Speaker notes: This presentation looks at randomization as a design tool, not as âstuff that just happens.â We will compare several technologies that create random outcomes: dice, cards, spinners, bags of tokens, and software random-number generators. A useful starting point comes from Scott Rogers, who argues that random systems in games are still designed systems. That idea matters: when a designer adds a die, shuffles a deck, or calls a random-number function, the designer is choosing what kinds of outcomes are possible, how often they occur, and how much control the player has. Throughout the presentation, keep asking: What does this randomizer make the player feel? Does it create tension, surprise, fairness, strategy, or frustration? [Sources] - Scott Rogers, *Level Up! The Guide to Great Video Game Design*, chapter 19, discussion of designed random systems. - Tracy Fullerton, *Game Design Workshop*, Part 1 and chapters on formal systems, prototyping, and playtesting. - Colleen Macklin and John Sharp, *Games, Design and Play*, chapter 2, âSkill, Strategy, Chance, and Uncertainty.â [/Sources]
Speaker notes: This is the first big idea. Randomness in a game does not mean the designer gives up control. The designer creates the boundaries of the random system. For example, a six-sided die cannot produce 9. A deck cannot draw a card that is not in the deck. A spinnerâs sector sizes determine the odds. A program can choose from a carefully limited list rather than from every possible number. Rogers makes this point strongly when he says that games may have random systems, but those systems are designed too. For students, this is an important shift in thinking: randomness is not the absence of design. It is one more thing the designer designs. [Sources] - Scott Rogers, *Level Up!*, chapter 19: random systems are designed systems. [/Sources]
Speaker notes: Macklin and Sharp separate two ideas that students often blend together: chance and uncertainty. Chance means that some event is determined by a random process. Rolling dice and shuffling cards are clear examples. Uncertainty is broader. A game can be uncertain even when it contains no random generator. In darts, players do not roll a die before throwing. The uncertainty comes from human performance. In basketball, players react to many changing actions even though the game is not using a random-number generator. Their roulette example sits closer to pure chance. Their poker example mixes chance with skill and strategy. This distinction helps designers ask a better question: âDo I want randomness, or do I simply want players to be unsure what happens next?â [Sources] - Macklin & Sharp, *Games, Design and Play*, chapter 2, âSkill, Strategy, Chance, and Uncertainty,â especially the darts, roulette, basketball, and poker examples. - Image extracted from the project EPUB: figure 2.13, darts and roulette. [/Sources]
Speaker notes: All five technologies can answer the same basic question: âWhich outcome happens?â But they do not feel the same to a player. Dice are fast and visible. Cards remember what has already been drawn. Spinners make probabilities visible through sector size. Bags make the hidden pool physical and tactile. Digital random-number generators can imitate any of these and can also create systems that would be difficult to manage by hand. The term âtextureâ here means the feel of the random system: how players understand it, how much they can predict, whether previous results matter, and whether the random process feels visible or hidden. [Sources] - Fullerton, *Game Design Workshop*, Part 1 and chapter 7: physical prototypes using dice, cards, paper, and other simple materials. - Macklin & Sharp, *Games, Design and Play*, chapter 10: physical prototypes can use dice, index cards, and tokens. [/Sources]
Speaker notes: Dice are one of the oldest randomization technologies in games. Fullerton opens her game design basics section by noting the long history of games and even mentions early clay dice. The standard six-sided die is useful because its outcome space is easy to understand: six possible faces. If the die is fair, each face has the same probability, 1 out of 6. Dice are good when you want repeated independent events. A roll usually does not change the die itself. Rolling a 6 now does not remove 6 from future rolls. Design questions: How many sides? Are all results equally useful? Does the player add a skill value? Do players roll once, reroll, or combine several dice? [Sources] - Fullerton, *Game Design Workshop*, Part 1: early clay dice and the history of games. - Cermak-Sassenrath, *Introduction to Game Programming*, chapter 25 exercise âThree Dices.â [/Sources]
Speaker notes: One die gives a flat distribution when each face is equally likely. Two dice added together create a different pattern. There are 36 equally likely pairs when rolling two six-sided dice. Only one pair makes 2: 1+1. Only one pair makes 12: 6+6. But six different pairs make 7. So the middle becomes much more common than the extremes. This is an important design tool. If a designer wants âusually average, sometimes extreme,â adding dice can create that behavior naturally. This probability example is mathematical elaboration rather than a specific textbook claim, but it extends the textbooksâ treatment of dice, chance, and probability. [Sources] - Macklin & Sharp, *Games, Design and Play*, chapter 2: chance and probabilities. - Cermak-Sassenrath, *Introduction to Game Programming*, chapter 25: three-dice programming exercise. - Distribution values calculated from the 36 possible outcomes of two fair six-sided dice. [/Sources]
Speaker notes: Jeremy Gibson Bond treats dice probability as something designers can understand by enumeration: list every possible roll, count the totals, and graph how often each total occurs. This slide applies that method to one, two, and three six-sided dice. With 1d6, every value from 1 through 6 has the same probability. The graph is flat. With 2d6, totals near the middle occur in more combinations. A 7 can be made six different ways, while 2 and 12 can each be made only one way. The result forms a triangular distribution. With 3d6, the middle becomes even more dominant and the extremes become less common. This is the beginning of the rounded, bell-like shape designers often want when they prefer average outcomes but still want rare surprises. The design lesson is more important than memorizing the graph: changing the number of dice changes the *feel* of randomness even when the dice themselves have not changed. [Sources] - Gibson Bond, *Introduction to Game Design, Prototyping, and Development*, Appendix B, âDice Probability Enumerationâ: enumeration of all possible rolls and graphing the frequency of sums. - Probability values in this graph were calculated by enumerating every ordered result for 1d6, 2d6, and 3d6. [/Sources]
Speaker notes: This graph extends Bondâs dice-enumeration approach to four six-sided dice. There are 6^4, or 1,296, possible ordered rolls. When we count the sum of every roll, totals near the center occur much more often than totals near the edges. The most common total is 14. A total of 4 requires all four dice to show 1, and a total of 24 requires all four dice to show 6, so those extremes are very rare. For game design, this lets us tune how wild a system feels. One die gives a broad, flat sense of possibility. Several added dice pull the game toward ordinary, middle-range results while preserving a small chance of dramatic highs and lows. Role-playing games often take advantage of this difference. A system based on several added dice can feel more predictable than one based on a single die with a similar numerical range. [Sources] - Gibson Bond, *Introduction to Game Design, Prototyping, and Development*, Appendix B, âDice Probability Enumeration.â - Graph values calculated by enumerating all 1,296 ordered outcomes of 4d6. [/Sources]
Speaker notes: Enumeration is exact, but it gets expensive quickly. For one six-sided die, there are 6 possible rolls. Two dice have 36 ordered possibilities. Three have 216. Four have 1,296. Five have 7,776. Eight dice have 1,679,616 ordered possibilities. Bond warns about this directly in Appendix B. Each die multiplies the amount of work. He gives the example that calculating 5d6 takes six times as many cases as 4d6 and 36 times as many as 3d6. The graph uses a logarithmic vertical scale so that both small and large values can be visible on the same slide. Students do not need to master logarithms here. The key visual message is simply that the number of combinations does not grow by adding a fixed amount; it multiplies. This matters for programming as well as probability. Exhaustively checking every possible result is practical for small systems, but a designer may need a simulation, a formula, or a more efficient algorithm as the outcome space becomes large. [Sources] - Gibson Bond, *Introduction to Game Design, Prototyping, and Development*, Appendix B, âDice Probability Enumerationâ: warning about rapidly increasing calculations; 5d6 comparison and 8d6 enumeration example. - Counts follow 6^n for n six-sided dice. [/Sources]
Speaker notes: Bond distinguishes between calculating theoretical probability by enumerating every possible roll and estimating practical probability by actually generating many random rolls. The blue bars show the exact theoretical probabilities for 2d6. The yellow bars show one computer simulation of 1,000 2d6 rolls created for this presentation. Notice that the simulated bars are close to the theoretical shape, but they are not identical. Randomness creates short-term variation. If we ran the simulation again, the yellow bars would change. If we increased the sample from 1,000 rolls to a much larger number, we would usually expect the overall pattern to settle closer to the theoretical distribution. This is useful for game programmers. Enumeration gives an exact answer when the space is manageable. Simulation is useful when systems become too complex to enumerate easily, and it also helps designers test what players are likely to experience in practice. Important source distinction: the textbook proposes repeated random rolls as a way to produce practical probability. The specific 1,000-roll sample shown here was generated for this deck using a fixed pseudorandom seed so the graph is reproducible; it is not a figure copied from the textbook. [Sources] - Gibson Bond, *Introduction to Game Design, Prototyping, and Development*, Appendix B, âDice Probability Enumerationâ: theoretical enumeration and suggestion to use repeated random rolls for practical probability. - Simulation shown here: 1,000 computer-generated 2d6 rolls, created specifically for this presentation. [/Sources]
