|
|
| Line 2: |
Line 2: |
| | <br /> | | <br /> |
| | <center>[[Image:proposal.jpg|200px |link=1718t1is428T14/Proposal]] [[Image:poster.jpg|250px |link=1718t1is428T14/Poster]] [[Image:app3.jpg|200px |link=1718t1is428T14/Application]] [[Image:paper.jpg|207px |link=1718t1is428T14/Paper]]</center> | | <center>[[Image:proposal.jpg|200px |link=1718t1is428T14/Proposal]] [[Image:poster.jpg|250px |link=1718t1is428T14/Poster]] [[Image:app3.jpg|200px |link=1718t1is428T14/Application]] [[Image:paper.jpg|207px |link=1718t1is428T14/Paper]]</center> |
| − |
| |
| − | == Problem & Motivation ==
| |
| − |
| |
| − | Overwatch is a popular online video game where players are put into two opposing teams of six and participate in competitive matches. Each player selects one of 25 hero characters, each of which plays a certain role in the team. To achieve victory, players have to support teammates and take down opponents. Players can choose between casual and competitive game modes, but our project will be focusing on the Competitive mode, where players are ranked based on their performance unlike in the other casual modes.
| |
| − |
| |
| − | As there are a myriad of factors that affect one’s performance in Overwatch, it is difficult for players to create effective strategies that give them consistent victories. Consistent winning is especially important in competitive Overwatch, where each loss worsens a player’s competitive standings. As such, this project aims to analyse the various aspects of Overwatch mechanics, and use data to create effective strategies that give competitive Overwatch players an edge in matches, so they can perform better against more experienced players. This will allow them to obtain and maintain high player standings efficiently in any given competitive season.
| |
| − |
| |
| − | ==Selected Dataset==
| |
| − | Our project will be using the official dataset from Overwatch, available on [....] as well as specialized Overwatch tracking site such as Overbuff.com and Master Overwatch. As raw data downloading options are currently unavailable, our group has chosen to transcribe the data to our own Excel spreadsheet.
| |
| − |
| |
| − | Our project will analyze the following attributes from the dataset:
| |
| − |
| |
| − | ** Hero
| |
| − | Pick rate:
| |
| − | Win rate:
| |
| − | Tie rate:
| |
| − | On fire:
| |
| − | Elims:
| |
| − | Obj. Kills:
| |
| − | Obj Time:
| |
| − | Damage:
| |
| − |
| |
| − | Additionally, our project will also use another dataset for the combination of different heroes’ ultimates. In particular, this will be a matrix that details the effectiveness of a hero’s ultimate when used together with another hero’s ultimate. Effectiveness is measured with the following ranges of value:
| |
| − |
| |
| − | Bad: counterproductive when 2 ultimates are played together
| |
| − | Ineffective: ultimates that are of similar types or types that do not combine well
| |
| − | Maybe: ultimates can combine to but not the best
| |
| − | A bit: ultimates combine to give medium effects
| |
| − | Defense: ultimates combine to give good effects
| |
| − | Nice: ultimates combine efficiently to give very good effects
| |
| − |
| |
| − | == 3) Approach ==
| |
| − |
| |
| − | Our approach to this problem is to first define a competitive player’s journey through a 2.5-month-long competitive season of Overwatch, then provide insights in each step of this journey to ensure the best possible performance. As the season begins, the player will participate in 10 ‘Placement Matches’ (PM). The player’s performance in these matches will determine his/her ‘Skill Ratings’ (SR), a value to denote player standings. Players are placed in different ‘Tiers’ based on SR as shown below (e.g. A player with SR between 1-1499 is placed in Bronze tier). Each post-PM competitive match thereafter causes the player’s SR to increase with each victory, or decrease with each loss, and the player’s Tier adjusts accordingly.
| |
| − |
| |
| − | <br />
| |
| − | [[File:Tier ranking.png|1000px|center]]
| |
| − | <br />
| |
| − |
| |
| − | It is especially important for a player to perform well in the PM because this is the best opportunity for a player to quickly obtain high SR, and therefore high standings. Post-PM victories contribute comparatively small amounts to a player’s SR, unless many victories are obtained, but this takes much more time. The exact Overwatch algorithm for SR calculation during PM is unknown, but it is speculated to be based on the player’s individual performance during the PM (as opposed to team performance). Some players lose all 10 matches but still receive a high Tier, such as Grandmaster, due to high individual performance. In this project, it is therefore assumed that only individual performance affects the SR a player obtains from the PM.
| |
| − |
| |
| − | This project will be separated into two sections. The first section will focus on the PM, To maximize a player’s individual performance in the PM, we will advise the most suitable heroes to select based on a player’s best skillsets and preferred playstyle.
| |
| − |
| |
| − | The second section of the project will focus on post-PM matches. In these matches, there are several factors affecting the SR gained, but the most dominant factor by far is whether the player’s team wins or loses. Therefore, maximizing team performance will be prioritized over individual performance. We have identified four main aspects that affect team performance, namely the player’s team composition, map-based strategies, individual heroes’ ultimate abilities, and the enemy team’s composition. Through analysis of data regarding these factors, we will maximize the chances of team victory by providing strategies highlighting the most effective hero combinations for the team.
| |
| − |
| |
| − | == Storyboard==
| |
| − | ==Timeline ==
| |
