Academia.eduAcademia.edu

Figure 3 – uploaded by Lazare Kolebka

See full PDFdownloadDownload figure
The K-rating, denoted as K, for Player 1 in this case, determines how much a player’s rating will change after one game. This revised K-value takes into account the number of games the player has played to balance the effect of each game on their rating. After conducting numerous tests22, a formula was developed for calculating the K-value for each player. For Player 1, this is expressed as:

Figure 3 The K-rating, denoted as K, for Player 1 in this case, determines how much a player’s rating will change after one game. This revised K-value takes into account the number of games the player has played to balance the effect of each game on their rating. After conducting numerous tests22, a formula was developed for calculating the K-value for each player. For Player 1, this is expressed as:

Related Figures (80)

FIGURE I: Ranking From the Players of Hays Brussels (February 2023). Calculation made by the author based on Appendix 1, FIGURE |: Data Collected with Google Forms - February 2023, p.41. 1.2. Limitations and Issues in the Initial System
FIGURE I: Ranking From the Players of Hays Brussels (February 2023). Calculation made by the author based on Appendix 1, FIGURE |: Data Collected with Google Forms - February 2023, p.41. 1.2. Limitations and Issues in the Initial System
This approach rewards risk-taking, as winning against a higher-rated player results in a more significant increase in a player’s rating, as shown in the table below: On the other hand, if a higher-rated player goes against their winning probability and loses against a lower-rated player, their rating will be significantly impacted: they will lose more points, and their opponent will gain more points.
This approach rewards risk-taking, as winning against a higher-rated player results in a more significant increase in a player’s rating, as shown in the table below: On the other hand, if a higher-rated player goes against their winning probability and loses against a lower-rated player, their rating will be significantly impacted: they will lose more points, and their opponent will gain more points.
FIGURE V: Expected Score Based on the Scenario Shown in Figure III Table created by the author to illustrate how the expected score calculation is done in a baby-foot scenario. Now that the expected outcome of a four-player match has been determined, this information can be incorporated into a new formula that considers multiple variables that affect the match and player ratings. As discussed earlier, the K-value can be modified to better fit the needs of the rating system. This new formula considers the number of games played by each player, reflecting their seniority as well as the outcome of the match. A dominating 11-1 win should have a greater impact than a closer 11-10 win.
FIGURE V: Expected Score Based on the Scenario Shown in Figure III Table created by the author to illustrate how the expected score calculation is done in a baby-foot scenario. Now that the expected outcome of a four-player match has been determined, this information can be incorporated into a new formula that considers multiple variables that affect the match and player ratings. As discussed earlier, the K-value can be modified to better fit the needs of the rating system. This new formula considers the number of games played by each player, reflecting their seniority as well as the outcome of the match. A dominating 11-1 win should have a greater impact than a closer 11-10 win.
Chart created by the author to illustrate the impact of the score difference on the point factor.
Chart created by the author to illustrate the impact of the score difference on the point factor.
After building the database system and processing all the existing data, the goal was to propose a solution that would make this system more sustainable by allowing players to register new games. A web application was developed to facilitate game inputs into the database, using data validation techniques to ensure data accuracy. The same logic used to process the Excel*®°? file was applied to sort, insert, and calculate the Elo rating for each new submission, ensuring proper storage of match information. By using this method, users can query the database at any time to obtain relevant information in real time, such as rankings or additional insights, through the web application. This strategy ensures sustainable and efficient data management and accessibility for future use. 4.3. Database Design and Modeling
After building the database system and processing all the existing data, the goal was to propose a solution that would make this system more sustainable by allowing players to register new games. A web application was developed to facilitate game inputs into the database, using data validation techniques to ensure data accuracy. The same logic used to process the Excel*®°? file was applied to sort, insert, and calculate the Elo rating for each new submission, ensuring proper storage of match information. By using this method, users can query the database at any time to obtain relevant information in real time, such as rankings or additional insights, through the web application. This strategy ensures sustainable and efficient data management and accessibility for future use. 4.3. Database Design and Modeling
Found below in Figure IX, the 'Player' table of the database model stores each player’s information, including their first name, last name, and employment status at Hays Brussels. This data is saved only once in the 'Player' table and subsequently linked to other tables using PKs and FKs. Managing the employment status of a player enables the display of lists featuring either active or all players as needed. The process of adding or deleting players is simplified by utilizing the 'Player' table, rather than changing both a Google Sheets and a Google Forms. Two types of relationships exist between tables in this database model: one-to-many and many-to-many. The relationship between the 'Player' table and the 'Match' table is many-to-many since a player can participate in numerous matches, and multiple players can be involved in a single match. A junction table called 'PlayerMatch' bridges this relationship, containing two foreign keys: 'player_id' (referencing the participating player) and ‘'match_id' (referencing the corresponding match). This structure ensures the accurate association of players and matches as demonstrated in the code below:
Found below in Figure IX, the 'Player' table of the database model stores each player’s information, including their first name, last name, and employment status at Hays Brussels. This data is saved only once in the 'Player' table and subsequently linked to other tables using PKs and FKs. Managing the employment status of a player enables the display of lists featuring either active or all players as needed. The process of adding or deleting players is simplified by utilizing the 'Player' table, rather than changing both a Google Sheets and a Google Forms. Two types of relationships exist between tables in this database model: one-to-many and many-to-many. The relationship between the 'Player' table and the 'Match' table is many-to-many since a player can participate in numerous matches, and multiple players can be involved in a single match. A junction table called 'PlayerMatch' bridges this relationship, containing two foreign keys: 'player_id' (referencing the participating player) and ‘'match_id' (referencing the corresponding match). This structure ensures the accurate association of players and matches as demonstrated in the code below:
FIGURE XIV: Elo Rating Processing. Diagram created by the author to illustrate the code in the Appendix XIII: Python Code for Elo Rating Calculation, p. 6 The output of the expected score processing’? is then used to determine the final parameters of the algorithm, such as the K-value, which depends on the number of games played, and the point factor44. These variables are then used to calculate new ratings for players and teams, which are stored in the database with the correct timestamp*>. This process is repeated for each match to calculate the Elo rating of each player involved. 4.5. New Ranking Proposition After processing and accurately saving all matches and ratings in the database, SQL queries can be utilized to retrieve rankings for a specific date. The data model’s design allows filtering the 'PlayerRating' table to display and sort only active players and those who participated within the desired time frame.
FIGURE XIV: Elo Rating Processing. Diagram created by the author to illustrate the code in the Appendix XIII: Python Code for Elo Rating Calculation, p. 6 The output of the expected score processing’? is then used to determine the final parameters of the algorithm, such as the K-value, which depends on the number of games played, and the point factor44. These variables are then used to calculate new ratings for players and teams, which are stored in the database with the correct timestamp*>. This process is repeated for each match to calculate the Elo rating of each player involved. 4.5. New Ranking Proposition After processing and accurately saving all matches and ratings in the database, SQL queries can be utilized to retrieve rankings for a specific date. The data model’s design allows filtering the 'PlayerRating' table to display and sort only active players and those who participated within the desired time frame.
Points-based ranking - February 2023 Elo-based ranking - February 2023 FIGURE XVI: Elo-based Ranking. Calculation made by the author based on lines 317-353 of the SQL query shown in Appendix XX, p.103-104. FIGURE XV: Points-based Ranking. Calculation made by the author based on Appendix 1, FIGURE |: Data Collected with Google Forms - February 2023, p.41.
Points-based ranking - February 2023 Elo-based ranking - February 2023 FIGURE XVI: Elo-based Ranking. Calculation made by the author based on lines 317-353 of the SQL query shown in Appendix XX, p.103-104. FIGURE XV: Points-based Ranking. Calculation made by the author based on Appendix 1, FIGURE |: Data Collected with Google Forms - February 2023, p.41.
After several months of development and testing, the first version of the app was launched in late March 2023 with basic features like uploading games and viewing rankings. This 5.2.2. App Releases and Updates The following CSS code was created by the author to illustrate the process of the compatibility on mobile. The complete code for the match processing can be found in Appendix XXXI: CSS Code for Styling, p. 126.
After several months of development and testing, the first version of the app was launched in late March 2023 with basic features like uploading games and viewing rankings. This 5.2.2. App Releases and Updates The following CSS code was created by the author to illustrate the process of the compatibility on mobile. The complete code for the match processing can be found in Appendix XXXI: CSS Code for Styling, p. 126.
FIGURE | : Data Collected with Google Forms - February 2023.
FIGURE | : Data Collected with Google Forms - February 2023.
FIGURE II : Google Forms of the Initial Solution (Desktop Version).
FIGURE II : Google Forms of the Initial Solution (Desktop Version).
FIGURE V : Google Sheets Ranking Deployed on Web (Mobile Version).
FIGURE V : Google Sheets Ranking Deployed on Web (Mobile Version).
APPENDIX VI: Data Cleaning Using Google Sheets
APPENDIX VI: Data Cleaning Using Google Sheets
APPENDIX VII: Sample Log Data Collected During the Elo Calculation Processing FIGURE VII: Sample Log Data Collected During the Elo Calculation Processing.
APPENDIX VII: Sample Log Data Collected During the Elo Calculation Processing FIGURE VII: Sample Log Data Collected During the Elo Calculation Processing.
Screenshot generated by the author to illustrate the project's development timeline. As of April 30th, 2023, there have been over 200 commits and 3500+ lines of code added.
Screenshot generated by the author to illustrate the project's development timeline. As of April 30th, 2023, there have been over 200 commits and 3500+ lines of code added.
This presentation has been created by the author to onboard the users to the new app in March 2023.
This presentation has been created by the author to onboard the users to the new app in March 2023.
This appendix presents the testing process for the algorithm, focusing on the documented tests performed to determine the optimal parameters for calculating the Elo rating. Although numerous tests were performed throughout the development process, not all of them were meticulously tracked. However, the tests presented here were carefully documented as soon as the initial version of the algorithm was established.
This appendix presents the testing process for the algorithm, focusing on the documented tests performed to determine the optimal parameters for calculating the Elo rating. Although numerous tests were performed throughout the development process, not all of them were meticulously tracked. However, the tests presented here were carefully documented as soon as the initial version of the algorithm was established.
APPENDIX XI : SQL Code for Database Creation
APPENDIX XI : SQL Code for Database Creation
APPENDIX XIV: Python Code for Excel Data Processing
APPENDIX XIV: Python Code for Excel Data Processing
APPENDIX XIX: Python Code for Web Application
APPENDIX XIX: Python Code for Web Application
APPENDIX XXIV: HTML Code to Delete a Match
APPENDIX XXIV: HTML Code to Delete a Match
APPENDIX XXVI: HTML Code to Display Games History
APPENDIX XXVI: HTML Code to Display Games History
Related topics:
StatisticsReal Time ComputingWeb DevelopmentMobile application developmentData ScienceData ManipulationData ArchitectureElo Rating
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved