Novacore aims (pun intended) to deliver an upgraded FPS experience to gamers today by leveraging the capabilities of machine learning.
Through dynamic enemy ai, Novacore provides hours of gameplay with NPC difficulty that constantly improves based on a player’s skill. This model of progressive difficulty is designed to challenge the player’s weaknesses from previous playing data.
In Novacore, players join a dynamic community where they can showcase their individuality through custom skins for characters and vehicles. The 3D assets have been expertly crafted using the latest techniques in mesh optimization and UDIM texturing, allowing for the creation of over 10,000 unique characters. With a vast selection of 3D parts to choose from, players can truly make their characters their own. Additionally, players have access to up to 10,000 customizable vehicles.
With machine learning, lessons can be tailored to a player's specific needs. Games like Minecraft and Starcraft have tapped into the power of machine learning as well.
In Novacore’s context, we first. identify a player’s behavior and aiming patterns.
We use this information as inputs into algorithmic machine learning. The system then evaluates how the player can collectively improve low performing player aim patterns. The patterns are recognised and rated, helping a player to achieve better aim during a high pressure scenario.
This data is then used to generate customized training scenarios to help players improve on their weaknesses.
High-Low Frontal Targets
In Team Battles on large multi-level maps, it is possible for one player to encounter two different opponents at the same time.
If both opponents are approaching on different levels, e.g. ground floor and rooftop, a player would most likely be shot down by his or her opponents.
In this situation, a player might first aim at a high platform. After shooting the target down, another target appears on ground level. The player has to immediately transition to aim at the new target.
However, most players are unable to adjust quickly enough after the change in level and end up getting shot down.
In Capture-The-Flag or Deathmatch games, maps are often more complicated. This allows for a variation of tactics and approaches. It is common for enemy players to approach from multiple directions.
When playing in such maps, it is possible for a player to encounter an enemy in front of them and behind them at the same time.
In this scenario, a player has to aim at the target in front. After shooting this target down, the player has to quickly turn 180 degrees to aim at another target behind.
Most players are unable to adjust their aim quickly enough after a drastic turn.
At the same time, the environmental pressure helps players to improve their aim gradually as they keep practicing. With machine learning, the time taken for a player to perform well in such scenarios is greatly reduced.
The system focuses on learning patterns that result in the lowest performance in the player’s aim. From there, the system creates pressure scenarios that are unique only to the player’s past performance.
By incorporating these reinforcement machine learning techniques into enemy ai, players of all FPS levels can enjoy a progressively challenging experience.
This, along with the diverse design of enemy AI and an array of terrains, creates an immersive single-player experience that is never repetitive.
Players have a refillable energy pack per game to use for special abilities. Energy tanks can regenerate slowly. Players are incentivised to move through the arena to pick up 25% energy tanks in order to regenerate faster.
With this energy pack, players can do the following: