Team selection like in the NASA: Krzepota and Skorko about esportsLAB and x-kom AGO research project

What do observations of players who may potentially join x-kom AGO look like? What will the ideal player be like? Are the tests not too complicated? Sebastian Krzepota, Sports Director of the organisation and Maciej Skorko, President of esportsLAB, talk to us about it.

I have seen an opinion somewhere that your team selection resembles NASA tests. Don’t you think what you are doing may be too complicated for creating a Counter-Strike team?

Maciej Skorko: Esport, like many other spheres of life, has been developing for years on the basis of intuition and knowledge of hundreds of thousands of people. Each and every esport player is an example of this process. They achieved it themselves, in competition with others like them, this is how it works and worked 10 or 20 years ago. the problem is that this approach does not guarantee repeatability and understanding of the importance of individual differences. Our experiences are our experiences, they are very personal and we are unable to make an objective assessment of reality. All this means that if we only base ourselves on intuition, we will not understand the nature of processes. So, if someone says that we are performing tests for NASA and not for an esport team I reply that we are doing it in order to build an objective picture of the people we are dealing with. esportsLAB is engaged in a project which tests the assumptions of a theoretical model which describes the correlation between the individual level of game skills and the characteristics of the player in the areas of mental ability, mental determinants and physical health. I am convinced that once we collect this information and combine it with individual assessment, we will be in a better position to understand what goes on in the game environment.

The effects of our observations do not guarantee success. They constitute an element which is needed to enable the appropriate people who collect information from various sides to make the correct decisions. There is a distance between what we as esportsLAB are able to deliver and how the information will be used by organisations and players. However, with our data someone may make a better decision for themselves. At present in esport top players have similar skills, they are testing the limits of human capability in terms of mastering the mechanics of the game. This is why understanding the factors that affect the variability of the level of performance may provide great advantages. At present organisations tend to have a better understanding of this, although they still often operate on wholesale basis – if an idea appears with new physical exercises the entire team is engaged. Does anyone think how this activity will affect this or that individual player? Maybe one of them would spend time more effectively by regenerating because he or she had been under more strain from competition or training. As long as there is a deficit of data collection and building of knowledge banks about competitors in esport, no one will be able to give the right answer to these questions.

Sebastian Krzepota: If people refer to NASA tests, I can only say that if you go into esport and look at the strain that the players undergo you will see that for a normal person this is „NASA”. Someone who does not play computer games at professional level is not able to understand the things that are absorbed at such speed by players in the game environment. If we look at the skills which are required from a person during the game, I can say that e-athletes are outstanding individuals in this respect. Outside the game we need to remember that like every one of us they have expectations from themselves, their team mates or the world in which they function. These expectations are mainly related to winning. There is nothing wrong with that up to a certain point, because you understand that in the entire pool of competitors who have the same objective only one team can win anyway, and the rest loses. If you do not have a longer perspective on your own and your team’s development, frustration, doubts, decisions to move on or even resignations frequently appear.

We decided to take a different path. First ,we defined the criteria which we believe are important in an e-athlete’s environment, now we are gathering information about players’ features and will try to form the team composition in a way that will ensure the best results in the esport competition environment.

After more than a year’s work in AGO and observing the team can you definitely say that the data collected by esportsLAB will be very helpful in selecting the players?

S.K.: The set of tests we selected in the respective areas are tests which have been used in population assessments for many years to determine trends, both in terms of the studied group as well as the population. Firstly, they are meant to help us to select player profiles, which constitutes the basic information for composing a team and planning team life. Secondly, such activities will enable us to observe micro-evolutions of the person, the esport player. We must remember that in competition it is possible to face an opponent who is theoretically better but has a bad day and can be defeated. It is also possible to come across a theoretically weaker opponent but lose the game because you have a bad day. This is why we believe that if we do not collect data and analyse them, our assessment will be based mainly on emotions and expectations, which is like reading tea leaves and chicken entrails. We do not want that.

M.S.: We started from the organisation level and what esportsLAB can do for x-kom AGO and other esport organisations, but the truth is that looking at these players we want to help them to understand what their performance level depends on and to what extent. Our main objective is to develop the players’ self-awareness, the coaches’ work methodology and the know-how of the organisation. At present, in the players’ assessment, quality (intuitive) analysis dominates, while quantitative assessment (data from the game environment) is limited to the level of mechanical operation. Some are better some worse, but at top level in our assessment they are all excellent.

When did the study concept appear?

S.K.: The idea was created by the players we worked with. They told us that something was wrong in this and that aspect. We knew it, but we wanted to leave subjective perceptions aside. esportsLAB took apart mental abilities, complex behaviours in the game environment, the players’ intuition and experience and mental determinants. To this we added human physiology, which is the same irrespective of whether you drive in rallies, play games or are a journalist. We selected areas on which we want to focus and adopted processes for collecting data about players, to see how far we can go and in what time. All this in order to be more effective both during the game and afterwards. You function differently when you play one important game a week and when you have three to play. If we want to be at the top this is what lies in store. This is why we decided to go ahead with the process, so that the expectations with regard to our objectives are based on concrete criteria.

M.S.: The software we are developing, at the end of the day, is meant to enable prediction of the esport form of each player. We want to be able to set an important tournament as an objective and develop the path to get there. We will not guarantee anything, but it seems to me that without these tools esport will not be able to continue developing at the current speed. We are now receiving important lessons, in which we adjust the research reports from our scientific experiences to the specifics of the work of an esport team and the requirements of the organisation.  

Does the esport community understand this?

M.S.: In my research work I have been in contact with many players who gave us a considerable amout of their time. We always met with a great deal of curiosity because they were themselves quite surprised to see how their passion and experiences changed their brain. The essence of our approach is to build new solutions by combining methods from various fields of science with the knowledge and intuition of players and coaches. In my opinion community scepticism disappears once we have the ability to show the results of data analysis, our summaries after the first months of our work. We are all only human and we are not able to remember how we felt or how we functioned in a game even five days ago, our memory is flawed and prone to distortion. This is why we need to create tools to enable us to collect quantitative data; some correlations, especially at the individual level, come to light only during analysis of a sufficiently large volume of data. A good example of this is the effect of the quality of sleep on the number of eliminations in CSGO games played on a given day. This correlation proved to be so strong that I found it difficult to believe that the analysis was correct and we repeated it just in case, with the same result. Conclusions such as these may be implemented at the individual level in order to be able to reach a better understanding of the competitor’s individual disposition at a level that is normally unavailable. These effects of our work tend to be quite convincing.

S.K.: I had mixed reactions. I met players who grew up in an environment where the assessment of performance was based on whether you win or not. If you do, you are good, if not, you can’t play. Already after the first two months of work according to the plan drawn up by AGO we were on a rising tide, we won the first LAN, went up to ESL Pro League, we won in Minsk, then the first instances of declining determination and impatience appeared. I had the impression that the players were waiting for a change in the team composition because that was what they were used to. In subsequent months we showed them that performance was a result of a number of factors, we did not know everything, but we wanted to build the best model with these particular players. However, when the worse times came and we knew that the players were not yet ready to adopt a different approach we decided to let them take the lead and do it according to their experience and treat our guidelines freely. We established how it ought to function from their perspective and collected data, assessments. It emerged under this model they were even less able to survive, one player left, the second said he wanted to go to the dugout.

We employed other players as we wanted new experience and an idea of what the differences will be with a new team composition. We had both positive and negative feelings. However, we wanted to objectivise it to be able to help these and other players. After Kap3er and SZPER joined we informed the team that we are starting to work with an external institution which will assist us. Everyone was pleased that a new system was being built, in which they would be able to participate. We started in 2019 by summarising the previous year’s work and setting guidelines for the coming months. We focused on setting sports goals and the workplan designed for achieving them. At the same time, with esportsLAB, we started to create a system of gathering and analysing data. The players were evaluated by scientists, coaches and by themselves. The players’ observations were, in many aspects, accompanied by long hours of disputes and creating a skeleton mechanism on which we wanted to base our work in the esport organisation in the future. All this happened at the time when we adopted a preparation plan for WESG. After we returned, we shared our observations with the players. We came to several interesting conclusions, one of them being that we are more effective when the players played together, in one place. This was explained among others by the players’ personality traits.

The next goal was Minor which led to Major. This goal was not achieved, so after we dropped out of the eliminations we received not only positive opinions, like after China, but also doubts. In our overall summary we decided to start from the beginning in our approach to team building. We retained two players who wanted to achieve their goals on the basis of the way we propose, and we said goodbye to three players who had similar goals but a different way to achieve them. We took all the lessons we could from the experience and we believe that each of the groups, those who left and those who remained, will make use of it in the future to improve their performance on the CSGO scene. We all remain open to the possibility that one day our paths might cross again.

M.S.: Reactions vary because for many this is interesting, sometimes perhaps strange. I am sure however, that as we continue on our path we will win over players and organisations to our approach, by the effects. At the moment we are in the beginning of this path, but we already have a large body of scientific research and a number of pilot studies. I believe that a good example is GruBy, who was the most sceptical of all. He was the only person who wanted to give up the entire undertaking and he had every right to do so because participation in measurement sessions is voluntary. I had an individual discussion with him to convince him to trust us, to stay in the project and work diligently. When he got the report, he was the first competitor who used it to continue developing his knowledge of himself in the context of performance level. However, today we need time and cooperation with the widely understood esport community in Poland in order to be able to create a system together, which will help people every day. 

On the basis of experience gathered with professionals and semi-professionals we also want to build a development model of a young person in esport, in order to understand at which point to decide whether he or she is more predisposed to play RTS type games or perhaps FPS type. If we started working from the bottom upwards, that is from amateurs, we would not be in a position to understand a great deal of issues. Largely because at that stage the level of performance is so varied that it is much more difficult to understand what this variability depends on. At the top however, the composition of phases and length of sleep will decide whether a player will have more or less frags. Such minimal things may determine a won or lost round. We will operate on a small scale but we believe that at the top level small scale makes a difference.  

What features must a competitor have to fit your picture of the best player?

M.S.: From esportsLAB perspective there is no clear answer at the moment of what the model e-athlete is like, who he or she is. This is mainly due to the fact that today no-one in esport has the information which could enable such an assessment. Nobody knows how to select five players in such a way that they actually supplement each other. These are very complex processes, and the essence of our approach is to gain knowledge by means of a continuous research process, with each and every iteration giving us a better understanding whether the criteria we chose are accurate. Time will show, the most important thing is that we will know why something did not work, why our theoretical model is not confirmed in reality. As a result of this approach and scientific method we will determine what can be done better next time in the same situation in order to achieve a better end result. 

S.K: In our view a player ought to be open to development, have the relevant features, skills and abilities in line with the system adopted by x-kom AGO to compose the team. And obviously have the determination and patience to achieve goals.

In short, these observations are an experiment for everyone?

S.K.: Apart from myself, because on the basis of the data gathered by esportsLAB with the team of coaches I will have to decide which players will suit the team profile defined by the coach. The chosen players will have to provide the lacking elements and together with GruBy and Furlan create a monolith. This composition will be given a Chance to develop in the organisation. I would not call it an experiment but the process of learning a “subject” – esport. The observation of competitors on the basis of selected tools will enable us to understand the requirements, formulate theses and implement the plan, and then make an assessment. For me, this is what constant development consists of.

M.S.: The role of esportsLAB in the process essentially consists in providing assistance. A comment appeared that we will invent a formula for an ideal team. We will not, because in many spheres of life human intuition cannot be replaced. A good example is NFL and the New England Patriots team, which won everything recently. The players of this team are controlled in an unbelievable manner, coaching staff is constantly watching over them, they collect key information every day. If someone asks why they are winning, probably only the coach, Bill Belichick, knows the answer. He can direct the coaching staff and the player team to use all information in such a way that on the playing field they achieve the best result. It is possible that many more years will pass before IT solutions will be able to replace people in the complex process of composing a team. Maybe this will never happen. 

I also believe that is a phenomenon in esport, no-one can replicate their success. And nobody knows how it actually happened. Was it that they were so well matched? Or did one of them have such features that made the other four play better?

S.K.: If data had been collected at the time, today they as well all other players in the world of Polish CS:GO would be at a different level. A higher one. 

M.S.: Obviously intuition is key, and it is the coach who will make the decisions at the end of the day, but he has to have ground to develop methodology; experience is priceless but in our opinion knowledge recorded in databases subject to constant analysis is of even more value. I appreciate the directions in which esport is developing, that it reaches out to sport psychologists, dieticians, personal coaches, but the effects of these activities do easily lend themselves to measurement.  

So what are you planning to do after you make your selection?

M.S.: We are creating the first versions of complex software for collecting data about the team and planning individual work time. It must be remembered however, that at present these tools are pioneer solutions and we have a lot of ideas how to evaluate them. I have been involved in creating tools for gathering data about people for a decade, as such I have a great deal of knowledge about assessing their effectiveness and adjustment to the situation. These tools will be used by the organisation on a continuous basis, because they are activities of research and development nature. This means that we are collecting data, trying to understand how to use them in everyday work of our players, coaches and organisation. Later we move on to the next iteration, in which our model is better suited to the data than in the previous one. In this way we will build tools which, I hope, will also be used by other teams. There is no better way to create such a solution than involving the greatest possible number of teams in the project. I have an idea, which I call the Esport Republic. I hope that within 3-4 years as a result of cooperation between many Polish organisations we will create a system of training enabling serial „production” of top e-athletes and coaches on a global scale.

S.K.: Once we have the tools created on the basis of the knowledge of esportsLAB researchers and programmers, our knowledge and player observation, we will be in a position to select the appropriate players. After 18 months we know what the key features for reaching the goal are. We also know that the waiting periods for success may be much shorter. We will select players who fulfil our criteria, and later on we will review how they function in the dedicated work environment. And either we will continue on the same path, or we will improve it. In any case we do not want to close ourselves to being better still. I believe this process is indispensable in an esport organisation.

M.S.: I would only like to add that our short term goal is not to teach professional players their trade, they are the experts in CS:GO, Dota 2, LoL or StarCraft 2. In this respect it is quite the opposite, we are learning a lot at every meeting. Nearly everyone in our team has some experience with esport at the amateur or semi-professional level, but it is not enough to, for instance, create an effective quantitative method of assessing the tactical abilities of professionals.

S.K.: Summarising, all that I say to the players from the first days of observation is that they are must be themselves. We will give them tools, we will show them the way, help them to understand themselves better. I must remember however, that at the end of the day not all of them will be able to stay with x-kom AGO, but we want every one of them to be able to use the information gained here for their own development. We will continue to work with those who stay on the basis of the gathered knowledge, and those who are outside the organisation will be observed by us from the point of view of their development.

Photo: x-kom AGO