Jump to content

Motivation and emotion/Book/2026/Game of dice task and decision-making

From Wikiversity
Game of dice task and decision-making:
What does the game of dice task reveal about risk-based decision-making?

Overview

[edit | edit source]
Figure 1. Dice have clearly defined odds to roll a specific number, but how Emile decides to bet changes his chances of winning or losing.
Imagine this...

Your friend Emile has money to spend on a night out and decides the two of you should go to a casino. At a dice table, a card in front of you lists the rules:

  • The dealer will roll one die with six faces.
  • Players can bet on one specific number and double their money, but they have to put in at least $50.
  • Players can also bet on four numbers at once, winning if the die lands on just one of the four. This also doubles their money, but they can only put in up to $5.

The odds of winning are not shown, but they can be worked out (see Figure 1). For the first few rolls, Emile spends his money on cheap bets. After a couple of small wins, he switches to the single number bet. He loses. He goes for the single number again. He loses again. By the end of the night, despite clear odds the whole time, Emile had lost almost all of his money on the riskiest, least likely bets.

Why do you think Emile made risky choices? How would you choose to bet?

Some choices carry calculable probabilities, like the chances of side-effects from medical treatments, the probability of developing health issues from smoking or drinking, or indeed, the odds of winning at a gambling game (Brand et al., 2006). In theory, such choices can be calculated, guiding people towards optimal decisions. In reality, people often choose to make risky decisions even when they know their chances of success are low (Schiebener et al., 2014). If the odds are known, why do people still choose bets they are likely to lose?

The Game of Dice Task (GDT) was designed to study exactly this kind of decision-making. By making the rules clear and the probabilities calculable, it removes ignorance as a reason for poor choices (Schiebener et al., 2014). This chapter examines theory and research on the GDT, which suggests that executive functions (EFs) protect people from the motivational pulls of risky goals, rewards, and losses.

Focus questions

  • What is the Game of Dice Task?
  • How do self-regulation and executive functioning guide risk-based decisions?
  • What motivational influences pull people towards risky decisions and how do these relate to the Game of Dice Task?
  • What are the limits of the Game of Dice Task?
  • How can research on the Game of Dice Task be used to improve risk-based decision-making?

The Game of Dice Task

[edit | edit source]

The GDT is a computerised gambling task in which players try to maximise a fictional capital by betting on 18 rolls of a single die (Brand et al., 2005). It measures decision-making under risk, where an option can turn out worse than hoped for and the decision maker knows, or could know, how likely that is (Schiebener & Brand, 2015). This contrasts with the Iowa Gambling Task (IGT), for instance, where players are not given clear odds and must learn which choices are advantageous as they play (Bechara & Damasio, 2005).

Before each roll, players bet on one, two, three, or four numbers. If the number rolled is in a player's bet, they win the associated amount. Otherwise, they lose that same amount. These rules are explained at the start of the game and never change. Players complete the GDT on a single computer screen (see Figure 2). After a bet is selected, a die rolling video is played. Feedback is given by displaying the amount the player has just won or lost, adjusting the total balance number and chart, and playing an associated win or loss sound. [factual?]

Figure 2. The typical presentation of the GDT, adapted from Brand et al. (2005). Clockwise from top left are: Die rolling video, most recent gain/loss with current total balance below, balance chart for a visual representation of current balance, round number, quit button, gains/losses associated with the possible bets, and possible bet choices.

The one and two number bets are risky and disadvantageous, since they win less than half the time, while the three and four number bets are advantageous, since they win half the time or more (Brand et al., 2005). Riskier bets win or lose larger amounts (see Table 1). The probabilities are not stated, but can be calculated from the rules. Overall score is not a player's final balance, but the number of advantageous bets minus the number of disadvantageous bets, ranging from -18 to +18. This makes results comparable regardless of luck.

Table 1. The four bets in the GDT and their associated odds. Expected value calculated using the amounts in Brand et al. (2005).
Bet Chance of winning Amount lost/gained Expected value Category
1 number 1 in 6 (16.67%) $1,000 - 667 Disadvantageous, most risky
2 numbers 2 in 6 (33.33%) $500 - 167 Disadvantageous, risky
3 numbers 3 in 6 (50%) $200 0 Advantageous, less risky
4 numbers 4 in 6 (66.67%) $100 + 33 Advantageous, least risky

Self-regulation and executive functions

[edit | edit source]

The GDT is primarily intended to measure how EFs impact risk-based decision-making. Hofmann et al. (2012) describe three EFs that are core to maintaining control over behaviour, or self-regulating: updating, which holds and refreshes information in working memory; inhibition, which overrides automatic responses; and shifting, which moves attention between tasks and rules. Self-regulation requires goals and the executive capacity to pursue them. This executive capacity may be weakened by stress or distractions, leading behaviour to follow impulses rather than rationality (Hofmann et al., 2012).

From somatic markers to a dual-process model

[edit | edit source]

Assuming that people will always make the best choice when the odds are known or calculable is assuming that they have "unlimited knowledge, time, and information-processing power" (Bechara & Damasio, 2005, p. 336). The somatic marker hypothesis proposes that automatic bodily signals like a quickening heart rate guide decisions before they are reasoned through, particularly when the odds are hidden (Bechara & Damasio, 2005). Brand et al. (2006) proposed that under explicit risk, EFs hold payoffs and probabilities in working memory, categorise the options by risk, and form a strategy. Somatic markers may be helpful, but are supposedly unnecessary [explain?].

Figure 3. A simplified version of the dual-process model of decision-making under risk, in which feedback from each decision can change whether the impulsive or reflective system leads. Adapted from Schiebener and Brand (2015).

Schiebener and Brand (2015) later revised this into a dual-process model in which an impulsive system of emotions and automatic responses runs alongside a reflective system of EFs and reasoning, with feedback changing which leads (see Figure 3). The model was developed as a post hoc explanation of how choices in explicit gambling tasks like the GDT happen, limiting its predictive validity.

Evidence for EFs in the GDT

[edit | edit source]

Higher GDT performance has been found in healthy adults who are better at inhibition and shifting, although all the studies reviewed were conducted in Germany (Colautti et al., 2022). People with Parkinson's disease made riskier choices than controls, with self-regulatory EFs linked to GDT performance in two out of three studies (Colautti et al., 2023). Other groups with executive impairments, such as Korsakoff syndrome and preadolescent ADHD, also make more risky choices despite normal IQ (Schiebener & Brand, 2015). Among 152 adults, general executive control was the only direct predictor of performance, but other EFs contributed through it (Schiebener et al., 2014).

In a study where 30 participants played the GDT with no feedback elements, they chose the riskiest bet more often and switched between risky and safe bets about twice as often as in the standard GDT (Brand, 2008). A purely calculating player would bet the same way in both versions, so players appear to use feedback to check and settle on a strategy. Further GDT research suggests this depends on EFs, though this may be because the GDT was originally designed to examine EFs. People with Korsakoff syndrome played equally badly with or without feedback, and after losing on a risky bet, only a third switched to a safer bet, compared with two thirds of healthy players (Brand et al., 2009).

Impairing EF also affects healthy players, where those who completed a 2-back working memory task while playing the GDT made more risky choices than those doing a 1-back task or just the GDT (Starcke et al., 2011). Conversely, calculating the odds before playing was correlated with safer bets (Pertl et al., 2017), possibly since options and strategies could be evaluated outside of feedback influences. Self-regulation therefore appears necessary for advantageous play.

Motivational pulls to risk in the GDT

[edit | edit source]

If self-regulation is necessary to hold back impulses, why are risky choices creating these motivations in the first place? Three theories can help explain what pulls players towards risky bets: the feelings associated with reward, sensitivity to rewards and punishments, and reference points and goals used to judge outcomes.

Risk-as-feelings

[edit | edit source]

If people only weighed the outcomes by which was more likely, players of the GDT would only pick the four number bet. The risk-as-feelings hypothesis [factual?] offers an explanation for why rewards make risky choices appealing. When outcomes carry strong feelings, people are more likely to compare the outcomes alone and disregard their probabilities (Pachur et al., 2014). The $1,000 reward [explain?] in the GDT may therefore feel appealing to people who are excited by the large payout, no matter its odds, aligning with Brand et al.’s (2006) suggestion that pathological gamblers found the large payouts of risky bets too appealing to resist. However, because larger payouts in the GDT come with low odds, no study has directly tested the pull of reward size in the task.

