This article discusses the properties of high replay value games. In the article a new abstract concept of Periodic Dilemma Generator (PDG) is defined and presented as a tool for analyzing and creating high replay value games. The article suggests that in order for a game to have a high replay value, it should incorporate a PDG system. More precisely, this article introduces a new hypothesis: having a PDG is a necessary and sufficient condition for a game's replayability. Significant amount of the discussions and examples in this article are connected with hyper-casual (HC) games, but the conclusions are directly extendable to any other genre of games.
Let's start our discussion with a frequently overlooked concept called replayability or replay value. Replayable games usually generate more downloads, have higher LTVs and, eventually, higher ROIs .
It is more than obvious that in order for a game to be replayable the game should be interesting every subsequent time you play it. This brings us to a very famous quote from Sid Meier. In the 1989 GDC conference, more than 30 years ago, the legendary game designer Sid Meier said: "A game is a series of interesting decisions". This phrase has become a huge topic for discussions. Some people have strongly agreed and cited in many books and articles, whereas others have argued if it is true at all. Therefore, in 2012, after 23 years of saying this phrase, Sid Meier returned to GDC with another talk to explain what are interesting decisions and to give more clarifications. And in his presentation he literally said: "And what makes a decision interesting? Ahhh... Kind of almost easier to look at it of what is NOT an interesting decision. For example, ....".  And he presents lots of very important thoughts, helpful examples but in his talk he does not give a clear and simple definition of what an interesting decision is.
Many other game designers and authors have discussed the topic of replayability, interesting decisions, and other related concepts based on different perspectives. For example, in her book Tracy Fullerton discusses the types of decisions.  She thoroughly discusses each type, explains advantages and drawbacks, but all this is not conclusive as, again, there is no clear formula of an interesting decision.
Wikipedia says: "Replay value or replayability is a term used to assess a video game's potential for continued play value after its first completion." .
E. Adams, J. Dormans are discussing replay value as a result of a large probability space, i.e. many different possible states in a game (e.g. in chess, SimCity, Civilization, etc). They mention "Large probability space makes games more replayable; as a player, you can be confident no two games will be exactly alike. This adds to a game's appeal, especially if the outcome of each play-through is as unpredictable as the first." 
On the other hand, Steve Swink in his book considers replayability a result of expressivity of controls. If the game enables the player to express himself in many different ways then the game has depth and, hence, it is replayable. He also mentions that goals and challenges are also important in terms of replayability. For example, leaderboards or competition with other players may give a player a reason to come back and replay. He writes "Adding additional sensitivity to the controls enables more subtlety and nuance to the interactions. Supporting these new, more expressive interactions with rules [goals and challenges] and context [spatial layout] enables the designer to craft the feel of the game at various levels, making it deeper." 
Yes, designers also use vague terms such as depth to explain the replayability. Why is this term vague? There is no strict definition or common understanding of depth. In my previous article dedicated to the concept of Most Memorable Moment (MMM) I have defined the depth as "The number of reasons you have to do the same thing" . But this is not a universally agreed definition (probably not very helpful either) and usually depth is understood only intuitively.
As you might notice, different authors explain the replayability reasons differently. We all understand what a replayable game is, but we have no good tools to make one. Although, since 1989 until now, more than 30 years, none of the authors mentioned above have introduced a concrete definition or formula to identify what an interesting decision is, anyways, all of them have discussed the topic from different angles, have developed the knowledge and eventually created a brilliant foundation on which, I believe, it is possible to build a solid theory.
In this article I will introduce and examine such a tool for analyzing and designing games. Throughout this article we will discuss HC games frequently. Why? For two main reasons:
- Games have four major constituent elements: Mechanics, Story, Aesthetics, and Technology . In this article we are concentrating on game mechanics and its replayability, so in this context hyper-casual games are the most distilled game genre. Indeed, HC games might have a small narrative but they don't have a story, the aesthetic part of HC games is only for providing a player with some context and helping to understand the Metaphor  and technology is not important in this context at all.
- The second reason is that our experience at Noor Games is mainly concentrated on the HC genre. At Noor Games we have been creating HC games since 2018, and due to the accumulated expertise in this area it is the best way for me to communicate our experience and findings to the readers.
However, the discussions and findings are not anyhow limited to HC genre. They are extendable and transferable to literally any type of games.
The Periodic Dilemma Generator
In a very boring game players may still have available choices and can make decisions. We, as game designers, need to understand what kind of choices or decisions make a game interesting and, hence, replayable. For this reason I want to define and analyze a new concept called Periodic Dilemma Generator.
Imagine a system, where you have a clear goal to achieve. The system is periodically presenting you with choices, and you need to decide which one to choose. Choosing between system-provided choices is the only mechanism to achieve or fail the goal. Imagine that in the suggested system the presented choices have the following properties:
- Choices are meaningful , i.e. they satisfy the following two sub-criteria:
- Informed choices - the system should provide you with both the cost and the benefit of every choice.
- Choices with consequences - the choices make a difference, i.e. they affect the system's state by bringing you closer to the goal, or taking away from reaching it.
- Choices are dilemmatic , i.e. choices are conflicting and in every situation there is no optimal answer.
- The outcome of a choice is unpredictable in terms of achieving the goal [3, 5], i.e. there is no choice, that choosing it will certainly achieve the goal. And also there is no situation, that in spite of available choices reaching the goal is impossible.
This system is an abstraction, and we will call it Periodic Dilemma Generator (PDG). The concept may sound a bit complicated to understand but the following examples will fully clarify the idea. But before we jump into examples, there is a very important statement. High replay value games have one or more PDG systems built into the core loop of the game. The main hypothesis of this article is: having a PDG is a necessary and sufficient condition for a game's replayability. In other words, PDGs are providing players with interesting choices.
Examples of PDGs
We should note that it is better to build the periodic dilemma generator into the core mechanics. Yes, PDG can be built in the level design, or in the meta-game, but first and foremost it should be built into the core part of the game mechanics - in the core loop of the game. In order to understand what it means to build a PDG inside of the core mechanics, let's analyze a couple of examples to demonstrate how PDGs act as a system built into the core loop of games or even in game genres.
In his book Steve Swink analyses fighter game combat mechanics on the example of Street Fighter II . He explains that when you press "one of six attack buttons" on a joystick the game starts to "play back an animation which changes the shape of the avatar". During playing the animation "the player is locked out of further input until the animation is complete". "For the “light” attacks, the duration is very short and will not interrupt the correction cycle of the walking mechanic. The heavy attacks can take almost one second to complete, however, disrupting the continuity of control."
So where is the PDG?
- Meaningful choices. In the menus of fighting games for every attack it is known how much the damage is, and how many frames the animation will play, i.e. how many frames the input will be locked. Hence, the player attacks are both informed and have consequences.
- Dilemmatic choices. The choice to attack is always dilemmatic. Player knows that during the attack either he will damage the opponent, or he will get locked by exposing himself to the opponent's attacks. Another dilemma is following: he can choose to attack stronger, but get locked longer. This clearly creates a risk-reward situation for the player. Jesse Schell calls it triangularity - when a player is faced with a choice between two actions: one that is low risk and gives a low reward, and another that is risky but has a bigger payoff. 
- Unpredictable outcome. Till the last moment of the fight neither of the players knows who will win the round even if players health points (HP) are significantly different (one is low HP and other one is almost full HP). A comeback is usually a possible option if player skills are evenly matched.
There are many fighting games such as Street Fighter, Mortal Kombat, Tekken, Killer Instinct, etc. All these games are inherently interesting, because they have the above mentioned PDG implemented in the core of the combat mechanics. But also these games are different: they have different settings, characters, animations, gameplay options, art styles, simulation precisions, etc. Anyways, all these differences would not guarantee any of these games to be a high replay value game.
Let's see where the PDG is hidden in shooter games. Obviously, it is in the shooting mechanics. Let's go point by point:
- Meaningful choices. Player is making informed decisions as he knows his HPs, how much health an armor adds, how much damage his weapon applies upon headshot or body shot. Choices have consequences. For example, when the player chooses to attack either he hits, and he gets closer to win condition, otherwise he might get hit and lose HPs.
