Gold medals and titles are won not only by the coaching staff and football players. Every professional football club has many employees who also make a huge contribution to victories, but their work is not always visible from the outside.
In this article, we will tell you about the scientific and sports activities of Zenit, which, among other things, helped the club to reach a new level, become the champion of Russia twice in a row, and last season to take the national Cup as well.
Three professionals from Zenit immerse us in details: the head of the club’s analytical department Nikita Vasyukhin , the head of the sports activity analysis department Maxim Gaidovsky and assistant Sergey Semak Alexander Nizelik .
Part 1. Selection. How to find a quality center back, why Barrios is the top for RPL and what is the difference between Argentina and Brazil
In the first two years of work at Zenit, Nikita Vasyukhin worked closer to the coaching staff and was engaged in the analysis of opponents. In the last two seasons, he has changed direction and now leads the analytical department of the breeding department.
In one way or another, all the players that the club considers as reinforcements pass through us, – says Vasyukhin. – Our conclusions become additional information for the people who make the final decisions.
There are two directions of work. First, we are looking for players in all positions – those who, from our point of view, fit Zenit’s style of play, will strengthen the team. These are players of different profiles. When specific requests (positions, profiles) come from the head coach, we offer players who, in our opinion, meet the requirements. The second direction: if several names descend into the department from above, they also need to be analyzed, to give objective information.
What tools do you use when analyzing data?
You have access to various platforms, including Wyscout. Advantage of Wyscout: there are videos from some championships that are not featured anywhere else. Not from the top leagues, but nonetheless. The championships from South America have better recording quality there. Wyscout also promises wide-angle formats (that is, not a TV picture, but a self-filmed technical recording) not only for RPL, but also for FNL – of good quality and for each game. For us it is very important.
When it comes to data, we work with Wyscout even more substantively than ordinary users. Zenit buys an API from them – this is a transcript of each match, in which all actions, their timelines, who performed them and in which zone are recorded one after another. Having an API for each game, you can create more complex metrics – for example, the non-shot xG model, which is very important for us.
Using the API from Wyscout, we can also split the match into possessions the way we want (according to our own algorithms), collect information that is not on the platform. For example, we created “crutches” to handle picks and rebounds there – and we do it more accurately from our point of view. Wyscout has the best API in terms of price / performance ratio, very good coverage, even the minor leagues we are interested in are covered.
Briefly for our readers: what does the API look like?
This is the system to which you connect, take data from it in the format you need. That is, to use them, you need to have a good programmer. Our department has such a specialist, he programs, creates neural networks, administers our platform, and interprets data.
You mentioned the non-shot model. Why is it so important to you?
It allows you to evaluate the players’ actions in terms of how they increase or decrease the team’s likelihood to score inside possession. And, accordingly, the likelihood of an opponent to score after this possession is evaluated in the same way.
The metric was created in order to better evaluate the players inside the attack. For example, center-backs rarely directly assist when hitting. But their transmissions are always important at the start of an attack. For creative players, the indicator can also be important, because often a player can give a pre-goal pass, which is not directly taken into account in the xG model. Or a footballer can often deliver the ball to good areas, but his partners are not able to convert a pass into a shot. And it won’t show up in xG either.
That is, using the non-shot xG model, we exclusively evaluate passes and “wounds” with the ball (dribbling, dribbling) in terms of how this action increases or decreases the chances to score or concede in this attack. Roughly speaking, non-shot is the mathematical expectation of the goal difference between possession and next possession.
We have developed this metric ourselves based on data from Wyscout. We make a framework that allows us to break all matches into ownership in our way. Each analyst firm may have a different perspective on what possession is when the other team passes possession. If in basketball it is clear which team is in possession of the ball, then in football there is no such thing, it has more transitions. We analyze the possession match as we feel it is right and evaluate how passes and injuries affect the situation.
Either the team strikes a dangerous blow, or misses a dangerous attack?
The strikes are not important here. The model is made in order to get away from them and create another metric. Rather, it is about how the team moves the ball into the danger zones and how effective the players are in this. And how well it defends its space. With non-shot, you highlight players who often lose the ball near their goal, which is bad. The metric allows you to rate players well by the ratio of how risky they play and how beneficial it is. For example, if a player sharpens a lot, then it is obvious that the accuracy of his passes will be much lower than that of a partner who does not sharpen. But at the same time, the effectiveness of his passes is high, even if one out of five passes (one out of five takes partners one on one), this is a very good indicator.
Non-shot is a complex metric. Evaluation of each action in shares of a goal. Then you can break it down into what actions the non-shot earns. It is popular among Western analysts to make a generally complex indicator that evaluates all actions in terms of the probability of scoring / missing: defensive, horseback fighting, and so on. But for us it is important to evaluate simply the progress of the ball with the help of passes and running. So we built the model this way.
Let’s say you need to find a quality center-back. What data will you analyze first of all on it?
First, it all depends on the profiles. You need to understand that a defender who is fast at the same time, works well with the ball, plays well one on one, strong in martial arts and at the same time not very young, that is, having gained experience – he costs huge money. Therefore, when you are given a profile, people donate something. For example, a footballer may be very good on the ball, but weaker at riding martial arts than the average high-level center-back.
For evaluating a ball player, non-shot is just very useful, we always watch it. As far as defense is concerned, it is much worse in numbers. The problem with studying center-backs is that the number of tackles and steals doesn’t tell us about their level at all. Exclusively – about the style of play. In my observation, center-backs who have few tackles and interceptions are often better and better than those who have a lot of these actions. As examples – van Dijk or Ivanovich, they rarely go to the selection, rarely intercept. They play more positionally, block the space, force the opponent to give passes to other zones. And raw data does not estimate it.
It is important for us to look at how good a player is at riding martial arts, this is a fairly objective indicator. How a football player plays heads-up and how many fouls he commits. Good center-backs tend to commit few fouls, and they know how to defend without contact. For example, the same Ivanovich, he has fewer fouls in two games on average, this is a very good indicator.
Naturally, we are trying to analyze the positional defense of space. We divide the field into zones, determine which of them the defender is responsible for (based on where he makes most of the touches), and see what the accuracy of the passes inside his zone, what proportion of the passes are transformed into shots within a few 10 seconds – also important for us indicator. We look at how the non-shot teams of the opponents are “spread” on the whole, how effectively the opponents move to the goal through its zone.
Although when studying central defenders, video viewing is much more important than all the statistics. Things like dynamics, speed, the ability to defend space, the level of decision-making – all this can be assessed only visually. Therefore, we are looking for wide-angle cameras so that the player is always in the frame and we can evaluate what decisions he makes when passing behind his back, how he unfolds the body, how he moves in general, what kind of dynamics he has, and so on.
Accuracy of transmissions within a specific zone, number of hits in 10 seconds – is this data also provided by Wyscout?
We calculate based on their API. A comparison can be made: API – as grain suppliers. Wyscout provides very good grain, but we bake the bread ourselves. That is, if you use their platform, go to the site, then you eat their bread. The bread is good – their platform also has almost all the data you need.
But we take the grain from them, and we bake the bread ourselves. This is good for us, because we can add raisins, coriander or something else to this bread. Today such bread, and tomorrow another. But the main thing is that the quality of the grain always remains excellent. Then we’ll figure out how to use it.
Let’s move on to examples. Lovren is the last high-profile transfer of Zenit. How was it studied? How were his famous mistakes analyzed?
Errors taken out of context a little. I’ve studied a lot of Lovren’s matches over the past year, including wide angle recordings, that’s more revealing. The player was analyzed according to the classic format, according to all the data that I listed: non-shot, horse martial arts, single combats below, space defense – all this was important. Usually the conclusion consists of a statistical report and a video of the player with comments, where his typical actions, pros and cons are analyzed. All this was given to the management. The management, based, among other things, on the report, but also on other factors (financial and medical components, the situation with a small gap between championships), decided to take it. And the main thing is that the coach likes the player. Semak praises him.
What kind of non-shot did Lovren have?
In the 2017/2018 season, Dejan became one of the best central defenders in the English championship in terms of this indicator. Then he started to play less. If we take a sample for the last two seasons, then his non-shot is in the region of the median. While it is important to understand that Liverpool’s style has also changed, van Dijk has come, who starts attacks much more often. Lovren was less involved in this stage (van Dijk took most of the promotion passes), made fewer difficult passes, and in general the team more often promoted the ball over the full backs.
