UmemotoRikuhei/w=48,quality=90,fit=scale-down){kind=small}
Atom ScottIntroduction
Hello everyone! This is Atom from Playbox.
In Parts 1 and 2 of this series, we explored how to acquire data and analyse actions on the ball. This time, we’re shifting our perspective entirely to focus on off-the-ball movement – in other words, what happens when a player doesn’t have possession.
The average amount of time a player spends on the ball during a match is around 3 minutes.— Johan Cruyff
Even with a simple calculation – dividing 90 minutes by 22 players – that works out to about 4 minutes per player. If we look only at actual playing time (roughly 60 minutes), it drops to under 3 minutes. Either way, it's clear that players spend the vast majority of the match without the ball.
How a player moves and behaves without the ball has a massive impact on the quality of attacking play. That’s why, in this final instalment of the series, we’ll be exploring OBSO (Off-Ball Scoring Opportunities) – a method for analysing attacking potential when players are not in possession.
This article was written by my junior from Nagoya University, Rikuhei Umemoto. He’s one of Japan’s leading football data researchers and the first Japanese presenter at the StatsBomb Conference.
This marks the final entry in our “Behind the Scenes of Football Analysis” series!
- Behind the Scenes of Football Analytics: Part 1 – How to Extract Match Data from Broadcast Footage
- Behind the Scenes of Football Analytics: Part 2 – Analysing the Ball Carrier (xG, VAEP)
- Behind the Scenes of Football Analytics: Part 3 – Analysing Off-Ball Attacking Patterns (OBSO)
Visualising Off-Ball Value Through Broadcast Footage
When it comes to tactical analysis in football, it’s common practice to use tracking data collected through specialised systems installed in stadiums. However, in reality, access to official tracking data is often limited, and the costs involved can be prohibitively high.
That’s where broadcast footage comes into play as a valuable alternative for data extraction.
In player analysis, it’s also crucial to evaluate not just those on the ball, but arguably even more so, those off the ball. Take a look at the following example to see why off-ball movement can be just as important.
Some of you might recall this goal—it’s still fresh in the minds of many football fans. Of course, the finish by No. 11, Mohamed Salah, was superb. But what we’d like to focus on here are the attacking players off the ball.
While Salah was in possession, imagine if the ball had instead been played to No. 7, Luis Díaz, on the far side, or to No. 10, Alexis Mac Allister, who was positioned centrally in front of goal. Had either of them received the ball and taken a shot at that moment, how likely would they have been to score?
In fact, at 0:09—just as Salah takes his shot—it appears that both Díaz and Mac Allister were in more open space, possibly in better positions to finish. This might lead some to wonder: Would passing have been the better option?
Being able to evaluate the positioning and potential of off-the-ball attackers like this could offer valuable insights—not only for improving future decision-making on the pitch, but also for enhancing the viewing experience from a fan’s perspective.
That’s why in this article, we’ll introduce a method that involves estimating pitch coordinates from broadcast footage to track player and ball positions, and then using a metric called OBSO (Off-Ball Scoring Opportunities) to visualise attacking patterns and evaluate the contribution of off-ball players.
Let’s break down how it all works.
What Is OBSO?
OBSO (Off-Ball Scoring Opportunities) is a metric developed by William Spearman, the current Chief Data Scientist at Premier League giants Liverpool FC.
What makes OBSO unique is that it estimates not only the scoring probability for the player on the ball, but also for off-ball attackers, based on the next potential event. In other words, it provides a probabilistic model that quantifies scoring opportunities for all attacking players on the pitch at a given moment.
OBSO quantifies the likelihood of each attacking player on the pitch scoring in the next phase of play.
OBSO is calculated by combining three key elements:
- Scoring Probability (Expected Goals) – the chance of scoring from a particular location on the pitch
- Occupancy Probability – the likelihood of a player receiving the ball in that location
- Transition Probability – the chance that a pass is successfully made to each location
By using this metric, we can evaluate the success probability of a pass and the scoring potential, based on player positioning, defensive pressure, and available space — all in an integrated manner.