- Dilemmatic choices. When a player starts an attack, he has a chance to hit the opponent, but also he has a chance to miss and expose himself for the opponent's attack. This is because he is out of cover and visible to the opponent, and the opponent has heard the shooting sound. The second dilemma is when the player tries to aim better, he needs to move less, which immediately makes him an easy target for the opponent. The third dilemma could be the headshot mechanics. In many shooters a headshot results in an instant kill, but body shots - don't. There is a triangularity here - headshot is a bigger reward (instant kill, better animation and sound effects, demoralization of the opponent player), but also it is more risky as the head is a smaller area to aim, and according to Fitts's law, it is harder and takes longer to aim . Body shot is less risky, as the body has a larger area, but also gives a less reward.
- Unpredictable outcome. Here also, like in fighter games discussed above, before every shot the player cannot be sure that he will achieve his goal, and even with low HP it is possible to kill an opponent who has a full HP.
- Meaningful choices. Player is making informed choices as the player knows what will happen when he presses one of four control buttons (accelerate, brake, steering left, steering right). And obviously those four choices have consequences - real time consequences.
- Dilemmatic choices. In every racing game the dilemma is the speed vs control ratio. The more you have speed the less the control is.
- Unpredictable outcome. In a racing game it is possible to have a predictable outcome, when the player lacks behind too much, or he is way ahead of his opponents. In these cases the game becomes uninteresting. Therefore, usually Rubber Banding is implemented for Racing Games. By the way Rubber Banding term originated from racing games. Imagine as if the two vehicles were connected by a large rubber band .
Economy Management Games
These are the games like all incremental games, idle games, strategy games, games where the player needs to build an empire, a city, an economy, a character, a vehicle, etc.
- Meaningful choices. In these games there is a cost for every new item to purchase, build or upgrade. In the game menus you can find clear information about what the benefit of the purchase is.
- Dilemmatic choices. The dilemma is to decide where to invest obtained resources. Game designers implement conflicting choices or even if there is an optimal choice, the designer hides that fact carefully in order to create a perception of a dilemma. For example, the designer might make ChoiceA good at the start of the game, then ChoiceB may become dominant, and then again ChoiceA may become the optimal. The player needs to always check what choices he has. If there is no perception that choices are conflicting a game may still be interesting to complete once, but please take into account that PDG is for making high replay value games. Obvious optimal choices will break the feeling of player's agency - he will not feel smart by making decisions anymore.
- Unpredictable outcome. In Economy Management games before every purchase or upgrade the player is not sure that he will achieve his goal. For example a player can't pass a level, because his vehicle or weapon has weak parameters. There are many upgrades that the player can do, and he knows each and every upgrade's cost and benefit in numbers. But the player is not sure that there is one particular upgrade that will guarantee him to win the level.
Hyper-Casual Game Examples
- Meaningful choices. The colors and the geometry of the helix clearly shows the deadly and safe sectors providing an informed choice. Choices have consequences as by turning the helix you either go down through the holes or fall on a deadly sector and die.
- Dilemmatic choices. There is a triangularity: the player can go down multiple steps and obtain higher speed, making it riskier to land on a safe sector, or go step by step and play less risky and less rewarding gameplay. The second triangularity is because the game suggests invulnerability (big reward) if you manage to pass multiple steps in a row without landing on any safe sector (big risk).
- Unpredictable outcome. You never have a situation in Helix Jump, thinking that it is impossible to win this level, or in spite of your moves you will surely win the level.
- Meaningful choices. The player sees the pipe shape, knows his avatar's speed and positions and speeds of other non-player characters (NPC). The control of aquapark.io is consistent, and hence the player has full information of costs and benefits of his actions. Choices, such as avatar movements and physical interactions between avatars create consequences.
- Dilemmatic choices. When a player decides to attack an NPC there is a triangularity - either he succeeds to kick an NPC and eliminates an opponent, or he risks to jump out of the tube and makes it harder to safely land back on the tube, especially if he attacks on the turns. Other triangularity arises, when the player decides to take a shortcut by jumping from the tube, in order to land on the lower parts of the tube or even directly into the pool.