Reinforcement sensitivity

[edit | edit source]

Reinforcement sensitivity theory proposes that the behavioural approach system (BAS), activated by signs of rewards, produces desire and excitement. The behavioural inhibition system (BIS), activated by conflict within the BAS, produces caution and risk assessment (Corr et al., 2013). Bets in the GDT could be seen as creating conflict, since they have a prize to approach but an equal loss to avoid. People with a strong BAS should be drawn to the risky bets, especially after wins, while a strong BIS should make them pick the safer bets. Unlike Schiebener and Brand's (2015) impulsive-reflective model, reinforcement sensitivity theory describes personality components, and the BIS detects conflict and produces anxiety automatically rather than disrupting impulsive behaviours as a result of EFs (Corr et al., 2013).

In a sample of 83 students who played a modified GDT that produced winning or losing streaks, the 20 participants high in reward sensitivity and low in punishment sensitivity made riskier choices after a winning streak, while other participants became safer. The 19 participants with low reward and high punishment sensitivity did the opposite, making safer choices during a losing streak, while other participants became riskier (Kim & Lee, 2011). This aligns with the predictions that a high BAS draws people's motivations to rewards and away from probability. However, the effects and sample were small, and were contradicted in a sample of 1,002 older adults, where higher reward sensitivity predicted less risky betting in the GDT (Sinclair et al., 2021).

Common questionnaires such as the BIS/BAS scales may not be reliable at separating these systems (Corr et al., 2013) and Sinclair et al. (2021) suggest reward sensitivity may appear in the GDT as exploring options and seeking feedback rather than going after rewards. Links between personality, motivation and explicit gambling tasks are rare and inconsistent, with only some impulsive traits showing associations (Schiebener & Brand, 2015).

Prospect theory

[edit | edit source]

Prospect theory suggests that people judge outcomes as gains or losses compared with a reference point, rather than by their final total [factual?]. It also proposes that people loss averse, feeling loss more intensely than gain, and framing effects influence people to change their behaviours based on how a problem is presented (Tversky & Kahneman, 1981). Pabst et al. (2013) tested this with gain-only and loss-only versions of the GDT. In gain-only, players started at zero and rolls either won money or nothing, whereas in loss-only players started at $1,000 and either lost money or nothing.

Betting on one number had the lowest chance of winning in both, but in loss-only, the single number had the best expected value (see Table 2). Participants facing losses made more disadvantageous choices than those facing gains, suggesting they were focused on avoiding loss instead of calculating the values. Each group's starting reference points may have also changed their betting patterns, since any loss was falling short of the reference point, while any gain was exceeding it. This also aligns with risk-as-feelings, where the more emotionally impactful effects of the loss led to disregarding the probabilities (Pachur et al., 2014).

Table 2. The odds and values in the gain-only (+) and loss-only (-) GDTs (Pabst et al., 2013).
Bet Chance of winning / losing Amount gained / lost Expected value (+ and -) Category
1 number + 1 in 6 (16.67%)

- 5 in 6 (83.33%)

+ $600

- $120

$100 + Disadvantageous

- Advantageous

2 numbers + 2 in 6 (33.33%)

- 4 in 6 (66.67%)

+ $400

- $200

$133 + Disadvantageous

- Advantageous

3 numbers + 3 in 6 (50%)

- 3 in 6 (50%)

+ $333

- $333

$166 + Advantageous

- Disadvantageous

4 numbers + 4 in 6 (66.67%)

- 2 in 6 (33.34%)

+ $300

- $600

$200 + Advantageous

- Disadvantageous

Goals may also act as reference points, where outcomes below a goal feel like losses. In the GDT, showing players a fake leaderboard increased risky betting despite heavy losses (Schiebener et al., 2012). The leaderboard showed scores up to $14,500, far above the real best of $3,500. Players who set higher goals made more single-number bets since they were the only way to approach such scores. Setting goals against the leaderboard reduced risk in players with strong EFs and increased it in those with weaker EFs, possibly because stronger EFs help revise unrealistic goals. Falling behind may also explain why pathological gamblers become riskier late in the game, when the riskiest bets were the only way to make up for earlier losses (Brand et al., 2006).

Check your understanding

1

The Game of Dice Task measures decision-making under risk, which is calculable probability.

True
False

2

According to Schiebener and Brand's (2015) dual-process model, what helps players avoid risky bets?

Somatic markers signalling danger
The reflective system, which relies on executive functions to hold and apply a strategy
Reward sensitivity

3

Whether people receive feedback has nothing to do with how they make decisions while playing the GDT.

True
False

4

Why did players make more single-number bets after seeing a leaderboard with very high scores?

The leaderboard explained the odds of each bet
High scores reduced their motivation to win
The scores acted as a reference point, so falling short felt like a loss that only risky bets could make up

Limits of the Game of Dice Task

[edit | edit source]

Several limits apply to the GDT. The GDT was designed primarily as a measure of EFs under explicit risk, so there is little research directly linking it to risk-based decision-making and motivation. Within its design, riskier bets in the GDT always have worse average values, so the task cannot separate a player who seeks risk from one who has not recognised the advantageous bets. The standard GDT also uses fictional money, which may weaken the reward and punishment signals the motivational theories rely on and encourage calculation (Pachur et al., 2014), so the task may understate how feelings drive real gambles. Further to this, the rolls are pseudorandomised so that each number appears exactly three times (Brand et al., 2005). Once a number is rolled, its chance of appearing next drops from 1 in 6 to 2 in 17, so the gambler's fallacy becomes an advantageous strategy in later trials. Finally, most evidence comes from small, mostly German samples (Colautti et al., 2022), and the task predicts real behaviour poorly. Among 226 adolescent gamblers, GDT scores did not distinguish the problem gamblers from others (Donati et al., 2019).

Improving risk-based decisions

[edit | edit source]

Despite these limitations, GDT research suggests several strategies for decisions where the odds are known, such as managing money, choosing medical treatments, or gambling. Working out the odds before deciding, and deciding away from distractions, reduces the load on self-regulation (Pertl et al., 2017; Starcke et al., 2011). Setting realistic goals rather than comparing against other people's results or earlier losses may reduce the drive for achievement that makes rewarding risks appealing (Schiebener et al., 2012).

These findings may be relevant for people with conditions that affect EFs, such as ADHD or Parkinson's disease (Colautti et al., 2023; Schiebener & Brand, 2015). Because several groups who struggled with the GDT had normal general intelligence, risky choices may reflect difficulty holding a strategy against the pull of rewards and losses, rather than a lack of intelligence or a personal failing. Identifying choices and thinking through their possible outcomes before a decision may help.

Conclusion

[edit | edit source]

The GDT was designed to test whether self-regulation guides decisions once the odds are known. Reviews and experiments suggest that it does, since people with stronger EFs make fewer risky bets, and loading working memory makes choices riskier. However, there is a tension between the reflective system of cognition and the impulsive system of feelings described in Schiebener and Brand's (2015) model. The effects of EF, framing, and goals have the strongest evidence in the GDT, while reward sensitivity evidence is conflicting. The main takeaway from the GDT is that people are not purely probabilistic decision-makers. Even when the odds are knowable, risk-based decisions depend on people's goals, their past experiences, their responses to feedback, and their capacity to maintain an effective strategy. Recognising these pulls can help [missing something?]

See also

[edit | edit source]

References

[edit | edit source]
Bechara, A., & Damasio, A. R. (2005). The somatic marker hypothesis: A neural theory of economic decision. Games and Economic Behavior, 52(2), 336–372. https://doi.org/10.1016/j.geb.2004.06.010

Brand, M. (2008). Does the feedback from previous trials influence current decisions? A study on the role of feedback processing in making decisions under explicit risk conditions. Journal of Neuropsychology, 2(2), 431–443. https://doi.org/10.1348/174866407X220607

Brand, M., Fujiwara, E., Borsutzky, S., Kalbe, E., Kessler, J., & Markowitsch, H. J. (2005). Decision-making deficits of Korsakoff patients in a new gambling task with explicit rules: Associations with executive functions. Neuropsychology, 19(3), 267–277. https://doi.org/10.1037/0894-4105.19.3.267

Brand, M., Labudda, K., & Markowitsch, H. J. (2006). Neuropsychological correlates of decision-making in ambiguous and risky situations. Neural Networks, 19(8), 1266–1276. https://doi.org/10.1016/j.neunet.2006.03.001

Brand, M., Pawlikowski, M., Labudda, K., Laier, C., von Rothkirch, N., & Markowitsch, H. J. (2009). Do amnesic patients with Korsakoff’s syndrome use feedback when making decisions under risky conditions? An experimental investigation with the Game of Dice Task with and without feedback. Brain and Cognition, 69(2), 279–290. https://doi.org/10.1016/j.bandc.2008.08.003

Colautti, L., Antonietti, A., & Iannello, P. (2022). Executive functions in decision making under ambiguity and risk in healthy adults: A scoping review adopting the hot and cold executive functions perspective. Brain Sciences, 12(10). https://doi.org/10.3390/brainsci12101335

Colautti, L., Iannello, P., Silveri, M. C., & Antonietti, A. (2023). Decision-making under ambiguity and risk and executive functions in Parkinson’s disease patients: A scoping review of the studies investigating the Iowa Gambling Task and the Game of Dice. Cognitive, Affective, & Behavioral Neuroscience, 23(5), 1225–1243. https://doi.org/10.3758/s13415-023-01106-3

Corr, P. J., DeYoung, C. G., & McNaughton, N. (2013). Motivation and personality: A neuropsychological perspective. Social and Personality Psychology Compass, 7(3), 158–175. https://doi.org/10.1111/spc3.12016

Donati, M. A., Frosini, A., Izzo, V. A., & Primi, C. (2019). The effectiveness of the Game of Dice Task in predicting at-risk and problem gambling among adolescents: The contribution of the neural networks. Journal of Gambling Studies, 35(1), 1–14. https://doi.org/10.1007/s10899-018-9796-5

Hofmann, W., Schmeichel, B. J., & Baddeley, A. D. (2012). Executive functions and self-regulation. Trends in Cognitive Sciences, 16(3), 174–180. https://doi.org/10.1016/j.tics.2012.01.006

Kim, D-Y., & Lee, J-H. (2011). Effects of the BAS and BIS on decision-making in a gambling task. Personality and Individual Differences, 50(7), 1131–1135. https://doi.org/10.1016/j.paid.2011.01.041

Pabst, S., Brand, M., & Wolf, O. T. (2013). Stress effects on framed decisions: There are differences for gains and losses. Frontiers in Behavioral Neuroscience, 7, Article 142. https://doi.org/10.3389/fnbeh.2013.00142

Pachur, T., Hertwig, R., & Wolkewitz, R. (2014). The affect gap in risky choice: Affect-rich outcomes attenuate attention to probability information. Decision, 1(1), 64–78. https://doi.org/10.1037/dec0000006

Pertl, M-T., Zamarian, L., & Delazer, M. (2017). Reasoning and mathematical skills contribute to normatively superior decision making under risk: Evidence from the game of dice task. Cognitive Processing, 18(3), 249–260. https://doi.org/10.1007/s10339-017-0813-x

Schiebener, J., & Brand, M. (2015). Decision making under objective risk conditions—a review of cognitive and emotional correlates, strategies, feedback processing, and external influences. Neuropsychology Review, 25(2), 171–198. https://doi.org/10.1007/s11065-015-9285-x

Schiebener, J., Wegmann, E., Gathmann, B., Laier, C., Pawlikowski, M., & Brand, M. (2014). Among three different executive functions, general executive control ability is a key predictor of decision making under objective risk. Frontiers in Psychology, 5, Article 1386. https://doi.org/10.3389/fpsyg.2014.01386

Schiebener, J., Wegmann, E., Pawlikowski, M., & Brand, M. (2012). Anchor effects in decision making can be reduced by the interaction between goal monitoring and the level of the decision maker’s executive functions. Cognitive Processing, 13(4), 321–332. https://doi.org/10.1007/s10339-012-0522-4

Sinclair, C., Eramudugolla, R., Brady, B., Cherbuin, N., & Anstey, K. J. (2021). The role of cognition and reinforcement sensitivity in older adult decision-making under explicit risk conditions. Journal of Clinical and Experimental Neuropsychology, 43(3), 238–254. https://doi.org/10.1080/13803395.2021.1909709

Starcke, K., Pawlikowski, M., Wolf, O. T., Altstötter-Gleich, C., & Brand, M. (2011). Decision-making under risk conditions is susceptible to interference by a secondary executive task. Cognitive Processing, 12(2), 177–182. https://doi.org/10.1007/s10339-010-0387-3

Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211(4481), 453–458. https://doi.org/10.1126/science.7455683

[edit | edit source]