For example, visualising OBSO through heatmaps, time series graphs, or player-specific tables enables detailed insights such as:
- Comparing Passing Options: Which teammate offered the highest expected scoring value at the moment the pass was made?
- Evaluating Dribble Impact: How did a ball-carrier’s dribble draw defenders and increase the OBSO of nearby teammates?
- Assessing Shot Decisions: Was there a teammate with a higher OBSO than the shooter at the moment of the shot?
This kind of analysis provides a quantitative framework for evaluating decisions like passing, dribbling, or shooting — with a clear view of how each choice affects the off-ball players’ scoring potential.
Because OBSO is built from the product of three interpretable probabilities, it remains not only mathematically meaningful but also easy to understand, making it a highly practical tool for future football analysis.
Example Analysis Using OBSO
In this section, I’ll walk you through an example of how OBSO can be applied to real match footage — from the broadcast video all the way to evaluating off-ball scoring opportunities.
As Japan have just secured their spot in the FIFA World Cup 2026, I’ve chosen four key moments from the AFC Final Qualifying Round to demonstrate the practical use of OBSO in match analysis.
Let’s dive straight into it!
AFC Final Qualifier – Japan vs Australia: Japan's Goal
This moment came from Nakamura’s skilful solo effort, which ultimately led to an own goal and earned Japan a crucial point.
In this analysis, we’ll focus on two key questions:
- Was Nakamura’s dribble leading up to the final pass effective in terms of creating scoring potential?
- How well did the off-ball movements of other players support the play before the final pass?
To answer the first question, let’s take a look at how Nakamura’s OBSO value changed during his dribble.
Figure 1. Change in Nakamura’s OBSO Value from the Start of His Dribble to the Final Pass
As shown in Figure 1, Nakamura’s OBSO value rose sharply after his second successful take-on.
As explained earlier, OBSO represents the expected scoring value at a specific location on the pitch for the next on-ball event. Between his first and second attempts, Nakamura operated in a wide but relatively low-threat area — near the left touchline and far from goal — so his OBSO value remained mostly unchanged.
However, after beating the defender on his second attempt and continuing to carry the ball forward, there was a clear increase in scoring potential. His controlled progression towards goal after getting past the defender also contributed to the steep rise in OBSO.
This analysis suggests that, from an OBSO perspective, Nakamura’s dribble was indeed effective in creating a high-value opportunity.
In addition, Nakamura’s dribble also served as a deliberate pause in tempo, allowing teammates to advance and position themselves for the next phase of play.
But did his teammates take advantage of that moment to exploit the available space?
To explore this, let’s focus on three players who entered the penalty area just as Nakamura played the final pass: Ueda, Kamada, and Tanaka
Figure 2 Changes in OBSO values for players No. 9 (Ueda), No. 13 (Nakamura), No. 15 (Kamada), and No. 17 (Tanaka) during Nakamura’s dribble leading up to his final pass.
In this chart, No. 9 represents Ueda, No. 13 is Nakamura, No. 15 is Kamada, and No. 17 is Tanaka. All three off-the-ball players show a clear increase in their OBSO values, suggesting that Nakamura’s decision to hold the ball and create time was effective in boosting their scoring potential.
Let’s take a closer look at each of these players:
Ueda (No. 9)
Throughout the sequence, Ueda maintained relatively high OBSO values compared to the others — even before Nakamura’s second successful dribble past the defender. This is likely due to his central positioning as the No. 9, constantly occupying areas in front of goal. After Nakamura’s dribble success, Ueda’s OBSO briefly dipped before rising again. This fluctuation coincides with a sharp run Ueda made behind the Australian defender. By attacking the space behind the back line, he effectively increased his own scoring potential — a strong example of intelligent off-the-ball movement.
Kamada (No. 15)
Kamada’s OBSO value steadily rose over the course of the sequence, with a particularly sharp increase just before Nakamura’s final pass. He had already found himself in space before Nakamura’s take-on and was moving consistently toward goal. His sudden change of direction right before the pass — accelerating into a more dangerous position — appears to have caused the late surge in his OBSO. Despite Nakamura choosing to pass to Ueda, the data suggests Kamada may have been the better option in terms of scoring potential at that moment.
Tanaka (No. 17)
Tanaka’s OBSO trajectory closely mirrors Kamada’s — with a sharp rise immediately after Nakamura got past the defender and again just before the final pass. Footage reveals that Tanaka changed direction just before Nakamura’s second take-on, moving into space behind his marker. The nearby defender subsequently dropped toward the goal to cover, creating a pocket of space between Tanaka and the goal. As a result, his OBSO value rose above Ueda’s in the final moment, indicating that Tanaka had also positioned himself as a high-value passing option.
AFC Final Qualifier – Japan vs Australia: Mitoma’s Shot
This sequence begins with Morita’s long ball, which Minamino controls and lays off. Mitoma then receives the ball, drives forward with a dribble, and ultimately takes the shot himself.
In this situation, we’ll examine the following three aspects:
- Was Mitoma’s decision to dribble immediately after receiving the ball a good one?
- Did his pause and change of direction during the dribble create a valuable moment for the team?
- Was his decision to shoot the right choice given the circumstances?
To begin, let’s evaluate Mitoma’s initial decision after receiving the ball by comparing his OBSO value to those of his nearby teammates at that moment.
Table 1. OBSO values for Mitoma and other players at the moment he received the ball.
Player (Number) | OBSO value |
Mitoma (7) | 0.02664 |
Minamino (8) | 0.02477 |
Ueda (9) | 0.04781 |
Kubo (20) | 0.01817 |
Figure 3. OBSO heatmap at the moment Mitoma received the ball.
According to Table 1, Ueda had a higher OBSO value than Mitoma at the moment the ball was received. Figure 3 supports this, showing available space in front of Ueda, with a denser (darker) area on the heatmap compared to the space ahead of Mitoma. This suggests that playing a direct pass to Ueda could have been a viable option.
However, the OBSO values for the other players were lower than Mitoma’s at that moment. This may indicate that Mitoma chose to carry the ball himself in order to increase the scoring potential of his teammates through his dribble.
Next, let’s evaluate whether his dribble — particularly the pause and cut inside movement — was effective in shifting the defensive structure and creating higher-value opportunities.
Table 2. OBSO values for Mitoma and other players at the moment he made the cut inside.
Player (Number) | OBSO value |
Mitoma (7) | 0.04319 |
Minamino (8) | 0.04538 |
Ueda (9) | 0.08042 |
Kubo (20) | 0.05574 |
Figure 4. OBSO heatmap at the moment Mitoma changed direction.
Table 2 shows that all players had higher OBSO values than they did in Table 1. This suggests that Mitoma’s pause and directional change was effective in creating higher-value opportunities for his teammates.
Notably, Kubo’s OBSO value more than doubled compared to when Mitoma first received the ball. As Australia’s No.19 had dropped deeper towards goal, space consistently opened up in front of Kubo — as shown in both Figures 3 and 4. This positional advantage brought him closer to goal and significantly increased his scoring potential, reflected in his elevated OBSO value.
Finally, let’s examine whether Mitoma’s decision to shoot was the optimal choice in that moment.
Table 3. OBSO values for Mitoma and other players at the moment he took the shot.
Player (Number) | OBSO value |
Mitoma (7) | 0.05434 |
Minamino (8) | 0.04518 |
Ueda (9) | 0.08122 |
Kubo (20) | 0.05597 |
Figure 5. OBSO heatmap at the moment Mitoma took the shot.
A comparison between Table 2 and Table 3 shows that while Mitoma’s own OBSO value increased, those of the other players remained relatively unchanged. This rise can be attributed to his movement further towards goal after changing direction.
However, at the moment of the shot, both Ueda and Kubo still had higher OBSO values than Mitoma. As seen in Figure 5, there was a fair amount of space in front of both players. This suggests that, had Mitoma chosen to pass, Ueda would have been the most optimal option in terms of scoring potential.
AFC Final Qualifier – Japan vs China: Mitoma’s Goal
This scene features Kubo drawing in defenders before passing to Doan, whose cross was met by Mitoma for the goal. In this section, we’ll examine:
- Whether Kubo’s hold-up play (or "pause") was effective
- The quality of Doan’s cross
Let’s begin by analysing whether Kubo’s decision to hold the ball and delay the pass had a positive impact.
Table 4. Comparison of OBSO values when Kubo received the ball and when he passed to Doan.
Player (Number) | OBSO values at the moment Kubo received the ball | OBSO values at the moment Kubo passed the ball |
Morita (5) | 0.02848 | 0.04196 |
Mitoma (7) | 0.0004084 | 0.001054 |
Minamino (8) | 0.02626 | 0.03477 |
Ueda (9) | 0.02335 | 0.04476 |
Doan (10) | 0.03321 | 0.03944 |
Kubo (20) | 0.03328 | 0.02974 |
Figure 6. Transition of the OBSO heatmap from the moment Kubo received the ball to when he passed to Doan.
As shown in Table 4, the OBSO values of all players other than Kubo — who was engaged in a one-on-one situation — increased during this phase. This indicates that Kubo’s hold-up play had a positive effect, raising the scoring potential of his teammates. In fact, Morita, Mitoma, and Ueda all made runs into the penalty area during this moment, which meant there were now more viable scoring options available.
Interestingly, Doan’s OBSO value also rose, despite him appearing to move away from goal. This can be explained by the fact that Kubo drew in two defenders — as highlighted in the live commentary — creating space for Doan to receive the ball in a position where he could immediately shoot. In short, Kubo’s delayed action made Doan a free and dangerous option.
Let’s now turn our attention to the quality of Doan’s cross that led to Mitoma’s goal.
Table 5. OBSO values at the moment Doan delivered the cross.
Player (Number) | OBSO value |
Morita (5) | 0.05406 |
Mitoma (7) | 0.001865 |
Minamino (8) | 0.03865 |
Ueda (9) | 0.02820 |
Doan (10) | 0.04058 |
Figure 7. OBSO heatmap at Doan’s cross.
As shown in Table 5, Mitoma’s OBSO value was lower than that of the other players. Since OBSO accounts not only for scoring probability and occupancy, but also for transition probability, this suggests that the model estimated a low likelihood of the ball successfully reaching Mitoma — who was positioned on the far side. In other words, playing a cross to him was considered difficult from a probabilistic standpoint.
Therefore, the fact that Doan managed to deliver that pass implies the quality of his cross was exceptionally high.
Conclusion
Leveraging broadcast footage opens new doors for advanced tactical analysis — even in environments where official data is unavailable. Metrics like OBSO allow us to quantify moments in the game that often spark debate: “Why didn’t he pass there?” or “What was the reason behind that dribble?” This adds depth to our understanding, whether from a fan’s or an analyst’s perspective.
The two methods introduced in this article enable detailed analysis using only standard broadcast footage. When combined with manual tracking, they can produce insights with a level of quality close to official tracking data — without the need for expensive dedicated systems. This approach is also adaptable to other team sports, such as basketball or rugby, where court or field markings are clearly defined, further contributing to performance development across multiple disciplines.
Looking ahead, advances in AI-driven automatic tracking and video streaming analysis could make it possible to display OBSO scores in real time. As a result, not only professional clubs but also amateur teams may gain access to deeper tactical insights using video footage.
Our company will continue to develop and refine systems for OBSO analysis based on broadcast footage, while also supporting its adoption by football clubs and media outlets. If you're interested in exploring these possibilities, please don’t hesitate to get in touch.
株式会社Playboxのホームページ 👉️ https://www.play-box.ai/
自動撮影・編集ができる手頃なAIスポーツカメラ「playbox」 👉️ https://www.play-box.ai/lp
お問い合わせ
Playboxへのご質問やご相談、ビジネスの提案など、お気軽に以下のメールアドレスまでご連絡ください。
