Kabaddi Analytics: How numbers and video can improve performance?
I got a chance to work with Bengal Warrior as a Sports Physiotherapist, it was a great experience, however, my intuitiveness made me drawn towards the analytics component in Kabaddi. Kabaddi is so intense a game that the team's percentage of winning the match keeps changing with the minutes.
Kabaddi is a team sport that involves physical contact. Two teams of seven players compete on opposite ends of a court measuring 10 by 13 metres (33 feet 43 feet) for men and 8 by 12 metres (26 feet 39 feet) for women. Five extra players are kept in reserve for each team in case of injury. The game consists of two 20-minute halves separated by a 5-minute halftime break during which the teams switch sides. A player from the attacking team, known as the “raider,” sprints into the opposing team’s side of the court during each “raid” and seeks to tag as many of the seven defensive players as possible.
Two perspectives can be drawn from sports dynamics, it’s both team games and individual performance games. For example, if a team’s defence is strong and two raiders are strong in performance, there is a high chance that the team will have extra advantages.
Let’s talk about raiders (two types — power raiders or agile raiders):
Raider Metrics
Raid Points — Points scored by the Raider include both Touch and bonus points
Successful Raids — Number of raids in which the raider managed to score a point
Super Raids — Number of raids in which the raider has managed to score more than 3 points
Super 10s — Number of matches in which the raider has scored more than 10 points.Do-or-Die Raid Points — A raid after two consecutive empty raids, where the raider has to compulsorily score a point against the opposition’s defence.
Average Raid Point: Average raid points scored by a raider during the entire session.
Let’s talk about defenders:
Defender Metrics
Tackle Points — Points scored by the defender include both Touch and bonus points
Successful Tackle — Number of tackles in which the defender managed to score a point
Super Tackle — Number of tackles in which the defender has managed to score more than 3 points
High 5s — Number of matches in which the defender has scored more than 5 points.Average Defender Point: Average tackle points scored by a defender during the entire session.
Assistive Point: A point scored by a defender during a tackle when a defender is assisting.
These are the metric that can be scrapped from different websites such as sportskpi, prokabbadi, kabaddi adda, or even any database.
Kabaddi Analytics can be bifurcated into three sections:
- Video Analytics
- Statistical Modelling
- Mathematical Modelling
Video Analytics plays a huge impact in analyzing an athlete during the competition as well as in training; let’s deep dive in:
In sports and science, video can be used in a variety of ways. Coaches and players are increasingly turning to the media to assess and correct techniques as well as analyse team and individual performance.
Analyze the performance of the team
Re-watching the game with the correct tools can be significantly more valuable than seeing the game on video.
First, an analyst must annotate the video (in real-time or after the game) with your chosen key points, such as raids, errors, and specific plays, as well as each player’s involvement. The coach or player can then filter and view their preferred facet of the game, such as all raids or tackles scored by a given player or all opposition team errors. Game highlights can be made fast, and you have immediate access to a variety of performance metrics.
Analytical Skill Techniques
Taging of an athlete is important in each phase of a match or training.
Skill video analysis is extremely effective for detecting and addressing issues with an athlete’s skill or technique. The following are examples of things that can be monitored and identified using video analysis:
- Raiders:
- Mechanism of Toe Touch & Release — Bonus Point
- Hand Reach Analysis — Touch Point
- Jump Analysis
2. Defenders:
- Catch Analysis
- Catch and Hold Analysis
- Catch and Lift Analysis
- Ankle Hold Analysis
- Dash Analysis
Apart from skill analysis, biomechanics analysis of an individual athlete can be incorporated for each position to name it:
I. left raider
II. right raider
III. left corner
IV. right corner
V. left cover
VI. right cover
VII. centre
2. Statistical Modelling
- A pre-historic data has a vivid significance in any spectrum of analysis, in sports also it has the same role. As mentioned above table, it can be used to predict the average raid point or raid point in a match using supervised machine learning models.
- Descriptive statistics can be used to determine the player performance through the session as mentioned below:
- Similarly, a successful player can be assessed using statistical formulas as mentioned below:
- Hypothesis Testing: Following are the key hypothesis that can be formulated:
- Markov Chain Algorithm: Kabaddi is a sport in which team performance changes a lot within a short frame of time. Most of the time in Kabaddi, we have to make tactical decisions based on the current circumstances of the team rather than past facts. As we all know, in a Markov Chain, we must forecast probabilities based on current events rather than past events.
3. Mathematical Modelling:
As a series of metrics are available in context with both raider and defenders, mathematical modelling can be done to ponder over:
- frequency — How frequent a raider or defender is performing?
- recency — How recent was a raider’s or defender's performance?
- per match performance — how many raiders or defenders per match performance?
whenever we are developing mathematical modelling: we should consider three criteria:
Constancy: does the metric measure the same thing over time
Discrimination: does the metric differentiate between players
Objectivity: does the metric provide new information
You can read about it more in Metrics Used in Sports Data Analytics
Happy Learning…!!
