predictive of performance, more so than the proportion splits. When it comes to proportion splits, the most predictive proportion split that came out was the proportion of feedback, as I mentioned before. So there was more of an emphasis on video feedback as the athletes transitioned rather than the volume of feedback, if you like. Yeah, I would be interested to talk about it, too. I have a question, I will ask it now, but you can decide to answer it later. So the question would be, you have monitored athletes for two years and then 12 months to see how they performed in weightlifting. Do you think it's possible that if you would follow up the athletes 5 or 10 years, the split between high and low performance could be different? So, in other words, could low performing athletes become high performing athletes? Yeah, the other way. Yes, no, no, absolutely interesting. Perfect question, Alex. In the sense that there are peer limitations to the study design in terms of the time of the observation window that we are allowed for. Generally speaking, when analyzing historical data, the past 20 years of historical competition records, transition from a lower performing to a high performing tended to not to happen after a given age window. So after junior, so from the age of 17 onwards, that tended not to occur. If you were in pathway from that early age, you can, of course, join the pathway later on and transition from high performing based on experience level. So if you came in as a 17-year-old with very little experience, there was a tendency to move from lower performing to higher performing. If you were in the pathway from early and show that actually your performance was suddenly from that early age, then transition to later on become high performing was harder or tended not to happen. So I guess at the competitive experience level or the amount of years involved, which we didn't necessarily control for when we studied. And so maybe we could have observed that some of these, some of the lower performing sample, if they had a relatively lower amount of competitive experience or time involved in weightlifting, then the chances of them then transitioning a few more years down the line would have been higher. But again, unlikely to happen, but it's possible, depending on the volumes of practice accumulated. So try to base your point. Talent development, I think, should not be. I think we've moved on from the narrative of talent development being just a simple eye test or, you know, a one-off competition performance to say, yes, that's the new future Olympic champion. Unfortunately, we need to begin to adopt more prospective longitudinal methodologies where we look at the progression of the athlete over an extended period of time, rather than, you know, very small, specific windows. Could you speak a little bit about the feedback and how it influenced or was related to performance? Yes. So as we mentioned, the proportion of video feedback that was received from a type of feedback rather than was received from coaches to the athletes. To the athletes tended to change throughout the study. And so what we found was that a high performing individuals actually began to shift into more forms of video feedback as the study progressed. I guess there's a number of implications that could be drawn from that. Perhaps athletes became more experienced as they transitioned through the study. And so they were more comfortable video feedback and extracting specific elements of performance from video than they were able to do right at the beginning of the study. And so there's a natural transition into video, or it could be coaching method. It could be coaching mechanism, maybe some coaches, which we didn't explore and study. Some coaches may have not enabled the athletes to be exposed to more video. That's something that requires further discussion and investigation, but nonetheless, that transition into receiving more video information tended to be predictive of high performance. Also, many of the athletes had coaches that were over longer distances, and so if they changed coach or they've relocated, they tended to be exposed to more video forms of feedback anyway. So again, it requires further investigation, but it does suggest that there should be an emphasis on video feedback the more experienced the athletes become. It could be also related to the transition from play, where it doesn't really matter how it looks to the practice, where it really matters how it looks. So yes, absolutely. Yeah, absolutely. I think there was also an emphasis on intrinsic or self-generated feedback versus externally generated feedback. I think we also found that at T1, 5 out of 6 of the high performance athletes reported that there was at least 20% of their feedback coming from their own self-generated whilst this was only happening in two of the 23 lower performing athletes. So 5 out of the 6 higher performing athletes were able to generate their own performance of feedback at T1, and the remaining 80% tended to come from their coach. And this was not found in the lower performing athletes, so perhaps they were more dependent on coach generated or externally external generated forms of feedback. Yeah, and a T2 that shifted to 25%. So it's a very old shift in the proportions there. But again, it does suggest that an element of being back all but to come from the athletes themselves. However, that doesn't discard the fact that especially in weightlifting, having the external eye, having the external form of feedback via a coach or video should be the large majority of talent development, training or feedback in developing athletes. We didn't talk too much about it, Adam. So Vriki, is for you to explain mastery approach? Achievement motivation, in essence, describes your motivation for achievement of competence in two different aspects. Performance is versus performance-based approach, which essentially compares yourself to a normative standard or others. And so those two, those dual elements of motivation can often characterize how we engage in a task. And so the approach and avoidance mechanism comes in when we are either characterized as approaching competence or outperforming yourself. Generally, do you either approach competence or you're motivated by outperforming yourself or outperforming others, or you engage in a task to avoid being incompetent? Very similar, but also very different. And so in a group dynamic setting, you can either approach being the best in the group. If that's a performance-based approach or being the best version of yourself, or you can avoid being the worst in the group or being the worst version of yourself. And so you can imagine that being a very different approach to training, and that can actually influence how you then approach and engage weighting training. And so what we found was that the majority of the higher performing athletes tended to have this mastery approach, so they were motivated by achieving competence that was focused on the development of their own performance. Whereas the reverse tended to be shown in the lower performing sample, whereby they were avoiding being worse than themselves. And so there was an emphasis on not doing it as badly as I did it the last time, whereas in the high performing, they were focused on outperforming what they did last time. So that's a slightly different mechanism there, but it also characterizes perhaps the attitudes towards training in some of the high performing individuals versus the lower performing. And this was also shown in the performance of the approach. So that was mastering the approach and avoidance, but in the performance approach, higher performing athletes were also motivated by being better than others. And this wasn't as prevalent in the lower performing sample. So yeah, that's the achievement motivation. More emphasis on mastery approach outperforming yourself and outperforming others versus avoiding being the worst. I think it's also worth discussing harmonious passion as well, just in the psychosocial profile, as that was the highest odds ratio goal. And in essence, harmonious passion describes an engagement in a passion of yours that is synonymous to your environment. So what that means in simple terms is you engage in something because you love it, you enjoy it, it's your passion. But also that passion is that doesn't result in any form of conflict in your social network or your environment, or it's also synonymous with values that surround you. Obsessive passion is actually the reverse of that. Obsessive passion is a passion that you enjoy doing, you want to do, but it actually can result in some form of conflict in your environment. And so I guess a few examples of that would be enjoying and listening to very loud music. That would be a very strong passion of yours, but it's not going to be conducive to your environment. You often may have problems at home, family and so on. The same occurs with weightlifting. So higher performing athletes tended to have this passion for weightlifting, but it was supported within their immediate environment. It was harmonious with their immediate environment. So through social media, through their clump, through their club environment, this passion tended to be fostered. That led to very little conflict, which I also think is quite my key. That's huge. I expected that it also influences family support, possibly with the under. Absolutely. Absolutely. There were many anecdotes whilst observing some of the athletes whereby parental support in terms of access to specific competitions. Some athletes may have had exams coming up and so they became an immediate conflict between training and exams. You can imagine that being quite a prevailing thing in your youth athletes. And so how is that handled? How is that supported by the immediate environment and by the support network of the athletes? Not saying that they should choose weightlifting over exams. I suppose the support mechanism would be slightly different for less harmonious passion individuals rather than the reverse. We have to talk about competition. So do you mean the preparation for competition? So these two features were features that also came out as predictive in the Great British Medalist Project. So two relative importance of sport was found to be a particular factor in super elite British medalists. And so we aimed a discriminative factor in super elite British medalists. So we aim to repeat that measure on paradigm in this study. And we found that yes, high performing athletes had a degree of relative importance of sport. So that was the perspective of sport relative to other life interests and responsibilities such as career studies and life choices socializing and so on. And so what we found was that the high performing athletes had a higher degree of relative importance of sport in their life. That coincides with obviously the investment and emphasis on the practices we discussed. But also from a social aspect, they placed a higher degree of weightlifting in their overall identity and how we specifically measured that was we got them to place two circles side by side. And we had just asked them to overlap the circles and higher performing athletes tended to the two circles were weightlifting and myself as a construct. And the high performing athletes actually overlap these circles much more so than the lower performing athletes had actually had a separate identity of weightlifting and themselves more so and the lower performing so that that would be a suggestion that sport tended to be higher than it had to be held in higher degree of importance in the higher performing group. And then total preparation for competition. And this also coincide with some aspects in protectionism in the high performing athletes as you can see in table six organization tended to come out as a very predictive feature in high performing athletes, as well as this aspect total preparation for competition, which essentially when you consider those two factors together, it suggests that the high performing athletes often left no stone unturned in preparation for their competition so that they felt you would feel typically that there was not much more that can be done in terms of the preparation for competition and this was more reflected in the higher performing group and the lower performing group often felt as though they weren't prepared for competition as they would have hoped. If you would have to guess how much is it surely as you describe the mental side is very important, the identity that aligning your body, your life to be able to perform and how much is that and how much is the physical attributes there. The front squat to body weight ratio and torso and all and so on. Wow, what are the proportions you think. Yes. So an interesting follow up analysis of this of this particular data set would be determining the weightings of those each of those factors. The fact that in this case, the back squat to body ratio didn't appear as prevalent in the model, the machine learning model could suggest that it's not as relatively important as these other factors that came out. There could be the possibility that this is an overfit of the particular date of the particular sample. So why would speculate that the more data that you would feed this study, the more there wouldn't be an emphasis on some of the physical attributes in weightlifting, more so than in comparison to other sports. So whatever proportion, the physical aspects in weightlifting are, and I think I've done, I think on my, on my, I wrote an article, I published some analysis on LinkedIn that showed that it was approximately 50% and 70% respectively of performance in weightlifting could be accounted for by body weight and age and competitive experience. So that remains that leaves 45% and 30% to be accounted by four by other factors. And of course, from this study that suggests that those additional factors are somewhat what makes the additional difference above the 45 and above the 50% and 70% in women. So yes, there is going to be a lot of emphasis on physiological development in weight instant athletes, but that physiological development has to happen as a result of some important antecedents. There has to be family values in sport, there has to be the emphasis on oral and total preparation for competition. There has to be an emphasis on that passion for the weight you think has to be harmonious and, you know, transition from play to practice and all of those, all of these elements definitely do contribute. But the relative weightings, I'd really love to follow up in that analyzing this data set again, just to determine the relative percentages of each. Yeah, that will be also, do you think you will, you will revisit the study, do the follow up? Yeah, I think so. I think so. If I can find the time to, I think it's worth looking at again, for sure, because also with the odds ratios at the statistical level, we're considering each of these attributes in isolation. So the odds rate shows themselves are considered independently of all odds ratio attribute, the logistically logical conditions. So I would really like to combine some of these. So I would really like to look at how the odds ratio would change if we considered the athletes in the data that had a front squat and a back squat. And they had maybe school, main play for sport participation. What would the odds ratio be as a result of those three factors combined and so on. And we can continue to build from there. That's probably a really good follow up, looking at combinations of odds ratios. Exciting time. Yeah, we're in a really interesting time. This particular study does enable us to explore opportunities for athlete development that otherwise wouldn't have been there. And so we could take an athlete at 12 years old and we could prescribe what their practice would typically look like, what their motivational approach is to practice would typically look like, how those proportions will change throughout as they transition through the pathway. And so we can build a really nice holistic pathway for the, for the athletes as they develop to enable them to explore as many opportunities as possible that they otherwise wouldn't have considered. Is that how the study is used now by weightlifting whales? Yes, so weightlifting have taken many of the parameters and recommendations from this. We produced a, from this study, we produced an executive summary report, which actually showcases the results in an infographic. In many cases, it reduces the findings to its most important aspects. We present this to the summary report. And now they have introduced recommendations for practice volumes at specific ages, exposures to different modalities of training as well. So flexibility as well as S and C. And so they're now beginning to incorporate these strategies, these recommendations into their strategies for talent development. Excellent. It's really exciting. Yeah, that's, that's the best. When the research is supplied and changing how the things are done, that's the best. Yes, absolutely. Anyway, so we can't predict the talent, but if you would like to non-idore athletes or try to look for talent within all of the athletes you are having, what aspects would you be measuring and how often. And the absolute name, of course, it is unrealistic to expect cultures to to monitor 648 variables. But are there any, apart from volumes, we mentioned already, and these time points where the certain modalities should be introduced. Anything the cultures could or should be measuring. I, from the data, the data suggests that absolutely we definitely don't need to begin profiling athletes on these whole ranges of data points. But the data does suggest that we ought to consider wider demographics and family influences in television development. And so if you have two athletes for instance that have very similar physical parameters and are exposed to very similar training modalities, I think you then ought to consider some of the external factors that could contribute to their engagement and attitudes towards their training. An athlete who has a father involved in father or mother involved in weightlifting will have a very different dynamic in terms of home values than the athlete that doesn't. And so how can you support the two and that shouldn't be a means for the talent selection. It shouldn't be a means for talent identification, but it's how can you support the one that doesn't, so that they're exposed to them. You know, there are ways that you could do this, you could, you know, mentorships. Of course, coaching comes into this, but it's how can you foster those values of the sport where, in instances that they wouldn't be exposed to them in their home or family environment. And that's the same for home place as well, town infrastructure and things like that. So we just want to ensure that we level the play and feel where possible using these findings. Yeah, that's interesting important aspect, the support that is required to some extent, because if it's a clear differentiating factor between high performing and low performing, this is something that will possibly cut off some athletes from the sport. So, yes, yeah, absolutely. Yeah, taken seriously. We know the training volumes are important. We know that that should be at the forefront, but it's who's going to achieve those training volumes and how wide they're going to achieve those training volumes. Those are those areas of consideration. Yes, they're the untongibles. So we can't touch, but are obvious, at least from the study. Yeah, absolutely. Absolutely. They're there. And they're, they're, they're, they're, they're repeated in many instances in the literature as well. So it just goes to show that these are factors that were all to be considered. I hope the viewers have enjoyed right. They'll bring into the study. They can, of course, contact me through through social media, Twitter and LinkedIn. And let's continue the discussion. Awesome. What do you like to mention your handle on Twitter? So it's easy for everyone to find. Yeah, of course. So it's Dior Anderson, LinkedIn and Twitter. And of course, I am, I'm continuing this, this area of research in building out some systems and technologies that can help to streamline and capture some of these holistic profiles. And so if you're interested in, in conducting similar investigations, then of course, reach out and be interested to open discussion. You want to mention the name of the company so people can Google and, and kind of. Yes. Yes. So it's, it's talent path for ID. So that's the startup that we founded a post PhD. So that's my post PhD life in which we are building or award winning sports technology. We won the sports technology award back in 2022 for the best coaching and performance technology. And so it's quite an exciting time and very excited to continue this work in understanding and developing talent. Is it across the sports or yes, so we have a few ongoing data capture systems. So we're working with many sports organizations. So from the pro level right away through to the grassroots in which we simply have players periodically complete some profiling using a very similar, but not the same methodology in profiling their development. Of course, it's going to be very specific to the to the sport in which we aim to understand their development and then support them on their journey. So we talent pathway is understanding for enabling them to understand their development and tying in the support networks such as their coach and so on. So we have work with them. And we also have work on going with national government body for canoe in Wales. And we also are continuing some work, not directly related to this, but a separate project with weightlifting wheels. We're by supporting their member perspectives on their program, the national programs that they're undergoing using data driven insights. So, yeah, we're expanding the code base and expanding the use cases across various sports. So we're excited to keep growing. Yeah, that's that's all sounds amazing. Last question. What is your favorite color question? It's always been blue. So I'm going to stick with it. Always been blue. Every M&M's blue, this sort of royal blue. That's my color of interest. I have no idea why. Is there any connection? Is there any background? I think that looks good. Yeah, I will tell you the background. First, the comment, the blue is pretty pretty popular with researchers. I have. Yeah, it is interesting. Vicki mentioned blue. She was more into sea blue. Oh, okay. Nice. The main person always, we should run some analysis. Reflective critical thinking. Maybe. I guess. Yeah. Yeah. So I'm asking because for every study, I invite an outdoor from, I make an infographic and I use the colors the researchers like so they can have the infographic that is awesome to them too. Amazing. At least I can do it. Yes, yes, absolutely. Love it. Alex, brilliant. - Okay, thank you so much for today. (clap) [BLANK_AUDIO]