Interview with Julen Castellano

After reading couple of excellent recent research papers on GPS analysis of Small Sided Games (SSG) in soccer by Julen Castellano et al. [ PAPER1, PAPER2, PAPER3 , PAPER4 ] I decided to interview him. Julen was kind enough to accept the invitation and share his viewpoints and research findings.

Mladen: I am really glad I have the chance to discuss GPS data and SSG games with you Julen. Before I starting picking up your brain can you please share with the readers who you are and what do you do?

Julen: I am Julen, Professor at the University of the Basque Country. I have Ph.D.s in Physical Activity and Sport Sciences (UPV/EHU, University of the Basque Country). I am also a Level III Football Coach.

I was a semi-professional football player for 15 years and another 15 as a fitness trainer. But I am not old because I simultaneously did the two roles in the same teams for years. In terms of coaching, I have worked in the academy of professional soccer teams.

My main research focuses on team sports, especially, in “fútbol”, football or soccer. My research areas are: performance analysis, training methods and evaluation, but they are focused on the game and the players’/teams’ tactical behaviour. I have done research on the physical and tactical aspects associated with sports performance in professional, semi-professional and youth football. I published over 30 articles in peer-reviewed journals, six books, 35 book chapters and tutored six Ph.D. students.

Mladen: What is the best way to measure training load in intermittent activities like soccer? What is the relationship between external (GPS data: acceleration and velocity) and internal (sRPE, %HRmax, TRIMP, bLA) indicators and how do they differ in different activities (for example match vs SSGs)? Which one is most valid, reliable and sensitive?

Julen: Measuring training load in intermittent activities is not easy. This has been debated for decades, yet now unlocked. They all have their adv’s and dis’s, and probably, if the team has sufficient resources, it will work with some of them; this would be better. Nowadays, with the improvement of technology like GPS having data on external load is a reality. I agree with respecting the principle of specificity, prioritizing, because it is assumed that performance improves more when training simulates the physiological demands and movement patterns of competitive matches. We must stimulate our players as specifically as we can. Velocity and displacement, but specifically acceleration, can be the main variables to measure players load. Soon (I hope) a new work titled RELATIONSHIP BETWEEN INDICATORS OF INTENSITY IN SMALL-SIDED SOCCER GAMES will be published. The conclusion of this study is that during training regimes of this kind it is necessary to consider a range of intensity indicators so as to obtain complementary information. This will enable coaches to assess more accurately the load imposed on players and to optimize the training process. The information obtained from indicators associated with high-intensity activity could be of interest, particularly when the aim is to assess specific training drills such as SSG rather than just training sessions as a whole. In SSG, it could be useful to combine both internal and external indicators so as to obtain a more accurate measure of the training load experienced by players. We have begun to see that these types of variables are more pertinent than others. But much more research is necessary to improve knowledge about it.

Mladen: Not sure if you are familiar with the work done by prof. Roberto Colli on power output during soccer [LINK], but what they basically did was to combine velocity and acceleration/deceleration data to get power output. Using velocity only analysis oversees short powerful accelerations/decelerations that did not reach high speed threshold, yet provide tremendous mechanical load characteristic for intermittent sports. What are your thoughts about using acceleration and power instead of only speed and how will this impact calculated external load?

Julen: Nowadays, the information available (to us) is enormous . One thing must be kept in mind, simplicity, we have to be simple. We must think that the trainers need little information (the most relevant) but quickly. Programs capable of handling this information would be included. But, a s a result of technological developments it is now much easier to evaluate automatically the external training load of several players at the same time. In this line, global indicators like player load (by Catapult sport , among others), calculated using the data obtained via the triaxial accelerometer incorporated within the GPS device, has demonstrated high reliability, suggesting that accelerometers are a viable tool for tracking activity changes during exercise. Now, we have other indicators derived such as player load 2D, player load slow, number of change of directions and their intensity… Absolutely, around this more research is necessary

Mladen: In most time motion analysis studies including the new ones using GPS, along with HR data all intensity zones are set up to absolute numbers (i.e. 10-14km/h, 80-90% HRmax). How would using individualized zones, for example based on max speed, MAS (vVO2max) or some other indicator affect calculated loads? Would that be more valid way to assess workloads of an individual?

Julen: Measuring individual zones we can use max speed, it could be an option, but we have to think that football is a sport of absolute values, in others words, for the game we need to know who is faster than another and not if the players ran at their 90% max speed. The paper by Buchheit M et al. (Match Running Performance and Fitness in Youth Soccer. Int J Sports Med, 2010) is a good example to understand this. I am sure both options have to keep in mind, mixing absolute and relative perspectives.

Mladen: The study done by Di Salvo et al. (2009) [ABSTRACT ] showed that high intensity activity in the game (assessed with total high intensity running distance; THIR ) was related to team success with teams finishing in the bottom five and middle ten Premier League positions completed statistically significant more THIR compared with teams in the top five. Also, the new study by Carling et al.(2012) [ ABSTRACT] showed that Repeat Sprint Ability (RSA) might not play crucial role in a elite match performance as commonly believed. What are your thoughts on this, especially taking into account that those conclusions were based on velocity based time-motion analysis?

Julen: Really, the apparent contradictory results of both papers are logical, why? Simply, it is not new, actual performance within a team-sport framework is a complex concept. Nowadays, there are more and more papers about the ‘contextual variables’ (like match status, quality of opponent, location…) and time-motion or playing tactics, for example. To summarise, there is a number of variables that could explain physical workload in soccer players, and combinations of these variables could be used to develop a model for predicting (from a probabilistic viewpoint) the physical activity profile in competition. Some of our research arrives at this conclusion: the player was to make more intelligent runs rather than simply running for long distances. The winning team will probably run less than the opponent, but I wonder, did they run less before scoring the goal? Or on the other hand, run more and once the team has scored a goal, use another kind of strategy to keep the advantage (and run less)? Often, to evaluate performance analysis researchers use all of the game to assess it, but during the game is there a relation to other contextual variables that influence the player’s physical and physiological demands, and one of them being match status. Maybe we should evaluate the physical performance whilst keeping in mind the score. Indeed, some papers suggest that effective assessment of soccer performance at a behavioural level needs to account for the potential interactions between situational variables. To answer the question, we can not to assess sport team performance using only physical point of view, we need more information, because among other things, used play style can be different and so enhances other physical demands.

Mladen: When it comes to soccer training, especially lately, coaches use SSG (small sided games) exclusively to develop soccer-specific endurance, even speed and power [click for more HERE]. What are your thoughts on such practices and can it be used for all levels of players (youth, adult, elite) and/or all positions. Is there a ceiling/plateau after which SSG cannot provide further stimuli for improving physical qualities and yield no transfer to a game? Can we achieve all needed physical adaptations by relying solely on SSGs as a method of conditioning/training?

Julen: Yes of course, I agree with you. The SSG can only help players in some physical qualities or areas and only to a certain level. When players get to one particular level, SSG cannot provide further stimuli to improve physical qualities. Players need other methods to improve their qualities. But we have to think that players need to optimize their qualities and not maximize their qualities. To underline, the play performance is more important than the physical performance. We must be careful! Attention should be paid when using SSGs in training programs because this training method probably would fail to provide stress on activity variables deemed to potentially promote adaptations for the development of game repeated sprint and repeated high-intensity activity. With all, there is no other option, this requires the tracking of players’ training load every day (if we can).

Mladen: Some coaches believe that 2v2 and 3v3 SSGs (and not 1v1, 2v1, 3v2 finishings) develop power of the players, thus negating the need for power/strength training. They base their rationale on how players feel after it (sore and heavy legs), but I believe that peak power output in those exercises is actually lower (or they spend less time and less occurrence at high power/speed/acceleration/deceleration zones) compared to bigger games due the proximity of the ball and opponents, but the frequency of medium-high efforts and zones is higher, thus the workload is higher on average. In my opinion this is “flaw of averages” and biased view. What is your viewpoint?

Julen: For an ideal performance in team sports, such as soccer, players need to optimize their technical, tactical, physical and psychological capacities. In this way, it has been suggested that the small games can improve the above mentioned skills of simultaneous and specific form. Nevertheless, although these situations of training replicate the majority of the demands of the competition and that they can be an exercise adapted for the development of some principles of the ‘play model’, they might provide a deficient stimulation of high intensity activities, requiring coaches and trainers to complement this training with other types of drills or carefully configure these tasks with the intention to provide the player with an ideal stimuli of training.

Mladen: Recent study by Buchheit et al. (ABSTRACT ; Slides HERE ) showed that we cannot expect linear connection with improving/decreasing physical qualities (MAS and Vmax) and changes in physical game performances. Taking this into account, how do we know whether improvements in physical qualities (MAS – maximum aerobic speed, acceleration, deceleration, agility, maximum speed, etc) yield improvement in physical game performances or do changes in tactical situations yield those improvements? Also, is there a certain threshold after which further improvements in certain physical qualities yield no game performance benefit? How do we know that?