- Unpredictable outcome. Aquapark.io's NPCs are programmed so that if you start the game and watch, you might even get the first place, and vice versa, if you play actively, you might not end the first. This is due to the above mentioned Rubber Bending effect in the Racing Games section. Therefore, until the last moment of the game, there is no idea if you are going to win or lose the race.
Physics based games usually have a higher replay value, because it is almost impossible to have the same physics simulation for every repeated session. Also random generation of the level plays a positive role in expanding the possibility space of a game. Both Aquapark.io and Helix Jump utilize these features.
We can easily and quickly analyze other replayable games too by examining their dilemmatic choices only:
- Push'em all's PDG operates on the fact that the closer the player's avatar is to the roof edge, the higher chances he has to push NPC off the roof. But there is a higher risk that either he will fall off, or enemies will kick him off the roof.
Crash Landing 3D's main dilemma generation happens because there is a limited fuel and the player needs to choose between consuming more fuel and obtaining more control, or vice versa.
Paper.io's PDG operates on the fact that the longer lines the player draws to cover larger areas, the higher is the risk to be killed by other NPCs - again a triangularity.
Archero is a shooter game, and, as described above, there is a dilemma of movement and stopping the movement. Should the player stop movement to start shooting and expose himself to enemy attacks? When should he do that? Where should he do that?
AdVenture Capitalist is one of the simplest idle games ever created. In his blogpost Anthony Pecorella has covered the math behind AdVenture Capitalist and from the article it is clearly shown that the game designer is deliberately trying to hide the optimal choice from the player by using multipliers . In his blogpost Anthony writes: "As you progress through the game, different generators take priority and have the biggest impact on income. The player gets to try to identify these priorities (they won’t always be obvious right away), and this provides more variety and less predictability."
Super Mario Bros. - Levels in this game are designed so that the best feeling you get is by rushing forward. It is not only about the feeling but also it is easier to pass the level with rushing if you have enough skill to rush properly. The designer has tuned the following parameters to make the rushing playstyle more favorable: enemy positions and speeds, star power-up behaviour, Lakitu and Hammer Bro and other enemy behaviours, Mario is jumping higher and longer while running, time limit on the level, etc. But it is also harder to play rushing. If you rush it is harder to control, but if you manage to control it is easier to pass the level. If you don't rush, it is easier to control, but the challenges are harder to overcome, and also there is a time pressure. So should the player rush, or not?  This is a very good example of how the PDG can be built not into the core of the game but into the level design. Enemies' behavior and the level design are periodically generating dilemmas for the player.
Pac - Man - This game economy is designed so that the enemies are becoming more and more aggressive over time till there is a relaxation period.  Over that time the player has a dilemma - to use power pills and attack enemies, or to take a risk and wait by hoping not to consume power pills quickly and save them for harder times.
Looking through the examples above, it could be concluded that the only way to create dilemmas is by using triangularity. In fact, triangularity is a really good tool for generating a dilemma. Jesse Schell says "I find that about eight out of ten times someone comes to me asking for help on a game prototype that “just isn’t fun”, the game is missing this kind of meaningful choice."  and, obviously, for a good reason. But triangularity is not the only way of offering dilemmas to the player. A simple example of chess will clarify this. Consider a situation, where you have a choice to capture two different pieces, where capturing one piece will create a strong material advantage, and capturing another one will create a strong tactical advantage. Now this is a dilemma without a triangularity. Another example, in a shooter game you can buy a low damage but a faster shooting weapon, or you can buy a high damage but a slower shooting weapon. Both have advantages and disadvantages, this is a dilemma, but without a triangularity.
As it was mentioned above PDG is an abstract concept, as it is very precise, unambiguous and you will never meet it in reality in the pure form. But the real world is much more complex and we need to consider that we have utilized this abstraction in order to simplify the situation and be able to deal with it. Later we will also consider the psychology of the player and his perception of a PDG system.
Using PDGs in Games
As mentioned above PDG is an abstraction, a simplification of real systems, but this is a very useful tool to model, analyze and understand how games provide high replay value and build our own replayable games. We will use PDG the same way we use point particle or ideal gas abstract and simple concepts in physics to model complex natural phenomena.
In fact, PDG is transparently describing what is an interesting choice. In order to achieve high replay value, games need to implement PDG systems to periodically generate interesting choices. After examining examples above we already should have a pretty good understanding of what is a PDG but we need to delve into the subtle details to be able to implement PDGs into our games. For that reason let's understand what it means to implement a PDG in a game. In order to understand that we need to not only discuss how PDG can be built into games but also how player's skills and psychology starts to play a role in the perception of the PDG. Let's first discuss PDGs as systems in the core mechanics.
Implementing a PDG in a game
So the player starts a game, sets a goal and waits for the game, as an implementation of a PDG system, to present him with choices.
- A game is an interactive system, hence players need to make decisions, or speaking another way, take choices to change the system's state. Choices presented by PDG should be meaningful, i.e. be informed and have consequences in terms of getting closer to the goal. The goal is not necessarily the win or lose criterion. It can be any smaller sub-goal a player might have.
- The choices presented to the player should not be obvious - it should be a dilemma. Yes, players feel smart when they see a good choice and take that choice, but if the game will continue to provide with obvious good or obvious bad choices, players will stop feeling smart or even interested. It is well known how dominant strategies ruin the game and how game designers need to balance the game and make sure that there are no obvious good choices. There should always be decisions that players make but they are not sure it was the optimal.
- The outcome of the choice should be unpredictable in terms of the player's goal. You might argue that this point contradicts to the point that player choices should be informed. You might ask if it is an informed choice, then how the outcome could be unpredictable? To understand this, think about chess. Before every move a chess player clearly knows what his move will change on the board. But in the long term he is not sure if he will achieve his goal. Professional chess players don't play the game till checkmate. If they identify that the outcome is predictable, i.e. checkmate or draw is unavoidable, they stop playing at that point. The game's outcome stops being unpredictable. In other words, the player should have a hope that he can achieve his goal, but not be sure that he can do that.
To summarize, a player needs to make not obvious choices, such that he knows the consequences of those choices, but that is not enough to predict the game's outcome.
The table below presents a schematic example of how a PDG works. On every step the PDG offers choices (column Offered Choices). One of the choices is being selected (column Selected Choice). As a consequence of the selected choice the system changes its state accordingly (column System State). In the next step again some choices are offered. This is done periodically - in a loop. For example, initially the system is in state S. On step 2 the PDG system offers 3 choices - A, B, C. Player selects B choice. After that the system state changes to S + B. All offered choices (A, B, C, D, E) are both meaningful (informed and have consequences) and dilemmatic, but in every state of the system (S, S + B, S + B + D) it is not predictable whether the user of the system will achieve or fail his goal. As long as these conditions are met, we will say we have a PDG system.
|Step number||Offered Choices||Selected Choice||System State|
|2||A, B, C||B||S + B|
|3||B, D, E||D||S + B + D|
Considering player's skills and perception
As I have mentioned earlier, projecting PDGs abstract concepts onto games is only one part of the job. The last part is to consider the player's skills and psychology. The player needs to perceive that the game is acting like a PDG. If technically the game is a PDG but the player fails to perceive it that way, then the interest will be lost. Here are some notes that we need to consider while applying PDG when designing a game:
- Meaningful Goals. The game designer cannot set an arbitrary goal in a game and hope that the player will care for achieving that goal. Also the designer cannot hope that the player will create meaningful goals for himself. For example, the designer cannot think that he will create a shooter game without an explicit goal and hope, that the player will define his goals by himself. You, as a designer cannot expect that the player will set goals such as advancing through the level by killing all enemies, or play so that the player will be unnoticed by enemies. The correct way of doing this is to create a conflict in a game which immediately challenges the player. E.g. in a runner game put enemies in front of the player so that they are able to kill the player. In this case the player will have to deal with those challenges. And he will define goals for himself like the ones mentioned above and those goals will already be both meaningful and interesting to achieve for the player.
- Achievable and clear goal. First of all, every game needs to have a clear goal. Activities that do not have a clear goal, are chaotic and do not result in Flow experience . So make sure in your game the player clearly knows his goal. The player not only needs to have a clear goal, but also needs to believe that he can achieve his goal.
- Informed choice. Based on the player's skills, the player perceives different available choices. The player can expand his choices by expanding his skills. Even if the choice is perceived as available, it is a matter of skill or luck to execute the choice and to get the benefit of that particular choice by paying the corresponding cost.
- Choice consequences. Even if the choice taken has had an actual consequence, but the player didn't perceive it, then the choice will not be perceived as meaningful. In order to get a feeling of Flow there should be a clear and fast feedback for every action . Missing a consequence is the same as not getting a feedback for a taken choice.
- Dilemmatic choices. Here again the player's perception is playing a significant role. If technically there is an optimal choice, but the player is unaware of that opportunity, then he might perceive the situation as dilemmatic and enjoy the game.
- Unpredictable outcome. If players perceive that the game's outcome is unpredictable they will continue to play. This is a typical case, when a good chess player plays with a bad chess player. At some point for the good player it is obvious that he has won, but the bad player still continues to play, as he does not have the same perception, and the game is still interesting for him.
- Player Skill or Competence - Over time the player learns that even though in every situation he faces a dilemma, he will find situational patterns, where one choice is typically better than the other choices . In these situations the player will feel himself in a less dilemmatic state. In other words, he will become skillful and will feel competent, which is good. But when the player will become so much competent that the game outcome will become predictable, then the perception of PDG will be lost and the interest will be lost too. There will be no replay value left in that game for that particular player.
- Safe to fail - When situations are dilemmatic, it is harder to blame a player for his failure, as he always can argue back, that the other options had drawbacks, whereas in situations when there is a clear optimal choice, the player is feeling more "unsafe" to make decisions.
- Puzzles - Based on my definition of a replayable game, choices should be dilemmatic. But puzzles are games that usually have optimal choices and usually have only one win condition. Therefore, puzzles have low replay value. You cannot collect the same Jigsaw Puzzle 10 times in a row, and feel fun, but you can play 1000 rounds of SimCity, CS:GO, DOTA or Chess and still have fun. This is why we cannot make games that are fun to play and teach players to science. Probably, this was the reason why Chris Crawford's has made his famous Dragon Speech and left the industry.  Science is not about dilemmas, it is about optimal choices - science is a puzzle for humanity to solve. Therefore, once it is solved, it is not interesting to replay, and after that it is not Safe to fail.
- Self-expression or Autonomy - Because the player is in dilemmatic situations, which means there is no optimal choice, this means that every player will take choices that express their own type of personality. There is no optimal choice, therefore some people will usually play safe (because they are that kind of people in life too), other players will play in a risky way, etc. They will feel autonomy to choose their own sub-goals, their own way of playing the game, their own way of expressing themselves. (Please compare with the thoughts from Steve Swink mentioned in the Introduction.) Playing a game becomes a self expression like an art. And on the other hand, this means that the game designer is not forcing the player the sub-goals he has prepared for the game. The player does not feel that he is obligated to do something, the player feels that he does what he wants to, and during the game, he can change and adapt his sub-goals too. He is free to choose, and none of the choices are wrong. Just over time he will recognize the patterns of situations when some solutions typically work better, he will become skillful, as mentioned above. And even in that situation he will choose to be competent or not - maybe he will feel that he needs to experiment new approaches.
- PDG is a perception. In his book Csikszentmihalyi writes "Enjoyment does not depend on what you do but rather on how you do it."  He writes how some activities are very good for inducing Flow state, and others are not. But even in very uncomfortable states, Flow conditions can be satisfied and the person can experience the Flow state. The same is true for PDG systems. The same system can be perceived as PDG for one person, and non-PDG for another. But there are systems that are very good at being perceived as PDGs - for example Football or Chess games, and others that are not good at that e.g. Rubik's Cube. But it does not mean that it is not possible to perceive Rubik's Cube as a PDG and it is not a must to expect everyone to perceive Football as a PDG.
- Aesthetic, religious, taste issues. People may not want to try a system for some reason, hence will not be able to feel its PDG properties. Another reason is that people can have some aesthetic, religious or other reasons not even to try an experience to even be able to feel whether it is a PDG or not. For example, one of my friends hates Brazilian jiu-jitsu (BJJ) for aesthetic reasons (a man hugging and rolling another sweaty man on the ground), but BJJ is a really Periodic Dilemma Generation machine - imagine like a real time Chess instead of currently played turn based one. Other people might not want to play a shooter game for religious reasons.
Summarizing this section we can conclude that the perception of PDG is highly dependent on a player:
- The same game, even the same situation can be perceived as a PDG by one player and a non-PDG for another one.
- The same situation in a game can be perceived as a PDG initially, then after playing and getting some skills the game might feel like a non-PDG for the same player.
- The same game can provide PDG experience at some moments of the play, and stop acting like a PDG at other moments to the same player.
PDG as a tool for a game designer
Please note that PDG is just a tool in the hands of a game designer. In spite of providing an abstract concept as a tool for designing games, we should not forget that creating a game is an art. There are no strict recipes, and every designer should use his intuition to understand when and how to use the tool. For example, a designer should decide what the dilemma generation period is. For one game it might be every second, for another game it might be every minute. The dilemma generation period might change during the game. Also how critical should be the consequences of the player decisions, should consequences be long term, or short term? There are many things that the designer needs to adjust for a particular game. But not implementing any PDG in a game might be a good reason to obtain boring and low replay value games.
PDG is also very helpful when you want to work on a new game idea. When you want to design a new game (if you are a HC game developer, then you need to make several games per month) you always have a question on what game to make so that it is fun to play. But by using PDG it is possible to transform the question to the following. Consider a PDG that works in some specific way, i.e. we have already an interesting system. So what are the design details of this system? For example, I design a PDG the following way: I have different paths, every path enables a different speed. The one that has a higher speed is also more dangerous to travel in. For this PDG the game designer needs to come up with a game design details (needs to do the detailed design of the system) so that the properties of the PDG mentioned lately will be satisfied and the metaphor of the game will make sense to the player. This is a totally different approach. Compare again:
- What game to make so that it will be fun to play? (There is no certainty or something to start with, right?)
- How do I design a system that implements the properties of a certain PDG? (This is certain, this is more like a clear task for a designer to start to work on.)
You might ask: "How should I come up with a PDG idea?" I think you would agree that this is much-much easier than "coming up with a game idea that is fun". Also you can reuse existing PDGs in different contexts and you will obtain different games.
It might be argued that this definition of replayable games is useless and does not express players' taste as we have seen many repetitive games in AppStore's top charts. For example, Pack Master, Dentist Bling both reached top #1 in the "US Games" category of AppStore and many other games that have very repetitive gameplays too. Aren't they replayable games, don't they have good retention values? No, those are not high retention games! Sounds paradoxical? Yes, it does. And it took years of experience for our team to understand this mystical truth. If a game is not a good one, what is the motivation to spend a significant amount of money on UA and bring it to the top charts? There are two explanations for this:
- The HC game publisher could find a very marketable but very repetitive gameplay, which means that both CPI and retention are low. This means that it is possible to obtain positive ROI not because of high LTV but because of low CPI. For example, compare Game1 with CPI \($0.18\), and average LTV \($0.65\) and Game2 with CPI \($0.05\) and LTV of \($0.19\). Some publishers call these low ROI games Ultra-casual and are OK to publish them, others are not interested in publishing positive ROI Ultra-casual games because of very low margins.
- There is fierce competition between publishers in the HC market. Newer or smaller publishers need to demonstrate to game development studios their capabilities. They need numbers such as N downloads, M top games published, etc to tell studios how powerful and successful they are. Therefore, they might even take games with ROI ≤ 0% and push to the top charts. (Fake it till you make it?) By the way, in such deals, studios might get no payouts or very small payouts.
Thus, reaching AppStore tops and/or getting millions of downloads neither means a game is commercially successful nor it means the game is good at all. And this is not specific to the HC genre. Therefore, as a side-note, game development studios need to pick games as references and publishers as partners very carefully.
There can be other seeming problems related to the concepts discussed above. In his GDC 2012 talk titled "Sid Meier's Interesting Decisions" Sid Meier says that his "series of interesting decisions" are criticized for example on Guitar Hero and Jigsaw Puzzle as these games don't present interesting decisions, but they are fun to play.  For the same reasons it might be attempted to criticize the replayable game definition written above. The following thoughts are meant to clarify that potential criticism:
- We should consider Guitar Hero not a game but a musical instrument. The operation principles of Guitar Hero are no different from real guitar playing. It is clear that it might be really fun to play piano piece or run 1000 meter distance over and over again, but the reasons, mechanisms of this experience should not be discussed in the context of games. This is well covered by the concept of the Flow .
- All puzzles are very low replay value experiences, as they have optimal solutions and have predictable outcomes, especially if the solution is already known.
- There are video games that are awesome games but are not replayable, e.g. Far Cry, Tomb Raider, The Last of Us, etc. They might be very enjoyable single time experiences. One might even replay the full game 5 times, but no one will play it for 1000 times, and equally enjoy the game, whereas you can play Chess, DOTA, CS:GO, Mortal Kombat as much as you wish and even enjoy more than the first sessions. For example Journey by Jenova Chen is an amazing game, it is an art, but by the definition above it is not a replayable game. It does not have a significant replay value. It is more like an interactive animated movie. These types of games are out of scope of this article since we discuss high replay value games.
PDG and Skinner Box
One reason why PDG might be so efficient recipe to create replayable games might be that it creates a Skinner Box  inside a game. For building the most efficient Skinner Box one needs a variable schedule/ratio of reinforcement. Hence, the outcome should be unpredictable. But if in the game the player has an optimal choice, then it means he will make that choice to reach his goal and it should result to a predictable outcome (otherwise it is a gambling, and gambling is the purest form of Skinner Box). Hence, the choices should be dilemmatic. And, of course, the choices should be meaningful, otherwise those are not choices at all - either the player takes random (not informed) actions from his perspective or the actions have random results (consequences) on the game.
An abstract concept of Periodical Dilemma Generator (PDG) is defined. We state that high replayability games have one or more PDG systems built into the game. Moreover, there is a hypothesis that having a PDG is a necessary and sufficient condition for a game's replayability. Of course, this is true, if there is a corresponding player perception about the system's PDG properties. PDGs are good tools for analyzing or making high replay value games, if the designer understands the tool and finds a good approach how to use it. Several games or game genres are analyzed to show how PDGs work and where the PDG should be implemented. It is clear from those examples that the replayability is usually the virtue of the core mechanics and the replayability can be achieved without adding new characters, weapons, levels or a meta-game into games. From the discussion above it is straightforward that all the conclusions are directly applicable on literally any genre of games.
- Game Feel: A Game Designer's Guide to Virtual Sensation by Steve Swink, 2008
- GDC 2012, Interesting Decisions by Sid Meier
- Game Design Workshop: A Playcentric Approach to Creating Innovative Games, Fourth Edition by Tracy Fullerton, 2018 (Chapter 11, Improving Player Choices)
- Replay Value Wiki
- Game Mechanics: Advanced Game Design by Ernest Adams and Joris Dormans, 2012
- The mighty MMM concept in hyper-casual game design, 2019
- The Art of Game Design: A Book of Lenses by Jesse Schell, 2nd edition, 2015
- Fitts's Law Wiki
- Rubber Banding Wiki
- The Math of Idle Games Part I
- Flow: The Psychology of Optimal Experience by Mihaly Csikszentmihalyi, 2008
- Theory of Fun for Game Design by Raph Koster 2013
- GDC Founder Chris Crawford's Dragon Speech
- Behavioral Game Design