Case Study: How Liga Pro transformed the collection & application of betting data

LaLiga and OSAI case study on AI data scout for table tennis. OSAI uses AI in sports to collect accurate betting data in real-time

Case Study: How Liga Pro transformed the collection & application of betting data
AI data scout by OSAI collects betting data in real-time

Accuracy - 99.31%

Data Latency - 14.7ms. Based on the 430 LigaPro matches in April 2022

Liga Pro - one of the biggest semi-professional table tennis tournaments - partnered with OSAI to transform the collection and application of the betting data.  Accurate and fast data collection is critical for Liga Pro. With the 24/7 format and multiple locations, betting data is a challenge.

Liga Pro was one of the first clients to implement AI data scout by OSAI. The technology uses AI/ML and Computer Vision for real-time betting data collection from the live stream. OSAI analyses low latency video feed at 120 frames per second and collects data after each frame, creating a real-time feed of accurate betting data.

Data points OSAI collects for table tennis

Serve

For every player, based on any format (default is bo5)

Ball in-game


Net

When the ball hits the net 

Score prediction

When a player wins a score

Let

Based on the referee’s gestures

Scoreboard OCR

OSAI recognizes the score additionally to inferring it from the events

Deep data

Ball 

Bounce on the table, speed, trajectory

Pose estimation

16 points of players tracking in real-time (head, body, arms, legs)

Tracking

Players position on the floor, walking distance during rally, game

Analytics

Heatmap 

Players position on the floor, bounces on the table 

Strokes count

Count (by game, total, by player)

Rallies length

Short, long rallies

Winning streak


Time

Pure in play, between rally, by player

Occupied zones

Zones on the floor (default are 3 — short, middle, long zone) on the floor while a player is playing

Accuracy of betting data produced by AI data scout

OSAI collects real-time betting data for Liga Pro matches across several halls.  To measure the AI data scout performance OSAI team constantly monitors the accuracy and latency of the system and ensures top-notch performance 24/7.

With thousands of games per month, producing reliably accurate data feeds is critical. Let’s take a look at a small sample of games in April. And compare AI data scout performance with human scouts.

The average accuracy of betting data collection across 430 matches was 99.31%. The minimum threshold was 97.4%. All errors in betting data are automatically corrected within 1 second. AI data scout collects data according to the official rules. Any discrepancies in the data collected are caused by the difference between human and computer decisions. To maximize the system accuracy, the OSAI AI data scout recognizes on-screen information (like score) to correlate the data streams.

It’s important to mention that for evaluation purposes we take the human scout's accuracy as 100%. In reality, humans are prone to errors and the feed of betting data they produce isn’t 100%. Typically there is a process in place to correct data errors, that cause data latency of up to 2s.

Speed of betting data collection

With the development of in-play and super live betting formats, the speed of data collection becomes paramount. The comparison of game event data input reveals that AI data scout performance is not just comparable, but better than the human scouts.

We’ve selected two data points to demonstrate. AI data scout by OSAI is 24.1% faster in identifying the start of the game and 51.2% in keeping the score.


OSAI slower than a scout

Equal

(diff. <1s)

OSAI faster than a scout

OSAI slower, %

Equal, %

OSAI faster, %

Event

>5s

5s

4s

3s

2s

1s

1s

2s

3s

4s

5s

>5s

POINT

1

0

1

0

9

28

412

210

205

45

5

1

7

4.2

44.6

51.2

BALL_IN_GAME

0

3

7

13

57

69

534

61

68

43

18

7

20

16.6

59.3

24.1

AI data scout vs Human scout — speed comparison of game event input

Data latency is another important factor in the speed of betting data collection. AI data scout is designed to minimize the data latency - it performs all calculations locally and only transfers the data feed to the cloud, from where betting data is distributed to the tournament organizers and betting providers. As of the measurement taken on 13th May, AI data scout has a 14.7 ms latency (8ms + network latency).

Optimizing the costs of betting data collection

Betting data collected by human scouts is not cheap, as it requires recruitment, training, and management of a large team with high turnaround times. Unlike human scouts, AI data scout is available 24/7 in any location upon a press of a button.

Depending on the game location and time of play OSAI helps to cut the costs of betting data collection from 40-80%.

OSAI - more than betting data  

OSAI analyzes the live video feed at 120 frames per second and collects accurate betting data like ball trajectory, athlete positions, ball speed, and more. This unlocks multiple opportunities to enhance content quality, improve game performance and grow viewership & engagement.

OSAI powers Liga Pro live streaming with its real-time Mixed Reality solution.

Liga Pro implemented OSAI rea-time Mixed Reality to enhance content quality and use immersive animations to visually explain the game and engage the viewers. OSAI designed animations that are consistent with the Liga Pro brand. These include ball tracking, player profiles, zones occupied, player positions, distance from the table, walked distance by the players, and more.

We continue to develop and test new features, like rally heatmap, and virtual LED boards. This allows to further maximize the entertainment value of Liga Pro content and enhance the viewing experience. OSAI and LigaPro work together on the design, testing, and iteration. And plan to launch it in production soon.

OSAI collects real-time diverse betting data portfolio. More importantly, OSAI understands the rules of table tennis, therefore the AI-powered sports technology is capable of fair & unbiased decision-making.

One of the key value propositions of Liga Pro is that the games are played at convenient-for-players times. Even if that means late night or early morning. There is a challenge that comes with it: finding and training referees for such unconventional game time. With matches played back to back, referees can be prone to errors caused by tiredness.

Liga Pro and OSAI collaborated on AI-assistant for referees to help them make fair and accurate decisions.

OSAI technology uses AI and real-time betting data to apply table tennis rules to the match that’s being played. The technology understands the accurate location of the ball and athletes and makes unbiased decisions.

OSAI solution includes an application installed on the referee’s tablet. The technology sends real-time suggestions, like changing the score and awarding points. The referee evaluates the accuracy of the suggestion and can change it in case of a difference of opinion.

LigaPro’s format with matches played 24/7/365 has a growing popularity among viewers and athletes of all levels. With thousands of matches every month in various locations, OSAI and Liga Pro are scaling up the partnership to cover all table tennis matches with AI data scout, real-time Mixed reality, and AI assistant for referees.


Interested to learn more about AI data scout for betting data and computer vision technology by OSAI? Get in touch today!