If Lovren played a right center back, then he mostly gave simple passes to Alexander-Arnold, and Trent himself promoted along the flank. Accordingly, the passes to the neighbor on the flank do not bring real benefit – that’s why Lovren had modest numbers. But we understood that the player’s non-shot could not drop so sharply as a whole for reasons depending on him. This was mainly due to the change in Liverpool’s style.
What was your verdict?
That the player is still good on the ball. Visually, this is confirmed, Lovren has a good first touch, he can move the ball, perform different types of passes: penetrating between the lines, long throw. In almost every match we see these actions. The first goal that Zenit scored in the game against CSKA came after a good throw from Lovren. Plus, in the same game, there was another good attack, where he gave between the lines to Dziuba, and Dziuba threw his heel onto Ozdoev. If the Maga had not slipped, the situation would have been dangerous.
Lovren is still good on the ball. As for the games at Liverpool, questions may arise about his actions without the ball, but, as I said above, everyone has some disadvantages.
Which of the current football players of the club did your department find?
In its current form, the department has existed for only the second season. All transfers were made earlier, when I worked more with the coach, and with him we watched almost all the players. Barrios, Douglas Santos, Karavaev – this is a collective merit.
We started working with Barrios and Douglas back in August 2018. The Colombian joined in February 2019, and Douglas in July 2019.
How was the study of Barrios?
At the time when we watched it, we did not have such a volume and quality of data. But we understood the profile well: we needed a dynamic defensive midfielder with good tackle and good positional defense.
We looked at Barrios’ tackle volume, adjusted for possession, and the quality of the heads-up game, at how well he controls the area. “Boca Juniors” under him defended the support zone well, allowed low accuracy of passes inside it, forced opponents to attack through the flanks. Barrios also had very good indicators in rebounds – and it was important for us to strengthen the middle in terms of rebounds, in our own penalty area in the first place. Wilmar added all the described qualities to us. Initially, when we compared him with other options on the radar, we understood that this was the very player we needed. The player who controls the pivot zone and takes most of the pressure off center-backs.
If we talk about the current Barrios, it is clear that from a defensive point of view he is the best defensive midfielder in the league, while his qualities with the ball are greatly underestimated. Yes, Wilmar does not give away a lot of complex passes with low expected accuracy: penetrating, behind the back, and so on. But he is very good under pressure, works great with the body, he has a good first touch, almost does not lose the ball under pressure. For example, Barrios gets fouled twice as often as he breaks the rules. Although for a defensive player who often enters the selection, this is not entirely typical.
Due to his qualities, the team, if it is under pressure, can leave the defense – he will deploy the attack or earn a foul on himself. Barrios has an excellent non-shot – not due to the fact that he promotes the ball, but due to the fact that he almost does not lose. He plays even with simple passes, but confidently brings the ball to his partners who are able to move it into the dangerous zone.
There were many candidates for the left-back position a year ago. Why did you choose Douglas Santos?
The work on the left-back was really great. Douglas Santos matched the coach’s profile in terms of the quality of the passes in the attack and in the final stages. That is, he has good passes that develop attack, and a quality cross. And also – a large volume in defense, he aggressively enters the selection.
Douglas had one of the best non-shots in the second Bundesliga. Although we noticed him when he performed in the first. When “Hamburg” took off, they continued to follow – and took it when the opportunity arose. Hamburg’s style highlighted the player’s strengths as the team spends a lot of time on the ball. Douglas both played wide on the flank and moved into the supporting zone, played a false flank defender – and there, too, he looked great in the development of the attack, acted aggressively in the selection. All this gave us a hint that Douglas could play in the support zone and become a generalist.
What primary statistical filter was used when looking for flank defenders to narrow the circle?
As I said, the filters are implemented based on Wyscout data, but on their own platform. We search by our own metrics. The footballer’s profile meant high-quality ball play in the development and completion of attacks, a high-quality pass was required. Accordingly, we paid attention to the expected assists xA, to the number of accurate passes into the penalty area, to the non-shot pass. And also on the volume of defensive actions and the quality of one-on-one play.
Are the last two stats for flank defenders not an indicator of style, as for center ones?
No, for the flanks they also talk about the level of a football player. For a full-back, tackle volume is important. There are special studies: how much the volume of defensive actions of the players of a particular position affects the number of moments allowed by the team, is there a correlation. For center-backs, the amount of defensive action (adjusted for ball possession using a special formula) does not correlate with the number of chances. Wing-backs and holding midfielders have such a correlation – and it is significant.
Flank defenders must be able to immediately meet the opponent aggressively on the flank, win one-on-one at least 70 percent of situations. This is important for us.
How do you calculate the adaptation risks when players move from other championships – weaker or seriously different from a stylistic point of view?
For example, let’s take a player from the attacking group. It is important to look at how he gets his chances and how it fits in with the style in which he will play in the new team. If the attacker gets most of the chances with the help of passes behind his back and runs out in space, and in our country he will have to play a lot with his back to the goal and in a saturated penalty area, then this may serve as a certain signal that such a player is not suitable. But you need to be able to separate his qualities from the style in which he plays. Even if he opens up a lot behind his back in his club, then you need to study whether he has any qualities at all that will allow him to play in new realities.
Next, it is important to analyze data in matches against strong opponents. If we are talking about the Netherlands championship, then you watch how a player plays against the “big four” – Ajax, PSV, AZ, Feyenoord. If he plays for one of these four teams, you study what he looks like in European competition. Also in the Netherlands, there are many football players with a large amount of strokes. It is necessary to look at the structure of dribbling: is it dribbling in tight spaces or throwing the ball past opponents and jerking at speed into zones? If the second, then you understand that in Russia this style will be worse, we have less space and more safety net.
We pay great attention to getting out of successful strokes. Often the players have a lot of dribbles, but the next actions have a high percentage of losses: inaccurate passes or further losses on the dribble. We are looking at what the player converts his stroke to.
Which style championship is more similar to the RPL?
If from the top 5, then France, there are also many martial arts, the emphasis on wrestling in many clubs. The Czech Republic is very, very similar, the players from the Czech league will adapt well here. More Scandinavia: Sweden, Norway, Denmark. To a lesser extent Portugal, although there are similarities too.
How about Brazil and Argentina?
Brazil has its own nuances. The pace and intensity are low there, few teams put pressure on. Is that Flamengo and the teams under the leadership of Sampaoli – last season, Santos, this year – Atletico Mineiro. For example, we can cite the following statistics: in the Brazilian championship, all but two clubs have passing accuracy higher than 80 percent. And in Argentina, on the contrary: all but two have below 80. This does not mean that Brazil has better players. It’s just that in Argentina, the pace and intensity are higher, where all teams put pressure and play aggressively after losses, they act better in transitions. That is, the share of transmissions made under pressure is higher there.
There are other specificities in Brazil. There is no full-fledged pre-season training, people instead play in the state championships. Therefore, when a footballer moves from Brazil to Europe, he needs time to adapt. A textbook example: Artur, who was hired by a fitness coach at Barcelona. The first season and a half, he almost did not play full matches, because he could not stand it.
And people come from Argentina – and it’s easier for them to adapt. The same example of Barrios – he was used to playing at a high pace with a lot of martial arts, he kept a high pace. Although the paradox: in the Copa Libertadores, Brazilian clubs usually show better results.
What else should be considered when studying players from the Brazilian Championship?
It is rare to find stories when a footballer over 25 moves from Brazil to Europe. After a certain age, Brazilians already it’s hard to change the approach to training and become stronger physically. From recent seasons, one can only recall defender Felipe, who moved to Porto in 2016 at the age of 27. He is now at Atlético. And so people leave before 25. Then the risky age begins.
Wide-angle cameras, which are becoming more popular, can over time replace live views, trips?
Not yet. They are hard to get, plus there are people with old habits who need to watch live. When you ride, you look not only at the qualities of a player, you also try to communicate with someone, find additional information: how he behaves outside the field, how he trains. This is all very important too.
But wide-angle recordings can really replace what you see in the stadium. The player gets into the frame all the time, even warm-ups are often filmed. How does he warm up before the match? How good is his attitude? Sometimes a football player in the warm-up can do what he has the opportunity to do in the game. For example, many people hit more than 90 minutes of play in these 20 minutes of warm-up.
Wide angle recordings are an undervalued market. Hopefully, over time, there will be more and more video from such cameras.
Which championships are you watching right now?
Wyscout gives an API for all the leagues they have. Our models process the data for the championships that we need, and we upload this to the platform. We have several dozen championships. Players from the top 5 are in principle already in sight, we all know them.
Personally, I primarily pay attention to the following championships: Portugal, Belgium, the Netherlands, Austria, Argentina, the underestimated place of Uruguay, to a lesser extent the Scandinavian countries and Turkey. We are also studying Brazil, but the market there is a little overheated. Although now the situation in the world is such that there are no countries that should be completely ignored. Good football players with the highest potential appear everywhere – in Colombia, Canada, Georgia, Uzbekistan.
Due to the existing algorithms, we made it so that any player who shows himself immediately gets on the radar. Even if a young guy appears in a league, spends 500 minutes, and his performance is above a certain mark, he is immediately offered by our system. And we start to watch it. The likelihood of missing someone is reduced.
Part 2. What an analytical report looks like about an opponent, what data it must contain, and how to build interaction during a match
Maxim Gaidovsky, together with assistants, is engaged in analytical support of the main team. His department also analyzes fitness data.
How is the interaction with the coaches organized in the weekly cycle in preparation for the opponent? – the first question for Maxim.
If we spent the previous round away, the head coach can get information about the next opponent already on the plane. We usually return the same day. If the opponent manages to play one more game in the same round, then we simply add a couple of pages in the next days.
The opposing team’s presentation is a 10-page report, it includes the estimated squad, information about the last matches, about the team’s players, about the structure of the game in different phases. The last couple of pages are devoted to just the analysis of statistical data. Plus, we are preparing two videos: on the standards and organization of the game. Both videos are divided into subcategories.
In the standards there are all subspecies: attack and defense, corners, direct free throws, free throws from the flanks, sometimes we add outs and start from the center of the field. Many analysts in other clubs also refer to the standards as starting from the goal. But we refer this rather to the organization of the game.
How do you break the organization of the game?
As a rule, we show a series of episodes for each phase of the game: high block, medium block, low block, transition to attack, start of attack , development, completion and transition to defense. But there are teams in which episodes of one or two phases may be absent. If, for example, a team is using high pressure, it will be difficult to find defensive moments in the middle block with its participation. Because they either take the ball away, or they cut off the pressure group, and they have to sit down immediately in a low block.
An important point: the head coach himself decides which episodes will be shown to the team. After all, he has to convey his ideas to the players. We adjust the presentation depending on what he wants to draw attention to.
Key statistical numbers that will always be in your reports on the opponent?
It all depends on the team, but as an introduction there is information about the height of the middle line of defensive actions – absolute value and indicator relative to the league.
That is, it can be 41 meters from its goal – the 14th indicator in the league, where 47 meters is the highest line.
How is this height calculated?
We interact with Nikita Vasyukhin, he has an internal resource in his department, where analysts process raw data that we purchase from statistical companies, and provide complex metrics. In addition to the line height, we also give the PPDA (passes per defensive action) indicator – the average number of passes that the team allows the opponent to make per one defensive action.
We also study the percentage of returned possessions in the first seconds after the loss. All this indirectly shows the intensity of the team’s pressure. Combining the three indicators, we can assume how far the pressing team will resist us. But we never focus on data alone. We always give our assessment after watching videos.
What platform do you personally use most often when studying data?
I believe that at any time you may need each of them – and it is desirable that all platforms be at hand. They compete with each other in details: some have wider league coverage, some have better shooting, some have increased efficiency.
Wyscout is certainly a valuable resource of information and videos for us too. They are trying to introduce new metrics, added xG to their reports, and introduced new sections on the platform. But personally, I don’t work with data. There is Nikita Vasyukhin, in my department there are guys with specialized education. The most important thing is that they understand how the data they use is collected and interpret it correctly. If the differences.
Let’s take the most common indicator – ball possession. Some companies consider it according to the Stop watch technology, that is, like a chess clock. One team owns the ball – he pushes one button, the other – another. If the ball is neutral, time continues to run. If the ball goes out-of-bounds, who owns it? Opta, for example, calculates ownership by a formula, by the ratio of the number of transfers. There will also be insufficiently accurate numbers if one team plays long passes and the other plays short. For the second team, the ball moves faster between the players. And a team using long passes will have a lower ownership percentage on that count than it actually is.
There is a lot of discussion about the xG model right now. Do you use this metric?
Yes – and quite often. We have our own xG model, by which we evaluate the quality and quantity of chances created in matches. I know that now many people criticize this indicator. The xG model is imperfect, it has many drawbacks – for example, it takes into account chances only in the case of a strike. If there was no hit, then this moment is not reflected in any way in xG. Also, the model does not take into account the level of the performer.
But if you look at the impact indicators when analyzing the data, then it is foolish not to pay attention not only to the quantity, but also to the quality, which is calculated according to this metric. No one would argue that xG describes a team’s probability of scoring better than just the number of shots. After all, you look at the number of shots on target or off target in TV broadcasts? Do you use them? Are you skeptical about this data?
I think xG is one of the most convenient quantitative metrics that are currently available to us. There are even more complex metrics – the same non-shot. Another question is that football people, coaches, are not used to talking about football in such categories. Although I am sure that if you correctly explain to them what exactly xG shows, then they will understand. Over the course of a year of working together, our coaches have become accustomed to this metric. Yes, they also constantly criticize her, but we do not focus only on her. So when we provide data about an opponent, we include xG and xGagainst in every match the opponent has played. And this more or less demonstrates the objectivity of the results.
Is it realistic to immediately identify the opponent’s weaknesses using statistics? Let’s say you saw that he missed eight goals out of 10 through his right flank. What will you do with this information?
In general, it is real. You can identify certain patterns if you have the skills to work with data. Even in your example, the data highlights a certain fact for us. But can we immediately say that the right flank of the defense is the opponent’s weak point? First you need to study who played in this zone at all – one player or each match is different. Next, you need to look at what type these scoring attacks were.
Let’s say the player was the same.
It is necessary to analyze all the mistakes he made, whether they were typical or situational. It is important to understand that a goal is partly a random event in football. It’s safer to focus on the quantity and quality of scoring chances: how they were distributed along the two flanks and the center. Then a wider sample will appear, the probability that this is a random distribution will be lower.
Your reports always contain data, through which zones were moments created?
Of course. In general, we have a lot of data on zones: the number of transfers inside the zones, the accuracy of these transfers relative to the league, the number of tackles … But this is redundant information. Previously, we just published it in the reports: we drew the field and painted over the zones with different shades. Blue is below league average, red is above average. But then, nevertheless, they came to the conclusion that they began to give these diagrams only if regularities were traced in them. It doesn’t always make sense to publish field layouts just like that.
To summarize: if you have statistics that eight goals out of 10 missed one flank, you still need to study all these goals and all the dangerous moments, find out and explain what is the weakness of this flank of the opponent’s defense. As a result, it will be possible to say that the statistics helped us to identify the weaknesses of the opponent. But analysis implies comprehension, not a bare enumeration of facts.
How do you calculate matches and break them down into the fragments you need?
Using the Hudl Sportscode program. This is a typical self-coding tool, that is, software that allows you to code matches yourself. To simplify terminology, encoding is the breaking down of a video into episodes. On the statistical platforms, the matches have already been coded by analysts, all corners, tackles, shots, crossings, and so on are marked there, everything is publicly available. Self-coding tools are a tool in the arsenal of the club’s in-house analyst, with which he can watch the match and independently code what he needs. Requirements and approaches are different everywhere.
These codes can then be digitized to generate a statistical report. Or you can make sure that you always have the episodes that you code at hand. Let’s say there is a certain moment that you notice all the time in matches. If you coded it in every match where you came across it, then you can easily get all these episodes and show them in one video.
Hudl has an online platform. You can enter users there – football players, coaches – and share videos with them.
Let’s say the opponent uses two different schemes – sometimes 5-3-2, sometimes 4-2-3-1. How does your report look in this case?
Such teams are not uncommon in modern football. Moreover, there are teams that change their structure not even from match to match, but during the game. The so-called “transformer” schemes. Pay attention to Leipzig, which can defend in 4-4-2 and attack 3-4-3.
In such cases, and when the schemes change from match to match, we just make two videos in one. In the game with these schemes, within each phase there are nuances that can be paid attention to. In each phase, we show separately a number of episodes in one scheme and a number of episodes in another. The video lasts a little longer than the standard one, and here animation and graphic processing come to the rescue.
Animation is a great time saver for football players. If you take three episodes and do not draw anything on the video, then the effect will be the same as taking one episode and indicating with the help of graphics what to pay attention to.
How do you interact with the coaching staff during the break?
During the match, we use Hudl Sportscode. This program has a video capture function. Our analyst is filming the match on camera from the upper tier so that all the players can be captured by the lens. During a match, I am close to the bench, receive a signal from the analyst, capture a video in Sportscode and immediately encode it. It is necessary to choose the points that we discussed during the preparation for the opponent and during the installation. Something that can be quickly shown to the team during the break.
When the whistle blows for a break, I go to the coaching room. There, as a rule, you have time to show episodes of five to seven maximum. The coach tells you which ones to put on the screen in the dressing room. Thus, the coach can adjust the actions of the players during the break, if necessary.
Part 3. How the coaching staff perceives and uses information from analysts during training
Alexander Nizelik has been working at Zenit for many years. Back in 2010, shortly after the arrival of Luciano Spalletti, he got a job at the club as a translator. In the future, Alexander managed to get a coaching education, work at the academy of blue-white-blue, travel with Sergei Semak to Ufa, after which he returned with him to the club where he spent most of his career.
In an interview with “Championship” Nizelik told how the training and organizational processes are being built in “Zenith”.
The analytical department has sent you a report on the opponent. What numbers do you study in it first?
The report is not only statistics, we study it in its entirety. There are very informative sections, which tell about the last matches of the opposing team, the details of the game, news are provided that make it possible to predict the composition for the game.
With regard to statistical data, we look at how the opponent attacks, and in which zones he does it more successfully, in which zones he defends better, and in which – worse. There are a large number of indicators. Of course, we focus not only on the report. Without fail, we watch the games ourselves. Often what we see when watching the game in its entirety is confirmed by data from analysts’ reports. Sometimes something doesn’t add up and you try to pay more attention to some specific points.
Do you yourself somehow come into contact with advanced statistics?
We have a large, organized and strong analytical department. Analysts work with platforms and advanced statistics and provide us with a finished product. We ourselves also have access to the platforms and use them as needed: watch a match or study a possible beginner in the selection process. Wyscout, for example, has a metric called one-on-one defense that is not found on other platforms. We resort to it when we study the defenders.
How many Zenit coaches watch the matches of the future opponent?
Everything. First of all, the head coach. We, assistants, try to help and also reveal some details. Teamwork is important here, everyone can notice some nuance. The main thing is that Sergei Bogdanovich ultimately has the maximum amount of information.
– How many matches are taken for analysis? All the latter, or just a few, against opponents with a certain style of play?
– Our analysts have already worked out a scheme – they are watching the most detailed three last matches of the opponent. Plus, if we played with this team relatively recently, then we definitely analyze our match, it can become even more informative. Of course, the opponents’ games against teams that use a game model similar to ours are also interesting.
How many theories are held at Zenit per week?
Depends on the length of the cycle and the amount of time to prepare. If we have a full week at our disposal, then at the very beginning we analyze our last match, and at the end of the week, on the eve of a new game, we always make a theory about the opponent. Plus, on the day of the match, we conduct a short lesson according to standards.
So you show your opponent right before the game? Not at the beginning of preparation?
Training work, of course, is built from the very beginning depending on what our game plan will be. In the course of the week, we can modify the work, convey information to the players on the field, in individual conversations. Often we make small video cuts, literally a few moments from training exercises, which show as clearly as possible what we want to achieve in certain game situations, and what, on the contrary, we want to avoid. These videos are shown on screens in the dressing room, and the guys can watch them before and after training.
Specifically, we hold a theoretical lesson with video on the eve of the match.
What is the maximum duration of one?
Varies, but on average does not exceed 15-20 minutes. Do not need anymore. Otherwise, concentration of attention is lost, information is absorbed worse.
Let’s say you play with CSKA in a week. You know that the opponents are pressing high and playing with three central defenders (take the team from last season). What will the weekly cycle look like?
In any case, it is built on the basis of our game principles. If we clearly understand the opponent’s model, then we simply try on certain days of the weekly cycle to simulate specific game situations on which we want to focus.
How do you achieve maximum perception of information in theoretical lessons? How to make the players not only see the video, but also understand everything at once?
It is important to understand that theoretical lessons are only a part of preparation. Do not expect that after watching the video (no matter how high-quality and detailed it may be), the player will enter the field fully prepared for any situation. Players absorb a large amount of information on the field during tactical and playing drills. This is where the main processes of perception take place.
Of course, we have all sorts of graphic tools that highlight the nuances of the video. Also, in our theories, we divide moments into phases so that the players get used to seeing the game in the same way as the coaching staff. Then the transfer of the information obtained on the video into training work and into the game goes better.
Are the reports on your matches and information about the opponents sent to the players?
No, they are used by the coaching staff for preparation. And the players are given information during the training process. Football players can ask for a video to analyze their own actions, and here we always meet them halfway. Also, the analytical department prepares individual videos of the opponent’s players. This is interesting and useful for the children when preparing for the game.
What is the difficulty of preparing for an opponent who changes his schemes? How do you prepare in such cases?
We are trying to identify the basic model of the game, it is still, as a rule, visible. But even if the scheme changes (usually there are two options), we must understand how to act in this case. As for our game, if the team knows its principles, then it can adapt on the field to any model of the opponent.
You have received a report on your opponent, you clearly know all the information about him. How is the game plan adopted?
The head coach makes all the key decisions, in all areas of our work – the final word is his. The headquarters is helping him. That’s why we are called helpers. Of course, there can be a discussion, we have a good coaching team.
As an example, let’s take last year’s match with CSKA (4: 0). You then surprised with a 4-3-1-2 formation with Malcolm, who was positioned under Dziuba and Azmun. Plus, the team came out of defense only with long passes, previously luring the opponent into pressing. It turned out cool! How did you come up with this solution?
There was nothing supernatural in it, it was dictated by the opponent’s style. CSKA, especially when playing at home, likes to put high pressure. The team deliberately tries to gain an advantage in the zone where the opponent’s attack starts, to create one-on-one situations at the enemy’s goal, even if they risk with one-on-one situations from behind. Dziuba and Azmun play in front of us, they fight well at the top, and Malcolm in the position below them due to his speed and mobility can play well on rebounds. We decided that this situation is beneficial for us.
If the opponent leaves behind one more player, respectively, he presses with fewer players. In this case, you can start attacks with a short pass, having a numerical advantage in the outfield player plus the goalkeeper. If the opponent, when pressing, risks playing one-on-one, does not give you the opportunity to start attacks from the bottom without pressure, of course, it is more profitable for us to deliver the ball forward as quickly as possible, where our attackers can surpass the opponent in single combats. This is a fairly logical decision.
The rights to the video belong to the National Sports TV Channel LLC. You can watch the video in Rutube “Match TV”.
Many admired your winning standard in the match against CSKA in the new season. Did you play that original prank?
In general, about the standards it is better to contact Anatoly Tymoshchuk, he devotes a lot of time to their preparation and does it efficiently. Let me just say that we have many options for the draws. And the best thing about that moment is that the guys acted great according to the situation.
Was there some improvisation there?
Yes, competent decision-making on the game moment. If you remember, then Vlašić was injured, CSKA had one less player on the field, and it became possible to play the ball. The skill of the players and the ability to quickly assess the situation came to the fore. The guys did it, played “3 v 2”, Karavaev ran into the penalty area on time, Douglas gave a timely pass, Azmun opened up great.
You will understand that in football, the schemes work up to a certain point. In the standards, by the way, they can still work, in them you can act as you want, because you have the possibility of the first move, you control your opponent and your actions. Otherwise, high-level footballers are guided by the situation and react to any changes, make cool decisions, as against CSKA. Our task is to make sure that, thanks to their work, the players make more correct decisions.
Schemes don’t work in the game either?
Our vision is that the principles of the game, not the schemes, work. We are working more on the principles, on the moments thanks to which you improve the quality of decision-making. The scheme is good when you are playing and there is no opponent on the field. If there is an opponent, more often than not he will not allow you to fulfill your plans. Then you need to make decisions based on the current new situation, and this can be done only if you rely on the principles of the game.