Julen: Absolutely, more and more studies focus their results in that same line. As I have previously commented in football the priority is not the physical condition. Prior to this, skills and decision-making are key for success, and moreover, all orienting to the team or tactical behaviour. During its history, football training has had different stages. The training methods depend on the era, especially influenced by winning teams. Probably, nowadays, if the German teams keep their superiority other kinds of training methods (and ‘play model’) will be copied. Regarding different styles, I am sure that a minimum level of fitness is necessary (footballers aren’t sedentary), due to the high pace of competition, increasingly during the last few decades. To assess physical qualities we must be very very specific, try to propose the evaluation of the same physical and physiological demands. The physical qualities should allow players to be prepared to keep their fitness for a whole season (long competitive periods within and between national and international competitions). Furthermore, training must be specifically adapted to player specificity (to their strengths and weaknesses), try to avoid unforced injuries. Permanently evaluating MAS, acceleration, deceleration, agility, maximum speed and others (e.g. in individual areas like biomechanical, physiological…) could help us to diagnose and make decisions regarding player rotation, overtraining, fatigue, recovery strategies, periodisation, influence of training loads on physiological responses and adaptations, risk of injury, inter-individual variability in the responses and adaptations to training, and a lot of more, that although they are not the most important in this type of sport we have to bear them in mind.

Mladen: Speaking about training, how important is to conduct specific intermittent intervals (i.e. 15/15 with changes of direction) for improving endurance and why are they better than more generic conditioning like 4x1000m or 4x4min? Wouldn’t too much of specific work (especially the one that includes a lot of changes of direction) yield specific over-use injuries? Is there time and place for generic training, like intervals on the bike or 4x1000m runs?

Julen: In my opinion both ideas could be valid. Considering both general and specific work, it is better to keep a balance. In Spanish we use one sentence that can sum up this: “todo no es ni blanco ni negro”, it depends. Each country, club, team and player has their own idiosyncrasy, so there is not a unique option. It will depend on multiple factors that I can’t list now, but everybody knows or suspects. Linking to the next question, I am closer to the Tactical Periodisation (“Periodización táctica”, Portuguese proposal) than other options, of course, proposed systematically and especially assessing players (experience tells me that it is unusual). This type of periodization involves games principles, every day, week and month. The training skill is almost always a game. This means not leaving aside the structural features of the game when preparing any task, as Ecological Dynamics argues: in designing practice tasks that faithfully simulate performance environments an important challenge is to share information and action, allowing emergent movement patterns. Playing football is different to simply running, jumping or changing direction, although to play football players have to run, jump and change direction. Decision-making comes before everything. But once again, although most training contents should imitate the game I disagree with only taking one type of model which would be a restrictive training approach. I agree with mixing methods.

Mladen: When it comes to periodization there are multiple solutions that include block training and concurrent training. How should a coach periodize the pre-season and in-season? What should one do during the long in-season to maintain fitness levels and avoid injuries? A lot of coaches use Raymond Verheijen rotation of SSG – what is your opinion on this?

Julen: In professional football the physical and physiological periodisation principles (that are both individual concepts) have most importance during pre-season, all players have to have a minimum level of fitness, as close to their previous years. Once an elevated level of fitness has been achieved, the team has to keep their fitness platform throughout the season. Independent from different options that the coaching team proposes for their players, the variability can’t be very high because each weekend the team must have a maximal performance (in some cases twice a week), in other words, all games have the same 3 points to win. From this point of view the block periodisation concept might not be the best option. Methodology adopted by individual sports could not be applied in team sports. Team sports need team principles, not individual principles. Another thing is that for physical performance coaches it is easier to program and assess individual qualities. This is necessary but not enough when it comes to team sports. Raymond Verheijen proposal seems a good option too, but I need to read more scientific evidence around this to evaluate adequately (in the same proportion for Tactical periodisation). In my point of view this proposal stems from the player and not from the team. Once again, we are trying to apply individual principles to team sports and I think we must begin from the mean, but it is only my opinion. Maybe, for this reason, when a team play better they are running less; although, it is true that training to run more is easier than training to run better.

Mladen: Thank you very much for sharing these insights Julen. A lot food for thought and some very important concepts. We are looking forward to new research papers from your group. I wish you good luck and a lot more studies.