In the last two years, it is increasingly possible to hear the thesis that the game industry is stagnating, many developers are engaged in self-copying, and games at the gameplay level do not offer anything new.

At the same time, since about 2020, we do not see any significant jumps graphically. However, games are becoming more and more demanding on iron every year, and it is not always clear where the resources are going.

In this article, I will try to understand whether games have really hit the ceiling of their development, and I will also try to find the growth points for video games.

I will warn you right away: I will not use cheat codes and touch on the topic of VR/AR games, cloud gaming, but I will try to focus on the classic PC (console) experience.

To begin with, the progress of video games is inextricably linked to the development of digital technologies, so it is quite obvious that the period of explosive growth of games from the mid-90s to the beginning of the 10s is directly related to another round of the digital revolution.

In recent years, the pace of game development has slowed down significantly. Every year we get more and more games that are similar to each other, and indisegment is no exception. What are dozens of clones of the same thing worth? Vampire Survivors. At the same time, all these games copy sets of mechanics without trying to develop them. But nevertheless, projects periodically appear, even in the AAA segment, which are able to surprise the player with a new approach and technologies.

For myself, I have singled out five directions in which games can develop in the future, and all of them in different variations can be combined in one project, and I will try to talk about them using the example of popular games.

1. Artificial Intelligence NPC

One of the key things that can diversify the gameplay experience is a complex multi-level AI of opponents or teammates, which can react differently to the player’s actions depending on the context.

This technology is especially important for stealth games. In most games, enemies have three states: patrolling, arousal, and aggression. In most games, enemies switch between these states quite quickly and easily forget even active fights with the player.

The game boasts one of the most complex and interesting AI Alien: Isolation. The alien has two levels of artificial intelligence:

• a macro level (the director) that controls the overall behavior of the alien, deciding when and how it should patrol or attack the player. This “director” determines the zones in which the player is located and sends signals to the xenomorph about potential victim locations;

• micro level (predator) that directly controls another. He acts on cues from the director, but navigates the level himself by responding to sounds, visual stimuli, and movements. However, the “predator” does not know where the player is in real time, so he needs to constantly search for him.

In addition, the xenomorph learns and adapts to the player’s actions. If the player often hides under tables or in closets, strangers will look there more often. Also, he does not follow fixed routes and often changes the direction and search tactics.

All this turns him into a difficult and unpredictable opponent. If you want to understand the behavior of the xenomorph in more detail, I recommend watching this video.

Of course, applying such AI not to one opponent, but to a large group of enemies will significantly complicate the design and balancing of the game, and for the player the process of passing will become extremely painful. Therefore, in my opinion, this technology has excellent potential in a slightly reduced format.

2. Procedural generation

With procedural generation, everything is not so simple. Yes, on the one hand, it helps to create large volumes of content (No Man’s Sky and Starfield) and refresh the gameplay experience on repeated playthroughs (basically any roguelike, e.g. Hades), however, the frequency of unique content drops sharply: points of interest in locations are often repeated, and after the first study, all subsequent content of a similar type becomes uninteresting to the player. When trying to create procedurally generated quests, everything falls into the most boring “give and take”.

You may have a legitimate question: if creating levels/locations and quests using procedural generation is a dubious business, why am I talking about it at all? The fact is that it needs to be handled very carefully, and at the moment, in my opinion, there are not many successful examples of the use of the procedure.

And one of the best examples is the Nemesis system developed by Monolith Productions for a series of games Middle-earth. Its essence is to create unique and dynamic enemies that can remember interactions with the player, adapt and even develop a personal relationship with him. For example, if the player defeated an orc in battle, at a new meeting he will remember it and want to take revenge.

The Nemesis system, in addition to creating unique NPCs with the help of procedural generation, possessing memory and capable of development, also has a dynamic hierarchy, where all its members have their own goals and tasks and, depending on the circumstances, are able to move through this same hierarchy, a system for adapting NPCs to player actions and tactics, as well as a recruitment system. Something similar, but less successful, Ubisoft tried to reproduce in Watch Dogs: Legionwhere the player could recruit and take control of absolutely any NPC in the game.

If you’re interested in learning more about the specifics of Nemesis, I recommend this video, as well as a talk by Chris Hogue of Monolith Productions about the philosophy and implementation of this system.

In essence, all of the above continues the ideas from the first point, while competently using procedural generation.

3. Number of NPCs

I think many players at the time of release Assassin’s Creed Unity In 2014, two things were surprising: the terrible technical condition of the project and the huge crowds of people filling both the narrow streets and the spacious squares of Paris. Moreover, the crowd was not static and reacted to the player’s actions.

Unfortunately, due to the poor technical condition, the focus of the players has shifted, and this truly breakthrough technology, in my opinion, has not received the attention it deserves. By tradition, I am giving a lecture on the design of the game process with the crowd and the peculiarities of the technical implementation.

However, five years later in 2019, the crowd technology got a second chance at the time of release World War Z by Saber Interactive and Days Gone by Bend Studio. Only this time, both studios offered the player not just to observe and minimally interact with a large number of NPCs, but to directly confront the crowd, which significantly refreshed the gaming experience of third-person shooters. About how the authors Days Gone created a horde capable of functioning under the conditions of an open world, you can read in this presentation.

Crowd technology continues to actively develop. In the same year 2019, a series of games started A Plague Tale from Asobo Studio, in which players were offered to interact already with a crowd of rats. And more recently, the same Saber Interactive again used crowd mechanics Space Marine 2.

4. Interactivity of the world

By interactive world, I mean a lot of gameplay interactions of the player with the objects of the environment and system design, where the game gives you freedom of choice and the game mechanics are applicable to many entities.

For example, if the game has an option to set fire to a tree, then it should be possible to set fire to a wooden door to enter a locked room. Oil should ignite from fire, and from water the same fire should go out. If a character is tired, they fight worse and move slower, etc. I think you know what I mean.

Separately, I would like to dwell on the topic of simulation of natural phenomena affecting the gameplay. I can only think of one person who truly realizes the potential of this system in his games, and that’s Hideo Kojima.

The Buryat genius began to implement these systems already from Metal Gear Solid 3: Snake Eaterwhere the Naked Snake can get wet, hypothermic, and is less visible at night than during the day. It is also important to consider the camouflage system and change it depending on the type of terrain. IN Metal Gear Solid V: The Phantom Pain he continued to develop these ideas, and now Vienna Snake (sorry for the spoiler, but the game is 9 years old) can hide behind sandstorms or tropical downpours.

In the future Death Stranding 2: On The Beach Kojima promises even more weather effects on gameplay, whether it’s a sudden snow avalanche or a flood that sweeps away players’ buildings.

The main problem with this approach is the complexity of the design and balance of such systems, as well as the abilities, levels and quests associated with them. Literally few studios are capable of this now.

So, speaking of system design, we can say that there are many representatives in the video game market who adhere to this concept. The brightest of which are sandboxes like Minecraft or Terrariaas well as the adventures of the series The Legend of Zelda.

But still, in my opinion, players lack big-budget story projects with high interactivity and systemic gameplay. This is evidenced by the stunning success of last year Baldur’s Gate 3 and growing interest in the upcoming release Kingdom Come: Deliverance 2.

5. Simulation

The simulation of everything and everything is, in my opinion, a difficult and practically impossible game concept, because it absorbs all the previous points and requires the perfect execution of each of them.

That’s what I thought until I met Shadows of Doubt — a detective simulator, created as if by the will of Warren Spector himself, which simply amazes with its capabilities. The player can enter any building or room in multiple ways, NPCs have schedules, jobs, personal interests, and relationships, and they react differently to the player’s actions. A small town literally lives its own life.

The only thing missing is directed investigations. Here they are procedurally generated and over time problems may arise, which we talked about in the second point.

It seems to me that this project is the most vivid example of the direction in which the industry should move, where the technical capacities should be distributed.

A complete simulation of a video game world based on an existing one is where you and I will all end up. Video games still have a lot of room for development, it’s just that the pace is not what it was 20 years ago.

If you liked this material, then I invite you to my blog in Telegram, where you can find more similar content. And I would also be interested to read your thoughts in the comments on this topic.

Thank you all for your attention!