How to improve game design with optimal strategies

    No matter how hard game designers try, in any game you can find mechanics or content that suffer from insufficiently detailed thinking. The consequences - from small exploits to unnecessary or even harmful features.

    A lot of tools were invented to help game designers: GDD formats, various checklists, user stories , portraits, agenda and so on. And still this is not enough.

    Therefore, below I will describe another method for evaluating ideas that will help to better work out the gameplay and take into account more nuances. For example, clans, prostitutes from GTA and not only.



    Let's start with the concept of "optimal strategy." This is a strategy that gives the maximum possible result with the minimum possible investment of resources. In man, the desire for optimal behavior is inherent in it.

    A simple example: banal laziness, as a way for the body to spend a minimum of effort if the expected result does not pay back the investment. If the work does not bring pleasure, we will be too lazy to do it. We will not become happier either, but we will spend less energy.

    In order for a person to start doing something, one must offer him a reward that is comparable or more valuable than the invested forces (this time we do not consider the whip method). But in this case, too, a person will seek to spend as few resources as possible.

    What does gamedev have to do with it?


    Game developers often start from the interests of the user. They are trying to understand what the player wants and what he will do if he is given some opportunities. The game designer comes up with mechanics (for example: “a player can do a trick in a jump”) and decides that they will use it exactly as intended: “to overcome an obstacle, a player will jump and do a feint in the air”.

    But this may not be enough, because from the point of view of the optimal strategy, doing nothing is much more profitable. Then we add motivation in the form of a reward: “a successful trick gives a little acceleration”, as in Mario Kart.



    At this moment, the player is no longer interacting with the game, but the game begins to affect the player’s behavior. It creates motivation to do what is needed. And this is often forgotten.

    An example is more complicated. In GTA, you can contact prostitutes. This is quite in the spirit of the series, but if the opportunity itself is not connected with game mechanics, the player will see a couple of times how it works and continue to do more useful things.

    Therefore, time spent with the ladies restores health (motivation). You can also take money back by force. We will not consider how ethical this is, we just note that in this case it is the optimal strategy (the player will restore health and spend nothing). This means that many are likely to do just that. Over time, money in the game will lose value and they will cease, but only because they will begin to optimize time as a more valuable resource.


    The bottom line is that not only the player interacts with the game, but the game sets its optimal behavior inside.

    Some projects go so far that they themselves begin to play with the gamer. On occasion, try Orwell or Superhot . It’s difficult to convey without spoilers, but at some point you begin to think: do you do what you chose, or what the game provoked you to.

    We introduce clans using the optimal strategy


    I mentioned above that development is often driven by the interests of the player. It is assumed that he has a priori a certain desire, which must be satisfied by adding mechanics. This approach intersects with typical software development. For example, if you are developing an application for a bank, you will have to satisfy the interests of customers - automate account statements, allow money transfers, add chat with support, and so on.

    In gamedev, I would look at the situation from the other side: each new mechanics, each unit of content is a new product and the player needs to explain why he should use it (or make sure he comes to this himself).

    Consider a hypothetical game where we want to add clan mechanics. With their help, players join in groups, play together, communicate, and share resources. And we want the majority of players to use clans, because practice shows: it is more interesting for players who have social contacts in the game to stay in the game.


    War Robots

    Usually the task sounds something like this: "We have different categories of players and we need to make the clans cover the scenarios for using each group."

    Take the abstract, but the most common categories of players:

    • Ordinary player. He plays alone, not for long, but several times a day. It makes up 75% of the audience.
    • Social worker. Playing with friends (real or virtual), rarely, but for a long time.
    • Payer. He plays a little, but efficiently, prefers to pay, instead of farming resources.

    For ordinary players in clans, an increased reward for the battle is provided. For social workers in clans there is a common chat, the ability to quickly assemble a squad and share resources. And clans can be expanded for in-game resources or inaps to complete joint quests faster and have a numerical advantage in guild wars.

    Is this enough to make clan mechanics popular and all players can use it?

    Not. We just answered the question of what can be done with mechanics if you start using it. But ordinary players do not automatically fall into the clan, social players most likely already have channels of communication with friends, and payers do not start paying, they should show them the buy button.

    Optimal player scenario


    The bottom line is that players are looking for the optimal approach to achieve their goals, but we do not yet know what is optimal for them or what goals they pursue. But with the help of analytics, we can find out how often they play, how long and how exactly they behave in the game.

    This habitual behavior can be taken as optimal - this is what the player does, regardless of us. He does not want to waste strength to relearn.

    Therefore, if the optimal use case, which we have determined for mechanics, coincides with the optimal scenario for its user, the mechanics will get the maximum coverage.

    Back to the clans. We still want to motivate players to use this mechanic and decided to add a reward for participating in clans.

    This can be done in many ways:

    1. The more games a clan player plays in a row, the greater the reward he receives (x1.2 for the second match, x1.3 for the third and so on).
    2. A clan player always gets + 20% to the battle reward.

    In the first case, the optimal use case assumes that the player sits down and plays for an hour or two. This use case coincides with the behavior of a social player - they have been playing for so long, and now it will also be profitable.

    But such mechanics will not attract ordinary players who want to play in short sessions. To use it, they will have to invest more energy and make decisions that they are not used to: should they play another game right now? Usually such decisions are not made at all.

    But the second implementation does not impose any restrictions on the usage scenario - it equally suits all types of players. It is safe to say that if we choose the second method, then more players will use clans.

    Of course, you can always make gigantic reward bonuses so that all are in a row broken into clans. But most likely it will harm the game if such an approach is not laid in the foundation: the gameplay will be tied only to clans, and the entire balance will be built by the clan player.

    Optimal Strategy and Content


    The approach from the optimal strategy is applicable not only to the design of the mechanic, but also to the design of the content. Let's say we make a new character for an existing MOBA.

    Let's call it Raptor - and as planned, this will be the first “tank” in the game with three teleporters. From other "tanks" it will be distinguished by unsurpassed mobility. We have just described the character’s USP - what distinguishes it from other class characters and the reason that players want such a tank.

    It may seem that we know what we are doing. The producers are happy - we came up with new content. Satisfied with the marketing - we came up with a promise that we will promote it. Satisfied designers - USP helps to focus and understand what skills Raptor can give. And now we have unveiled the hero, but ... no one is playing for him.

    Because, as with the mechanics, before coming up with the USP, it was necessary to evaluate all the “tanks” existing in the game: what characteristics they possess, how players are used to playing them. Then we would have noticed that all the "tanks" in our game are slow. And it’s not only about the characteristics - the decision-making time when playing for the “tank” allows players to have delays of up to 2 seconds of thought (compared to 0.5 seconds for “killers”). And the USP that we came up with (a teleport with three charges) requires a high decision-making speed, which is unusual for our typical “tank” players. As a result, attacking players do not play with the Raptor (they do not like the characteristics) and tankers do not play (the speaker does not fit).

    If a player wants to learn how to play for a new character, then the optimal behavior will be to choose a character for whom you do not need to relearn. This will require less effort and time. That is (unfortunately or fortunately) the new content is more likely to be popular if it is as old, but “with pearl buttons”.

    Although, as in the previous example, you can always make a new character so strong that they rush to play for him in general. But in the long run this will do more harm than good. A striking example is League of Legends. In 2009, they had certain problems with the balance of new content. The most memorable was the release of fighter Xin Zhao, which was simply unstoppable. As a result, he was even picking up support - everyone wants to win.

    And even whole markets


    Go up one level and look at the game market as a whole - you will notice the trends already described. Namely: popular genres exploit the same formula, polishing individual details or introducing innovations in small portions. An open world and towers, perks and weapons improvements in shooters, cells and shelters in tactics - all this wanders from game to game so that the mass user does not have to relearn and deviate from its optimal scenario. The same logic works here.

    It's very cool to come up with a completely new avant-garde game, which has no analogues, and which is cool to play. But. If your goal is mass, then remember that the player strives not to spend a lot of energy on choosing a new game. Already familiar details and mechanics will help him in this (or a well-known brand - with him you have more opportunities for experimenting with gameplay).

    But it is not all that bad.

    Game developers do not always can or want to do things optimal for the player. The industry needs innovation, new games, mechanics and new content, not self-repetitions.

    Sometimes we want to increase the duration of the session, sometimes we need a new character in the game that will turn the meta and the idea of ​​the gameplay upside down. Such things can and should be done. Just remember that every step you take from the optimal behavior of your player reduces the audience of your idea. And here it is necessary to evaluate what is more important: the benefit from those who nevertheless begin to behave according to your scenario, or greater coverage.

    I would consider such "motivating" mechanics as "unleashing the potential" of the game, and not an unambiguous way to improve its performance. Let's say you have launched a new game. She is good and players regularly play one match a day. You can well do a daily quest for 2 games a day - then the average number of matches played will increase. But it works more precisely on the new project, where the audience has not yet reached the maximum of its potential.

    And if the project is already mature and has an established audience, mechanics and gameplay cycles - then it becomes more difficult to motivate players to do something unusual.

    Why else are optimal scripts useful?


    See how the optimal scenario for using your mechanics (or content) and the optimal scenario for player behavior correlate. How far are you ready to go? After all, if they do not match, then in extreme cases problems can arise. A few examples:

    • When the real optimal scenario for using mechanics does not coincide with what the designer intended, an exploit appears. To avoid this at the design stage, you need to make sure that the intended use is really the most profitable. As an example, the pumping system The Elder Scrolls, where skills swing as they are used. And you can pump stealth without overcoming some obstacles, and spending a couple of hours to sneak along the wall, being behind the enemy. This is clearly not what the designers wanted.
    • When the game rewards the player for the behavior conceived by the designer, but really effective differs from him, this leads to frustration from the fact that the player is not rewarded for the "correct" behavior. A good design feels natural - the player does what he wants, and he is awarded for it.
    • When you have several options to choose from, and the best of them is only one, this creates the presence of one “right” option and a set of useless ones. In MMORPG, guides for quick pumping very quickly appear: you need to go to such and such locations with such and such a crew. As a result, some locations are filled with players, while other players enter by mistake. And the team spent resources on their production. In some cases, this can be corrected by balance. And sometimes, without harm to the game, you can simply remove what will not be in demand. You can minimize production costs without doing too much. But this is a talk about dominant strategies, which deserves a separate article.

    Not a panacea

    In fact, this text is not an attempt to force one to consider strategies and try to think through everything in the world at the stage of writing the DGD. It is useless and will not work for everyone. The article is an illustration of two simple ideas that too often fall out of sight.

    • Think not only about what your player will do, but also why he will do it. Focus not only on opportunities, but also on the motivation to use them.
    • Consider how convenient it is for the player to use these opportunities. The less convenience, the less players will use them.

    A good design is one that meets your intended goals. Do whatever you see fit. The main thing is that the end result does not come as a surprise to you. If you do not forget about these things, then the design will become better, and the resources invested in the development will be spent optimally.

    Also popular now: