Progressão
Progressão is a book, a podcast, and a long-term thinking project focused on football, learning, and skilful human behaviour. Our work approaches football from a complex, holistic, and ecological perspective, where players and all football actors are understood as living beings always in correspondence with their environment.
Progressão
#197 DISCUSSÃO MUNDO: Rodrigo Picchioni – data, scouting, and what we can't measure
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Data scouting. One of the most intriguing and fast-developing corners of professional football. Most people have seen a heat map. Very few know what a Lead Data Scout actually does on a Tuesday afternoon, or what the method can genuinely explain about a player, and where it hits a wall.
Rodrigo Picchioni built analytics department at Atlético Mineiro and now works at AS Monaco. In this conversation he crosses two worlds that rarely talk to each other directly: the rigour of data analysis and the deeper question of what we can do with data, and what not.
In this episode: what data scouting actually is, how the Monaco model differs from what Rodrigo built in Brazil, and what the hardest things to quantify reveal about the limits of the field. The Moneyball promise — has analytics democratised football, or have rich clubs simply bought both the data and the best players? A provocation from @ThePurist on the felt connection between players and audience that precedes analysis.
The most interesting version of this conversation is not data versus the eye. It is what it means to understand a footballer at all, and whether we are getting closer.
🌍 More at progressao.fi
📷 Instagram @progressaofi
👥 LinkedIn @Project Progressão
Holdistic, skillful jackal kyrin jail management. Welkom Diskustan Mundo. Our näistä guest is Brasilian, has worked on three kontinentä todella kuin lead data scout at AS Monaco. He built Atletico Mineros Analytics department from scratch ja won the League and Cup Double with them in 2021. He has spent his career trying to understand what data can tell us about a footballer and the game ja what it can never. Rodrigo Piccioni, welcome.
SPEAKER_01Thank you, Janni. Thank you for the invitation. It's a pleasure.
SPEAKER_00Let's first start with the ongoing World Cup and of course Brazil is there. And big expectations as always. Carlo Ancelotti, new manager. Cele São with superstars, many good players. Rodrigo, what's gonna happen?
SPEAKER_01I don't know. I think I think national team football, even more the World Cup, it's extremely unpredicted and random for the I'd say especially for the tournament format. So I always do very bad on those prediction games, you know. Always do very bad. It's very hard. We try to come in a logic way, and there's no logic in this.
SPEAKER_00True. Many things can happen throughout the tournament and and in one individual games. Uh, what do you think about the Brazilian team? Like, what can be the expectations for you?
SPEAKER_01It's obviously a strong team. I think we were hit with some difficult injuries that uh hurt the squad. So in key positions, like positions that we have been struggling uh for the past cycles, like fullback. So we lost Militown, who was like the only short thing in as a fullback. We lost Estevão, we lost Rodrigo. Okay, up front, we we have many, many talented players, but uh I mean those are guys that were likely uh gonna be starters. I think that the team is solid. Again, I think the biggest concern in Brazil right now is is around the fullbacks, and it's already an interesting uh interesting reflection that we sometimes discuss with some Brazilian peers that Brazil is is hardly producing few fullbacks these days in the past. Why is this?
SPEAKER_00What do you think?
SPEAKER_01Okay, one one theory is this is if you get the 90s and the 2000s, the main formation that Brazilian coaches and Brazilian teams used to play was like a 4-2-2-2, 4-2-1-2-1. So they were developing a lot of fullbacks in the in the youth. So the powerful guys, the fast guys, they would go and they would dominate the the sides of the pitch. Today, what we think happens in the youth academies is that when you have players with these profiles, fast, dynamic, powerful, they put them in. So there is a big lack of of Brazilian fullback development recently. And we are seeing uh we we are repeating a World Cup. Great players, historical players, uh had great careers, but we are repeating a World Cup with Danilo on the right, he was the right back, also less World Cup, Alexander on the left was also the left back, left World Cup. Four years later, you know, they are both they already left Europe, you know. They and they they still have yeah.
SPEAKER_00Okay, okay. When we think about the coaching staff and head coach, of course, Ancelotti is a club manager by DNA. How do you think he has started in the national team and what expectations you would have of him?
SPEAKER_01I think there has been a an adaptation period. I'd say like uh euphoric starting period. Okay, we have Ancelotti, then a plateau of adaptation period, and the recent friendlies were encouraging uh leading up now to the World Cup. I think he was getting to know more the players and understanding how he would fit all the puzzles together. Again, national team football, what is important? I don't know. Uh we can say Andrew is a great leader, a great manager of people, uh able to form the Eagles, and one can argue that this is what really matters. Uh he has pragmatic tactics, okay. But is it? I don't know. Because you have examples in in all sides. Then one one can also argue okay, this is gonna be a very hard World Cup for the legs, it's gonna be extremely hot, long duration track. How much does him have in a more I don't know if it's the right word, a more conservative playing style that okay Brazil is not probably gonna press like crazy? Is is is this a benefit in this World Cup? Maybe I don't know.
SPEAKER_00The answer is I don't know. We will see now during the tournament. Uh speaking about data, do you have any any data about how national teams perform under club culture coaches? Has this been studied at all?
SPEAKER_01I I have never approached the study, it was never something that was that I was particularly interested, to be honest. Uh so I don't know. I think again, the randomness of of this. If we if we take Brazil 2002, we won the World Cup with Filipão Scolari, who was uh until then purely uh uh a club coach. And he comes in, he creates the Scolari family, that's what they call, and Brazil, uh it's very similar actually. Brazil struggled in that World Cup to even qualify. Uh and he comes in, like Ronaldo two years without playing. He creates that environment. Okay, we take it all. But then you have Argentina winning with a coach that never coached the club before. He came to the youth ranks of the of the Federation, Scaloni. True. You have Jorgin Lowe winning with Germany. I'm not even sure if he coached the team before. I mean, we have we have positive and negative examples in on both sides. In the other side, you have a guy like Capello coming to England and be a complete disaster, you know.
SPEAKER_00So true. There there are different examples of this. When we think think about data, data scouting, and if you analyze a national team tournament, do you analyze it in the same way? You already said that it's random, it's different than uh than a league, sample sizes are different, adaptation time, all this is so compressed. How can we analyze World Cup?
SPEAKER_01From a professional point of view, I am extremely, extremely skeptical of using short tournaments and especially important national team short tournaments like Euros, like World Cup, like Copa America, for example. The first aspect for me is low sample size. You know, even if you are okay, we can we can see how a player will perform, let's say, a you know, eventually we'll touch about this, but league translation, you know, players facing against different uh harder competitions is something hard to have opportunities to see. And then okay, but then you can you can see this in a World Cup scenario, yes, but then for me it's it's one game. And I'm confident that we with our eyes can analyze the behavior of a single player in a single game in much more depth and confidence than a data model. If I have power, the the biggest leverage that we have with the data models is that we are able to see big, big chunks of game. So we have uh large sample sizes that the human eye can process and find patterns and an algorithm can. Uh, but for a single game, uh single player, sometimes in a single game, I'm much more confident in our capacity to observe a player. So, from a data perspective, I'm extremely skeptical. Besides the sample size, I think there is a huge emotional burden in these games, and it's also connected to the sample size because you have a low sample size and it's not really a meaningful one, maybe. Uh, so we want a repetition to take away these this effect. So I'm skeptical, and I I I don't use it to be honest. It's nice, it's nice. Oh, let's see, but more in a curiosity, recreational type. Okay, let's see who did what in the World Cup. It's nice, but not as an evidence to support processes in a professional.
SPEAKER_00Very interesting. It's kind of um paradox that we have. World Cup is, you know, people watch it a lot, and it's so much in the frame, and people analyze it and they want to make uh, you know, conclusions, uh, predictions and conclusions based on that. But at the same time, we know these problems that you already brought out.
SPEAKER_01I think players can get you you can pay a World Cup premium on a transfer fee, for example, because the player had a great World Cup and it's what you said, it's the hype, right? So everybody's talking about the player, but the the confidence for me is low.
SPEAKER_00We have good examples of this that somebody played very good World Cup, he was you know bought to a new club and he don't make it there because he was he was short time in a few games, he was very good, and everybody thinks that he's not a super player, and then he was not. Let's go into data scouting. So, you have built analytics departments at Atlético Mineiro, for example, and now you operate in Alas Monaco to very different environments. Uh, can you please tell us what is data scouting?
SPEAKER_01To be honest, I don't like the name, I think it's too narrow. Uh, I like something more broad like scouting analytics. Um, it encompasses better the how I see the role because I think the traditional view of the role is okay, this is a profession that will identify players and build some reports, some nice visualizations, and to support this uh smaller scope of the scouting process. But for me, I see the role of as being in a position to support and impact the entire uh scouting process. So, yes, doing reports, uh identifying some players, but a lot of it for me is building systems and processes and automating systems and automating processes that support the scouting operations uh in order to enhance other scouting practices that are not related to data auto. So I also see my as okay, how can I generate create systems and tools and automations that will allow the scouts to be able to perform their traditional work better and with more time and with more tools to use. So I see a lot of the rolling around this.
SPEAKER_00I think it's a very this is a very good goal because then you're actually you're actually helping the scouts, you know, do their core work well. So you're not giving them all the different kinds of things that you have to do this and this and this, but actually reduce things and make it simpler so that they can do their work well.
SPEAKER_01I I'd say exactly this, uh Yanni. It's about reducing I call non-artistic workload because the art is to watch the players, is to watch the game, is to flag the players and be really deep and watching video. Let me watch this type of this is the art. So everything around I try to take off this burden from them, you know. Which game should I watch? What uh kind of player should I be aiming tournament X? You know, uh, which player played uh X amount of minutes, which players debuted, like this, all this everything that is around this. Ah, how can I make sure that I'm conveying the message that I want in my scouting report? No, your the art for me is watching the game. Uh, and this is what I I try to give them the the focus. And I I think the the other part also is about research. I do a lot of research.
SPEAKER_00What kind?
SPEAKER_01Can be like ad hoc research about specific players. Let's say there's a player that we have a theory and we want to go deep on it. Let's test the theory, let's see if our hypothesis is sustainable, if there is evidence around it, and we find something okay. So there's this type of research that's like player recruitment specific. But then there's also research about the uh the environment, you know, the market environment. Uh, are there leaks that we should uh be focusing that we are not are there, methods that we should implement and that we are not are there, peripherical aspects of the transfer market that we can explore better? Are there uh deal structures, for example, this kind of stuff, like doing broad research as well around everything that's related to player trading?
SPEAKER_00Uh, you already mentioned that uh players in different leagues, the context is different. So, how how do you translate this information from one league to another? You know, that you have a player who plays in, for example, let's say Nordic leagues, and you would scout him and try to understand how he would do in uh Central European good leagues.
SPEAKER_01So there are there are both public and and private models around this. We have data science techniques to to approach this problem in a in a good way. Uh, I haven't yet seen a model that it's that it's perfect, uh, but there are different ways to approach this. You can create, for example, chains of transfers and see how player performance outputs have have regressed when going or improved when going from one reason from one uh competition to the other, and then you can create like a synthetic chain of transfers to estimate transfer transferability, even if the there has never been a transfer to that league. So, for example, let's say there has never been a transfer from Finland to Brazil, for example. Um, but there has been a transfer from Italy to Finland and from Brazil to Italy. So you are able to infer this uh the synthetic uh transfer chain in a way, which is very nice. This is a model that is public actually by the guys in Gemini. It's it's a very nice read. In Atletico, we did we did uh a mixed model, and I thought it was a very interesting uh exercise. We would update every six months. So essentially, we we gave a questionnaire to the practitioners, to the to the domain experts, which were the scouts. Uh, and then we had okay, this is all the leagues we cover. I want you using whatever method you want. Some will use one type of method, one will use another type of method. Yeah, I want you to rank these leagues based on your knowledge. So we would take this as a part of the signal, and we would combine this uh with advanced data modelings in order to find okay, this is the almost like a wisdom of the crowds uh approach.
SPEAKER_00Yeah, very nice. I think we had a similar type of uh process in Independiente del Valle, if I remember correctly, that the coaches had to kind of rank many, many different leagues. And then when you get this all info from, I don't know, 100 coaches, for example, you might get some info from that.
SPEAKER_01And and it's nice also to see the the differences, you know. People experts judging sometimes the same thing, and just especially when you go further down the chain, like okay, how do you rank uh I don't know, uh second division of uh Portugal versus uh I don't know uh Poland? Okay, um you know so true, true.
SPEAKER_00You can get very interesting answers from there. When we think about the different contexts, you have been recently in uh in Brazil in Atlético Mineiro and now in As Monaco and Europe. Is there a difference between these contexts and if yes, what kind?
SPEAKER_01The the biggest difference is uh here in Monaco we are MCO, so we have uh we have Circle Bruges with the group as well. This is the biggest one. Uh we are operating with with two clubs. Um additionally, we have a very strong operation with EDG, which is the underlying team, our elite development group. So essentially the the operation is is around three clubs instead of one. So this changes the the workflow. Another thing that was different in the first time to me is that the scouting operations in Atlético, for example, Brazil is a big country, and we basically deployed scouts. Uh we had like the central department that was based in Belo Horizonte in the in the training center. Then we had one scout in Rio, one scout in Sao Paulo. Uh, but the bulk of the operation was internal. Uh and and here, and I think it's European, it's a more aligned with European culture. You know, we have scouts deployed in different regions of the world. Uh, and this dynamic of having experts in each region of the world uh was something new for me and very interesting to see as well.
SPEAKER_00How has that been working? Like a different you have now people in different time zones and different cultures and everything, so is it easy to Yeah.
SPEAKER_01Um most of them are are based in Europe, so the time zone is not really an issue. And we try to consistently integrate these scouts uh internally, like doing seminars, you know, we do constant meetings. So I think that even though they are not here and there might be different cultures, the way that the club approaches decreases this this barrier because we are we are very tight together. It's not like oh we only guys once a month. You know, we are constantly talking, doing meetings, they come here, so it works very well.
SPEAKER_00So, how do you see data? What can it describe what happened and can it explain why?
SPEAKER_01Yes, I I'm very confident that it can describe what happened. I think that's yeah, we have we have very good confidence that it can describe facts and be descriptive. Uh can it explain why? Like, can anything explain why?
SPEAKER_00It's a good question, yeah.
SPEAKER_01I'm more skeptical on this. Like the the chain of events is so big and random and multidisional and chaotic and multifactorial that yeah, I think we can make hypotheses on why, I think we can uh create approximations on why. Uh I think we can make uh inferences, and I think data can do this as well. Human observation came with a different approach again, using the big strength which is large sample sizes. So, like, no, I I'm not confident that data, you know, black and white answer can tell you why, but uh I'm also not confident that human observation can.
SPEAKER_00Very good, very interesting. Philosophic philosophical question also. When we think about event versus continuous data, do you see there is a difference, you know, how what they can explain and what you use?
SPEAKER_01I think the biggest gain here is being able to see the other players on the pitch, so you are able to have more context so you can model better football actions.
SPEAKER_00How do how do you define football action, by the way?
SPEAKER_01Well, when we are using tracking data, uh any action can be a football action, right? It's just a matter of how we categorize. So a run behind the depth can be uh can be a football action, an approximation support the ball holder can be a football action. Uh when we are talking about event data, we are talking about football actions on the ball. Uh and depending on the provider, some defensive actions. And then when you put a layer of tracking data inside on top of this, you can have more details about that football action that will generate subcategories of football actions. So, for example, uh you have a pass. A pass is a football action. Uh, when you put the tracking data on top of it, uh you can have okay, was that pass in between lines or not? Okay, so this is a different football action that's uh with additional event data, you can't you can infer, and maybe you know, using the location of the pitch, game situation, and maybe in a large enough sample you will be able to have a decent uh approximation, but uh with the tracking data you are able to to do more uh to have higher confidence in this kind of details. And then, of course, the other thing I spoke was off ball, you know, runs, which is this is uh a big, big game, you know. Uh what is happening that the events don't see, yeah, and this is only with tracking today.
SPEAKER_00Uh, in your opinion, what are the hardest things to quantify at the moment?
SPEAKER_01For me, it's off the pitch is adaptation. And again, it's a philosophical because I think that's true both for data and for traditional scouting.
unknownI think.
SPEAKER_00Have you seen a lot of situations where you feel that now we have a good player but he's not adapting?
SPEAKER_01Yeah. And and adaptation here can be personal adaptation. I I think not not many factors need to go wrong for a transfer not working out. And these factors can be both personal factors, like the guy can't find a good house to live, uh, or his wife doesn't like the city. But it can also be football adaptation problems. Okay, uh the coach decided to play him in a different role than the scouting department was involved. Oh, there is a kid on the academy that we didn't foresee, and he's coming up, he's taking many things, like injuries, of course. So so I think that today we are pretty good uh both from data and from traditional observation, and especially when you put them together. Uh, we are pretty good at describing what happened. I don't think there's anything like challenging about this right now. We are pretty good at saying, okay, this happened. The biggest challenge now, okay. What will happen if he comes to our contest? And and this is the adaptation, this is what I mean by the adaptation. This is the biggest hard. There's so many factors. There is league translation, there is team style translation, there is personal adaptation, uh, there's training adaptation, and so on and so forth.
SPEAKER_00What kind of advice would you give when a player comes to a new club? How to make sure that this adaptation goes well?
SPEAKER_01This is 100% not my area. So it's not something I I have the property to speak. I know clubs are investing a lot on this, you know, in player care, in getting to know the player on a personal level beforehand, going very deep on their families, on their day-to-day, to make sure they fit. I I think if you talk to an expert on this, I think you there will be arguments. Okay, you should give the player everything and make their lives easy, and there will be arguments, oh, he it's gonna be better that it's struggling, so I don't know.
SPEAKER_00Yeah, this is already a good point. No, we can go to next topic, which is still related to data scouting. So, have you ever had a situation where the data was kind of telling that, oh, this is the player, let's sign him, but the I was saying otherwise, or vice versa.
SPEAKER_01Yes, many times. And this is expected, this is needed, and this is wanted, like this needs to happen. If the data and the eye are every time seeing and coming to the same conclusions, uh there is no point in doing both, right? Uh, extremely important that there are disagreements in the process, and that these disagreements trigger not only additional analysis on the specific subject of the player, but a continuous loop to refine the data models and refine the traditional scouting process so that we we start to refine and and I won't remember exactly the situation, but there has been situations that we understood limitations on our data models because of situations like these, and we were able to improve our data models based on this.
SPEAKER_00You put it very very nicely so that there is a kind of debate between data and and uh and the scout who is going to see the player, but that debate or contradictions or conflicts can be used in a positive way to improve your work.
SPEAKER_01Yes, 100%. Uh there was actually a recent uh conversation I had with a scout that I I really wanted him to explain me very deeply his thoughts on the player, because his initial perception that he provided was very different from what I was seeing in the data. So that triggered a very strong okay, like this is in a level that there might be an adjustment needed, and I really need to understand what's going on here. You know, so these kinds of situations help the process to be more mature.
SPEAKER_00So this is uh anyway, we are talking about data, but we are talking about human work. So combining this together. There is a quote that the purist had in X recently, and he was saying: the irony of analysis is that when you are really watching good football, it doesn't need explaining. There is a felt connection between players plus audience that transcends strategy or intention. It is the same connection felt in a theater or concert hall, and it cannot be faked. What do you think about this quote?
SPEAKER_01I think that when you see this and feel this, this is our at least mine, primal self of loving football. And when this happens, these are unique moments, these are unique games sometimes. Like, I don't know. I think I felt this like in the in the Champions League semi-final, you know, of uh Bayern and um Paris. And and afterwards you might out of intellectual curiosity, okay, what is the data saying about this game? But why when you are watching that happening? Uh it's like doesn't come any anything about data. Then it's just pure football. I think my background was football first, data second. So in these situations, football uh it's everything that see like yesterday watching the opening of the World Cup or the game. I go in my system, okay. Ah, let me just uh you know, let me see if there's some interesting players here that I don't have much of an opportunity to see. Let's see what the data is saying about some of the starters.
SPEAKER_02But honestly, when it starts, it's just like okay, this is this is just uh amazing to watch, regardless of anything.
SPEAKER_00This discussion is related to, for example, Matjas Manna who Lionel Scalone in the Argentinian national team. They have been speaking a lot about relational ideas or relations between the players and such things, and how can we help those relations emerge in a better way? So this is a little bit related to progress also. And our message that skill is perceived relationally before it is understood in some other ways. And that you can see as a coach also, we are thinking about tactics and how the players are in the right positions. But then when you see that it starts to kind of flow, it starts to work, it goes into some level that you cannot anymore like put into words even easily. That you feel it a little bit like messi when he plays. You cannot put messi into a metaphor, for example, it's just messiness or something like this.
SPEAKER_01Yes, and I I obviously didn't play at good level. You you played in better level, right? Semi-level, let's say. I played mostly futile, but I think playing there were moments that for sure me, you much less than Matthew, but we felt this, right?
SPEAKER_00For sure.
SPEAKER_01Yeah. Like, oh okay, this wow, this this clicked really well. Was almost like I was doing automatic, and if if I have to explain, I can't. How how did I did this?
SPEAKER_00And how did a group of players did it together? Like, how did they find each other in that that way?
SPEAKER_01How did that guy was made you know, made a first touch pass back to me? He knew I was gonna run there, like yeah, exactly. Yeah, it's it's it's and yeah, can't explain, it's really truly impressive.
SPEAKER_00How how can how could you or how could we um then understand this phenomenon in scouting? Like, for example, three players together, there's something more than they would be alone. Can we find ways to analyze these synergies? I'm skeptical to be honest.
SPEAKER_01I think this connection, if I'm trying to be the skeptical voice here, I think this connection is not like divine grace, you know. I think this connection is intuition, and intuition in this sense, I think is experience. Experience in the sense of having visual cues that you have ex that you have been exposed to, uh similar patterns unfold like these variations of these of these patterns have happened before to you and probably to the people around. That uh our subconscious is able to process this in understanding what's the pattern that it's unfolding there now, and it's able to to show a likely action. So like I think this is a very subconscious process, and and and the process itself, I'm I don't think it can be measured, but like, does it matter to be measured? Because what matters is the output in the end of the day, you know? So for example, if if if these players have this connection and they say, okay, I I feel like I'm I'm playing here uh in a superior level because I'm with these guys, uh doing this combination, doing this one, two, and going in the depth. I feel like I'm doing this. But if the end output, the pass, the the whatever, the end output of that chain of events is not positive, consistent, or if it's consistently negative, then maybe this intuition needs some fine-tuning. So at the end of the day, we can see the output, you know?
SPEAKER_00Yeah, but I think it matters in that sense that when we are scouting players and we are bringing a new player into our team, so how could we understand that this new player with that old player can they build a new better synergy than they had before? That would be interesting. Like, how can we help players to synergize in a way and new players coming in to better synergize? Like Rafael Paul, this Paris assistant coach said that what they are trying to do, and he has his article also in synergizing phenomenon. So, can we scout players in that sense? You know, that maybe we have two players that are close to each other, but we could understand that the other one will synergize our team more.
SPEAKER_01I think that these kinds of relationships they go beyond like technical being complementary technically and stylistic. I think it's also on a personal connection level.
SPEAKER_00Yeah, for sure it's much more than only this. We had a situation in the Baltic Cup recently with Estonia under 21. So we had two games in three days, which is not a good format in this level. But anyway, we had this situation. So we took two different squads, like 22 players, and then we spent a lot of time actually thinking how do we build these teams? And I think that very good coaches that I have worked with have been able to do this. That that's something that coaches might have, like they can think that these two players together they will find each other and such things. And we were trying to do this, and actually, in some part, I would say that we did it. So we put some players together that had been had not been playing together, but they kind of, you know, you have this feeling that probably these players will find it. Of course, in the training, then you will see that if this is true or not, and finally in the game. But yeah, this is an interesting phenomenon to think about in Scout.
SPEAKER_01And for this, we go back when we're translating this with scouting, we go back to adaptation, adaptation with teammates, adaptation with the system, affinity. In Brazil, I don't know if this is true for other countries, but in Brazil we we have some like very specific noises that we do when we are playing uh between us.
SPEAKER_00Yeah, I have I have heard I have heard those a little bit when I played with Brazilian players. Yeah, yeah.
SPEAKER_01And and this is mostly between because I played in the US. I had like a Brazilian teammate, so we could understand each other because of this. So you know, we have specific noises that okay, we understand what's what he means there, you know, and this plays a part too.
SPEAKER_00Let's go and think about the future then. We have a lot of data providers. You are working with Gemini at the moment, if I understood correctly, at least. Uh, then you have other for sure other data providers too. So tracking data now covers biomechanics, off-ball position, such things. How do you feel? Where is the field moving at the moment? And what are the next steps?
SPEAKER_01Yeah, uh, here we work. Uh Gemini is uh uh is our partner in solution for both the data infrastructure, some um front-end matters as well, and that's some advanced modeling. Also, they have very capable engineers there to support us. And we work with uh event data provider and tracking data provider. Uh today we consume the the physical output of tracking data, we don't consume the raw tracking data. Do have this raw tracking data by the league, so the league provides us this. And then for the scouting purposes, we use uh physical data output of this. So this is already a first point. I I don't think tracking data is is democratized, and because it's not democratized, it's not properly used. Uh it's so expensive for clubs. Uh it has implications not only on buying this data, but also the compute necessary to process this data is very high at scale, of course. So I think that's uh once this kind of data becomes cheaper, and I think it will, because there are other developments in computing in the world that will have a uh butterfly effect on this, become cheaper to store, to process this data. I think some new layers of analysis that maybe already exist in a select few clubs in the world that are able to do this, but these new layers of analysis will be expanded uh more broadly and new things will come. Biomechanics, I think, will be the next uh step, uh the next revolution of how we analyze players, but also will have for me, and this will be more important, uh will have uh a big effect on player development, you know, around ideal, if there's such a thing, ideal technique of football. Um I think this this is the pattern we have observed in other sports, and I think it's likely that football follows this as well.
SPEAKER_00Can you tell a little bit more about this? Like what what has been used, what has been done in other sports, and how could you how could that relate to football?
SPEAKER_01Up until now, so June 12th, uh every single sport analytics progress has been the same. Three stages. First stage is can we find better players? Scouting. Okay, is there statistical techniques that enable us to find undervalued players that the traditional observation doesn't see it? So this to exemplify, this was Money Bowl. If you watch the movie Moneyball, it's about this specific part of the data revolution in the game. The second stage is about and and this first stage it's it's solves in football, I'd say. Like there's very little competitive advantage that you can gain on this today in football, at least in high-level football. Today, uh everybody's is more or less okay, able to find inefficiencies in the market. So they are not even inefficiencies anymore. So this is like solved, okay, for me, if you use it smartly, etc. Uh, then the second stage uh is can we play this game better? Is there more efficient ways to play the terminal determined sport? So in Rainbow, this was about uh shifting the the the I don't know a lot, but I know they they started to shift the positioning of the players that were catching the ball. But the main example for this is basketball. Okay, so three point shot. They found inefficiencies, they found efficiencies that were not being explored in the game. And I I'd say football is we are there now, we are starting, and it's it's taking longer than other sports because it's more chaotic, it's more difficult. I think we've found something. I think set pieces is today what we can represent better as being okay, this was the biggest low-hanging fruit that people were not truly exploring. And I think we are reaching okay a level where what Arsenal did, for example, what Michelin has been doing for years and years, what Scandinavia is more oriented than this. Um, so I'd say this is where we are in football today, and then baseball is already very deep on the biomechanics uh player development. So we went from can we find better players, can we play the game better, can we develop better athletes? Baseball has been stabilized on this already, so they have very thorough biomechanics improvements that they make on that they make on players around throwing, around uh uh hitting, setting. Uh so they have found optimal ways to hold the ball, what kind of curve, what kind of angle should your arm be when you are throwing, and the effect that this has on the velocity of the ball, leaving your hand, going to the bat, and the velocity of the ball is obviously correlated with basketball. Uh, I'd say starting to be very advanced on this. So there is even some I don't remember which team is it now, but they basically turned their gym and their courts into closed biomechanics labs. So every action that is being made there, they are able to collect biomechanical data and they are able to improve it. I don't know exactly what's being done around like optimal technique there. Football, we are for sure not there yet. Uh, I know Ajax did some work with this, especially around goalkeeping, where I think the actions are more closed, so it's easier to start. Baseball is obviously very closed actions.
SPEAKER_00This is the difference between the sports, yeah.
SPEAKER_01Exactly. Uh so I think football will take longer, but um I think eventually we'll we'll be able to improve.
SPEAKER_00This is a very interesting discussion related to skill learning. So there is some debate in baseball, also basketball, on how the player's skills actually develop, and football the same thing. So we might get into an understanding what is an optimal way, for example, to hit the ball in certain kinds of pitch or kick the ball in certain kinds of situations. But football that's why I think we are not there yet because it's more complex. There are more skills that we don't even understand yet. Like if we put a dokument of skills in football so often these dokuments actually lack many really important skills and then they can be variable. I remember Moises Caiseda, for example, who was in IDVäno midfielder and really good in duelling, and how he goes into the duell is so variable that and in so different ways that I don't think normally football coaches would understand how this is done and how this is trained. So he learned it from some other place, not from the coaches. And this variability of the skill like how can I do some like how can I achieve the same goal in different ways is so interesting in football that there is a lot of things to improve in this for sure.
SPEAKER_01And football has less standardization around uh tools of of the game, so ball, pitch. I I I was actually hearing, I don't I'm not gonna remember where, but some guys were speaking about the the World Cup ball and or no, the Premier League ball, how it was for the first time a Puma ball, and and they were like, Okay, I've heard from some players at the ball, so and we are seeing more goals from outside the pitch, and then they said that there's a university in England that is doing studies around balls and how balls are affecting uh different so like even this level you can go, okay, and then we go, uh, have players found a good way to hit a specific ball? You know, how does this change? And there's factors like baseball and basketball, you don't have this, you know.
SPEAKER_00The grass can be higher, lower, exactly, can be artificial or normal turf and such things. Yeah, very interesting. Well, let's go into in the final part of the episode: uh, football game, football styles, tactical plurality. So there is, for example, discourse on relationism, positionism. What do you think about this discourse of different playing styles?
SPEAKER_01I think when it started, maybe because I'm Brazilian and I was seeing one of the catalysts there who was Denise. I was especially interested in okay, this is different and this is nice to watch. It's very curious. Uh, I'd say today, it's especially from a professional point of view, is not something that I not that I don't pay attention, but that I value a lot uh one or the other. Like I kind of don't care because I think. It's uh like I like a specific playing style for a variety of reasons. And the in possession dynamics of that playing style, how how is that gonna be achieved? It's not something that I'm particularly interested as long as the output of that of that in possession style that I expect, that I like, not that I expect, that I like, is achieved. So it's gonna be by having a position or a relational dynamics. I I see I think they can be achieved both ways. So it's not something that to be honest, recently is that I have okay given a lot of thought and uh paid a lot of attention. I I'd say now that I'm off the pitch, so I'm not coaching, I take a more macro macro observation around playing style.
SPEAKER_00I understand. From a data perspective, do you have info on which style is currently most efficient globally? Or is that a useful question? It's not really easy to put playing styles into a bracket.
SPEAKER_01Yeah, no, I think that's yes, you have you have directions that but the difficulty is to extract team strength from this player strength. So are the are the styles that are more efficient better because they are better styles, or because they are styles that the best teams with the best players in the world are using? So then we pose a new question: okay, do we need to measure style against expectation? Uh and then it starts to get messier. Then you have to model what is expectation, how much is good enough of an expectation to beat? Uh, is being on par and expectation good enough? And even answering all these questions, when you arrive on this versus expectation analysis, the data has very little correlation with efficiency on anything. So there is basically like examples on all sides.
SPEAKER_00So is this a good news for coaches with different ideas that you can win in different ways?
SPEAKER_01Yes, yes. There are aspects that I don't consider playing style that I consider simply playing better football with set pieces. Like if you are not using well set pieces, you are leaving goals in the table. Period. You know, this is not about playing style away or playing Bs. When you have the set piece opportunity, take the best out of it.
SPEAKER_00So finally, um thinking about my home country, Finland, and current working environment, Estonia, we haven't been selling players with big money and a lot of players out from these countries. So please give us now good advice. How can we do that better?
SPEAKER_01If I knew that, uh Yay. If I had the concrete answer, my my my experience in Finland was very good. My experience in Finland, I think, showed a good side of being in this position of not being a big football nation because there's just so much more room and openness to try new things and to test hypotheses that means when you go to to higher uh nations, uh it's much more conservative. Honestly, my experience in Finland, I thought the players had very good declarative knowledge of football actions, football patterns, tactical patterns, what they had to do, like their declarative uh they they understood very well, and maybe it's because of this environment as well.
SPEAKER_00What would you be looking for, or what would you say that would have to change in maybe Wakehouse League, a Finnish League, Premium League, Estonian League for us to sell players, sell more players?
SPEAKER_01For me it's then there is social cultural and and uh demographic, right? Like the first answer you would give, okay, it's demographic, okay, it's population. Uh but then say you have examples, you have Portugal, you have Uruguay, but then okay, it's population playing football and competition with all these sports. And I think this might be like an angle, you know. Uh I I uh I think hockey is the number one in Finland, for example.
SPEAKER_00In media coverage, yeah. Football is number one in uh in uh amount of players, yeah.
SPEAKER_01But for example, I think football is number one in amount of play in in the US as well. But there is a phenomenon there that when you start going up the chain, even if your base of football players is higher, the top athletes go to sports because there's more coverage, they pay more, etc. I don't know if you have anything, um, but uh it's an angle because like in terms of development, research, uh openness to new ideas, what I've experienced there is to be honest, is much better than Brazil. Much better, like years, years ahead. You know how how we are thinking player development and what kind of cares do we need. But I think that Brazil maybe has natural things that in the same culture, everyone wants to be a football player, everyone, every single, every single boy that is born and girl sometimes wants to be a football player.
SPEAKER_00Yeah, exactly. It's so important in the and in that environment.
SPEAKER_01And it it's also a social proof. So when you are in school in Brazil, if you are good in football, you have many friends, you are a nice guy. So this is a big thing. You want to play in the football. If you are in the football team of your school, you you have friends, you you know.
SPEAKER_00Okay, very interesting. In the future, we hope to have more players from Finland and Estonia also that you could scout, hopefully. But uh until that um thanks thanks a lot for the discussion, Rodrigo. It was very pleasure, it was a big pleasure, very nice insights, and we hope also that Brazil goes far in the World Cup. Thank you, thank you, and Yani.
SPEAKER_01It was my pleasure, it was very nice as always talking to you. Like we want to be here all day.