Speaker notes: Cards are powerful because the random system has memory. When a card is drawn and not returned to the deck, the remaining deck changes. That creates ârandomness without replacement.â If four aces are in a 52-card deck and players see three of them, they know only one ace remains. The probabilities have changed because of previous events. Macklin and Sharp use poker to show how chance and strategy can interact. Players know only part of the deck state, so they use probability plus guesses about opponents. Cards also let designers attach unique rules to individual random outcomes. A die face is usually a number; a card can contain an event, action, item, character, or rule. [Sources] - Macklin & Sharp, *Games, Design and Play*, chapter 2: Texas hold âem, shuffled decks, probability, hidden information, chance, and strategy. - Gibson Bond, *Introduction to Game Design, Prototyping, and Development*, chapter 7: collectible card games and cards with printed rules. [/Sources]
Speaker notes: A spinner makes weighted probability visible. If one section covers half the circle, players can immediately see that it should occur about half the time. That makes spinners useful for younger players and for games where the designer wants the probability to feel concrete rather than hidden in math. A spinner can imitate a die by making equal-sized sectors, or it can make some outcomes more likely by changing sector sizes. Design caution: a physical spinner must be built so friction, pointer shape, and player technique do not strongly bias the result. The device itself becomes part of the system. The project textbooks discuss chance technologies more often through dice and cards than through spinners specifically, so this slide extends their broader chance/probability framework to a common physical randomizer. [Sources] - Macklin & Sharp, *Games, Design and Play*, chapter 2: chance and probability framework. - Fullerton, *Game Design Workshop*: physical prototyping and tangible game components. [/Sources]
Speaker notes: A bag of tokens combines two useful properties: hidden information and a changing pool. If drawn pieces stay out of the bag, the system behaves like a deck of cards without replacement. If pieces are returned, each draw can remain statistically similar to the previous one. Bags are also tactile. Players can physically feel that there are pieces inside, but they cannot see which one they will select. This can create suspense very cheaply in a paper prototype. Designers can also add or remove tokens during play. That means the probability system itself can become a resource that players manipulate. [Sources] - Fullerton, *Game Design Workshop*, chapters on physical prototyping and hidden information. - Macklin & Sharp, *Games, Design and Play*, chapter 10: paper and physical prototypes using tokens and simple objects. [/Sources]
Speaker notes: Cermak-Sassenrath introduces Processingâs random() function as a way to make graphics, audio, levels, and other parts of a program less predictable. In Processing, random() returns a floating-point value in a range. If we want a die-like integer, we convert the result to an int and choose our bounds carefully. The important design idea is that a digital random-number generator is not automatically a game mechanic. The game mechanic comes from what the program does with the number. A programmer can use random values to choose enemy types, item drops, positions, colors, level pieces, or many other features. The designer still decides the range and how results affect play. [Sources] - Cermak-Sassenrath, *Introduction to Game Programming*, chapter 12, section âRandomnessâ: random() makes graphics, audio, levels, etc. less predictable and returns float values. - Gibson Bond, *Introduction to Game Design, Prototyping, and Development*, chapter 36: randomized item drops and selecting random items from a list. [/Sources]
Speaker notes: This is one of the most useful comparisons for students designing games. âWith replacementâ means the selected item goes back into the pool before the next selection. The probabilities reset. A normal die behaves like this because rolling a 4 does not remove the 4 face. âWithout replacementâ means the selected item stays out. The next event has different probabilities because the pool changed. A shuffled deck usually behaves this way during a hand. This distinction affects strategy. Systems with memory let players learn from what has already happened. Counting which cards or tokens are gone can become part of gameplay. [Sources] - Macklin & Sharp, *Games, Design and Play*, chapter 2: poker, hidden cards, and changing knowledge of the deck. - Fullerton, *Game Design Workshop*: hidden information and chance as formal game elements. [/Sources]
Speaker notes: Randomness is not only about what probabilities exist. It is also about what the player can see. A visible spinner communicates odds directly. Dice expose the set of possible faces. A deck hides order but often reveals what cards have already appeared. A digital system might hide everything unless the game tells the player its percentages. Macklin and Sharpâs poker discussion shows how limited information creates uncertainty. Fullertonâs Battleship prototype discussion shows that hidden information can create suspense and variance even when guessing is central. When designing, ask whether players should be able to calculate the risk or whether uncertainty itself is part of the experience. [Sources] - Macklin & Sharp, *Games, Design and Play*, chapter 2: poker, probability, and hidden information. - Fullerton, *Game Design Workshop*, chapter 7: Battleship prototype and hidden information creating suspense and variance. [/Sources]
Speaker notes: This is the comparison slide to use when students are choosing a randomization technology for a design. Dice are excellent for repeated numerical outcomes and quick physical play. Cards are strong when past outcomes should affect future probabilities or when each result needs its own rules. Spinners make weighted probabilities obvious. Bags are useful for tactile hidden pools that can change over time. Digital RNG is extremely flexible but can become invisible or confusing if the game does not communicate enough information. The point is not that one technology is better. The technology should match the player experience goal. [Sources] - Synthesis of Fullerton, *Game Design Workshop*; Macklin & Sharp, *Games, Design and Play*; Cermak-Sassenrath, *Introduction to Game Programming*; and Gibson Bond, *Introduction to Game Design, Prototyping, and Development*. [/Sources]
Speaker notes: Fullertonâs play matrix gives designers a way to discuss how much an outcome depends on skill versus chance. The other axis considers mental calculation versus physical dexterity. The matrix is not a scorecard where one corner is âbest.â It is a discussion tool. Chess sits strongly toward mental skill. Blackjack includes substantial chance but also player decisions. Other games mix these factors differently. For a student project, the useful question is: Does the current amount of randomness support the audience and experience we want? If a game feels unfair, one possible change is to reduce chance or give players more ways to respond to it. If a game feels solved or repetitive, adding uncertainty might create variety. [Sources] - Fullerton, *Game Design Workshop*, chapter 9, âThe Play Matrixâ: skill/chance and mental calculation/physical dexterity axes. - Image extracted from the project EPUB: figure 9.8, the play matrix. [/Sources]
Speaker notes: These are three common reasons to use randomness, but randomness can also create frustration if players feel they had no meaningful choice. Fullertonâs playcentric approach emphasizes player experience goals: designers should be able to state what they want players to feel and then test whether the game actually produces that experience. Rogers also warns against treating randomness as a substitute for design. Random outcomes should contribute to the experience rather than merely produce noise. When students add a random system, ask them to finish this sentence: âI am adding randomness because I want the player to feel ______.â If they cannot answer, the mechanic may not have a clear purpose yet. [Sources] - Fullerton, *Game Design Workshop*, chapter 6: player experience goals and intentional design. - Rogers, *Level Up!*, chapter 19: random systems still require design. [/Sources]
Speaker notes: Both Fullerton and Macklin/Sharp emphasize fast physical prototypes. Random systems are especially easy to test this way. If you are making a digital loot system, you do not need to write the loot code first. Put item names on cards or tokens, create the same odds you plan to use digitally, and test the experience at a table. Fullerton recommends preparing variations for playtests and changing as little as possible each time so you can understand the effect of the change. Gibson Bond similarly describes iteration as choosing a specific design goal, making a change, and testing again. Examples of one-variable tests: change a d6 to a d8; remove two strong cards; double one spinner sector; add one rare token; change a digital drop chance from 10% to 20%. [Sources] - Macklin & Sharp, *Games, Design and Play*, chapter 10: prototypes as playable questions; use dice, index cards, and physical objects early. - Fullerton, *Game Design Workshop*, chapter 9: prepare variations and change as little as possible each time. - Gibson Bond, *Introduction to Game Design, Prototyping, and Development*, chapter 7: iteration and reducing randomness when needed. [/Sources]
Speaker notes: Give students a few minutes to choose one event and design a randomization method for it. Require them to answer four questions: 1. What outcomes are possible? 2. Are the outcomes equally likely or weighted? 3. Does the random system remember earlier results? 4. What information can the player see? Then ask a fifth question drawn from the playcentric framework: What should this random system make the player feel? Students should be able to explain why they chose dice, cards, a spinner, a bag, or digital RNG instead of simply saying that one is easier. [Sources] - Fullerton, *Game Design Workshop*: player experience goals, formal systems, prototyping, and playtesting. - Macklin & Sharp, *Games, Design and Play*: chance, uncertainty, and prototypes as playable questions. [/Sources]
Speaker notes: These three questions summarize the deck. First: outcome space. What can actually happen? A d6 has six faces. A deck contains a known set of cards. A program can have an enormous or tightly limited set of results. Second: probability. Are outcomes equally likely? Are they weighted? Do the odds change after each event? Third: information. Does the player know the probabilities? Can they see what has already happened? Can they use that information to make a better decision? If students can answer these questions, they can analyze most random systems in games and make more intentional design choices. [Sources] - Synthesis of Macklin & Sharp, *Games, Design and Play*, chapter 2; Fullerton, *Game Design Workshop*, chapters 7 and 9; and Cermak-Sassenrath, *Introduction to Game Programming*, chapter 12. [/Sources]