Tip
If you’d rather not do the math yourself, a helpful calculator for computing win probabilities based on USCF Elo scores is available at www.bobnewell.net/nucleus/bnewell.php?itemid=279.
[I]f cause and effect aren’t possible, better that there at least be some reward for all the suffering.
—JEFFREY KLUGER
Novice designers’ games have a common theme—a heavy focus on randomness. Randomness itself is neither good nor bad. Randomness for the game designer is like salt for a chef. Sometimes you want to prepare something a bit salty. Sometimes adding salt only hurts the dish.
The “saltiest” games are those that are determined entirely without any intervention from the player. For these games, a combination of die rolls, card draws, or other random processes determine the games’ events and winners. LCR is a game that is completely determined by dice rolls. You cannot be “good” at LCR because no skill is involved in determining the winner.
The card game War is another prime example. In War, players split a deck of cards and then draw cards from the top and compare the faces. The player with the highest-ranked card collects the lower-ranked card from the opponent. Play continues until one player runs out of cards, or, more likely, until the players get tired of playing. Theoretically, and if a computer knew the decks of both players, it could tell the players who would win before the first card was turned over. In War, each player should win with a probability of 0.5, whether the player is a genius or a household pet.
Games with high randomness are usually suited for children with cognitive abilities that make a game with actual decision-making too taxing. Hi-Ho! Cherry-O, Snakes and Ladders, and Candy Land are popular children’s board games, and none require decision-making.
However, many high-randomness games are played by adults. One of the highest stakes casino games is Baccarat, which is a game that is completely random. The player makes a bet and then cards are drawn. The player and the house each get two cards. The hand value is the ones digit of the sum of the two cards. For instance, if the player draws a 7 and a 5, the hand’s value is 2 because 7 + 5 = 12 and the ones digit of 12 is 2. At certain times, the player or house draws a third card, but the algorithm is too needlessly detailed to cover here. At no point after the bet does the player get to make a decision. The player with the highest hand wins.
Famed sociologist Roger Caillois wrote in his book, Man, Play, and Games, that in games of pure chance “destiny is the sole artisan of victory, and where there is rivalry, what is meant is that the winner has been more favored by fortune than the loser.”1 It’s difficult for adults to feel satisfaction in games of pure chance between human players because the rivalry in the game was not settled by forces they themselves put in motion, but by the chance of a random selection.
1 Caillois, R. (1961). Man, Play, and Games. Champaign, IL: University of Illinois Press.
Blackjack is similar to Baccarat. Both games are dealer-versus-player interactions where the side with the highest hand total wins. However, Blackjack allows the simple decision for the player to press his luck by adding another card that may or may not make the player automatically lose by going over the maximum hand total of 21. This small change allows a player’s skill to affect the odds of his winning or losing and thus creates a more popular and meaningful game.
On the other end of the spectrum are games that are completely skill-based. In the purest skill-based games, the player with the higher skill level should always win. The world champion at Darts should always beat me. Chess is considered a highly skill-based game; however, even it is not deterministically skill-based. The Elo rating system that Chess and other games use is a rating of the probability that a player will win. You would expect that an expert-level Chess player would always beat a novice-level tournament Chess player. However, the Elo system predicts that, 3 percent of the time an expert with a rating of 2000 would lose to a novice rated 1400.
Tip
If you’d rather not do the math yourself, a helpful calculator for computing win probabilities based on USCF Elo scores is available at www.bobnewell.net/nucleus/bnewell.php?itemid=279.
Proponents of European-style strategy board games often favor games that lean heavily toward skill. One of the most popular European-style board games of all time, Settlers of Catan, uses the results of dice rolls to distribute resources. Some locations, such as those that pay out on a roll of 2, are probabilistically less likely to give resources, so they are less valuable than ones that pay out on an 8. The player who chooses a position that pays on 2 and gets greater than expected returns because of lucky rolls is often seen as an unfair result because the player is being rewarded not for strategy, but for random chance. This upset so many players that a popular variant uses a deck of cards with die-roll results printed on them in the expected frequencies (that is, one-sixth of cards have a 7 printed on them, while only one of the 36 cards has a 2). This variant mitigates the luck factor of choosing a territory that has a low probability of producing product.
An issue with a game that is entirely skill-based is that if two identical players play again and again, the game plays out the same way every time. The player with the higher skill level should always win. However, if there is no sense of mystery to the game, why play at all? If I play Magic: The Gathering with a professional player, I know my probability is low, but I still have a chance. If I competed in a sprint with an Olympian, I know I have no chance of winning. Why go through the motions?
Note
In college, I was paired against a professional Magic: The Gathering player at a prerelease tournament. His Elo was 1993. Mine was 1595. I won. It was the talk of the tournament. The Elo scores predicted that I had a 9-percent chance of winning, which makes the win seem less Cinderella-like.
