Combining everything #TeamRICOCHET has developed over the course of the last two years with new Machine Learning advancements, RICOCHET: Anti-Cheat™ is preparing for the launch of Call of Duty®: Modern Warfare® III with a stronger and faster process to combat cheating.
Machine Learning, in combination with client and server-side systems that continue to evolve and grow, helps advance both the speed and accuracy of our prevention techniques and detection systems.
How Does #TeamRICOCHET Use Machine Learning?
Machine Learning advancements have been integrated into our tech to help with efficiency and speed in prevention, detection, and removal of cheaters. Machine Learning advancements enhance our team’s ability by:
- Examining client and server data to find new cheat behaviors
- Issuing account challenges to validate abnormal behavior
- Collecting and collating problem accounts for action
In short, Machine Learning helps us anticipate behavior better and operate with more effectiveness, with our team validating for accuracy.
Machine Learning works in concert with our team, providing information to make account decisions – but Machine Learning systems do not issue bans.
Machine Learning also helps enhance existing tools. One example of how we’re using Machine Learning to accelerate our anti-cheat capabilities is with the Replay tool.
Machine Learning x Replay Investigation Tool
Earlier this year we announced a replay investigation tool that captured gameplay data so it could be converted into video internally, allowing our teams to review player matches for problem behavior. This tool has been beneficial since it launched, but the team wanted to drive toward a new goal: Speed.
On average, a #TeamRICOCHET teammate could review somewhere in the ballpark of 700 replay clips in any given day.
Some clips are easy: the most egregious “rage hacking” is simple to spot, but the Replay Investigation Tool was helpful to identify hackers who used tools to give them a slight advantage that was harder to spot in-game, such as wall hacks.
For the launch of Modern Warfare III – and across all titles protected by RICOCHET: Anti-Cheat – the #TeamRICOCHET team is activating Machine Learning processes to increase the efficiency and strength of our anti-cheat efforts.
For the Replay Investigation Tool, a Machine Learning model is trained to identify suspicious behavior like wall hacks or raging (plus many others), and immediately prioritizes and alerts the team to review the issue for account action. A single PC running the model can review up to 1,000 clips per day – a number that grows exponentially when multiple computers are tasked with operating this specific Replay Machine Learning Investigation model.
We’re just getting started on Machine Learning integration for the Replay Investigation Tool, but we’re excited to see how it evolves over time. A major focus for this and many advancements is Ranked Play modes across our titles, combating anyone attempting to jump the ranks of the leaderboard unfairly.
This is one of the many ways Machine Learning helps identify and prioritize issues for our team, allowing Team Ricochet to develop new prevention strategies, detection techniques, and mitigations.