In the book The Well-Played Game, Bernie De Koven writes “when the concept of fairness is spoken of in relation to playing games, it is used more as an emergency measure—a semimagical word which, when evoked, gives the utterer the chance to win, too.”2 Fairness is key to how players perceive randomness. If they believe that the randomness limits their chances at victory, they can view it as less fair. If they believe that the randomness enhances their chance at victory, they can view it as enforcing fairness.
2 De Koven, B. (2013). The Well-Played Game: A Player’s Philosophy. Cambridge, MA: MIT Press.
Sid Meier found that players felt that when they had a 1:3 army disadvantage in Civilization (that is, one unit for every three of the opponent’s), they should win 25 percent of battles.3 However, when players had a 3:1 advantage, they felt that their advantage was overwhelming and that they should never lose such a matchup, despite the magnitude of the differences being the same.
3 Graft, K. (2010, March 12). “GDC: Sid Meier’s Lessons On Gamer Psychology.” Retrieved December 8, 2014, from www.gamasutra.com/view/news/118597/GDC_Sid_Meiers_Lessons_On_Gamer_Psychology.php.
One way to design around the limitation of a preconceived victory (or defeat) is by limiting randomization to before the players are exposed to the field. This is the preferred design of many popular European-style games. For instance, the well-regarded game Agricola limits randomization to the initial occupation and improvement cards dealt to the players at the start of the game. The game also contains a random element to the order in which the actions of the game appear.
Some players dislike the random dealing of cards at the start of the game, so tournament Agricola players often use a drafting mechanic in which players take cards from a limited group of cards to smooth out the differences in power at the start of the game. Even if the cards were completely deterministic, the random deal of actions available to players causes games to play out in different ways. Since all players receive the effects of that randomness more or less evenly, it’s seen as fair, and no player can gain a large advantage by random chance alone.
FIGURE 11.1 shows two possible distributions for starting cards in Agricola. For simplicity, I’ve assigned a value of 1 to 10 for each card (higher numbers are more valuable), and I limited the number of starting cards to five. In a completely random setup, Zack is dealt cards with a value of 39, while Glo is dealt cards with a value of 20. Zack has a better chance of winning just by the luck of the draw because he has the more powerful cards.
IMAGES OF CARDS FROM AGRICOLA © 2006-2015 LOOKOUT GMBH. USED WITH PERMISSION.
FIGURE 11.1 A randomly dealt start can result in wildly different hands by luck alone.
FIGURE 11.2 shows the card distribution when drafting using the same cards and values. Both players look at the five cards they were each dealt and then each chooses the one with the highest value. The players then each pass the remaining cards from their dealt cards to the other player, who chooses the highest from the remaining four. The players repeat this process until all cards are drafted. The process self-balances and mitigates the randomness of the initial deal. Each player picks the best card from the ones remaining.
IMAGES OF CARDS FROM AGRICOLA © 2006-2015 LOOKOUT GMBH. USED WITH PERMISSION.
FIGURE 11.2 A more even distribution of cards can be achieved by drafting the cards.
At the start, Zack picks the Chamberlain (value 10), and Glo picks the Chief (value 6). Now they switch hands. Zack’s highest value choice is now the Hedge Keeper (value 6) and Glo’s is either the Seasonal Worker or the Chamberlain (both with value 9). At the end of the draft, Zack has a total value of 30 and Glo has a total value of 29, making the power of their hands nearly identical.
Many popular tabletop games are highly strategic while offering significant randomness. Spiel des Jahres winner, Dominion, involves getting a new random hand of five cards every turn. However, the player has a great deal of involvement in which cards go into her deck, so she is often changing the probabilities of what will be in her hand on the next turn.
Out There is a rogue-like space exploration game. Players need certain elements to repair their ship and jump to new sections. Yet, sometimes random events demand elements that the players don’t have or cannot get at their location. Thus, players feel as though they are being punished for something for which they could not prepare. FTL is a game with similar mechanics that mitigates this by having the difficulty centered on objectives that would be easier if the player had made certain earlier decisions. The difference is in the player’s reaction to random elements insofar as their perceived degree of agency. The perceived degree of agency is high in FTL and low in Out There. This is despite the prevalence of random events within FTL.
• Randomness itself is neither good nor bad. Instead, examine how the randomness in a game affects player choice and, thus, how it affects a player’s flow state.
• Although completely luck-based games often fail because complete randomness does not provide enough challenge for the player to enter flow, completely skill-based games can fail if the challenge required by the skill causes too much frustration by either being cognitively overwhelming or not providing any paths to victory.
• Often what matters is not the level of randomness inherent in a game’s mechanics, but rather how that randomness is perceived by the players.
• When players lose because of randomness, often they feel as though they had a lack of agency in the result. However, when they win because of randomness, they are less likely to object.
• Drafting is a mechanic that allows players to self-select resources to gain a more even distribution.
