A Walk in the Forest to Discover the Bayesian Universe !
Game Flow
Game Flow
This article provides a detailed explanation of the gameplay mechanics of the game we previously introduced.
In the story we propose, the quality of the mushroom depends on the presence of a certain molecule. To convey the statistical aspect inherent in this type of measurement, at the beginning of the workshop, the student is invited to conduct an experiment and is asked to measure a sample of mushrooms.
With a click, the positive or negative presence of the molecule on a particular mushroom is revealed. The final proportion of mushrooms containing the desired molecule is calculated as the game progresses.
Initially, we offer a deterministic version where the desired molecule appears 100% of the time when the mushroom is good. This experiment can quickly become boring once it is understood that the result is entirely predictable. However, it becomes much less obvious when we restart the game in a probabilistic situation.
This time, statistically, the desired molecule appears two out of three times when the mushroom is good, and only one out of three times when the mushroom is bad. The number of positive and negative measurements being random, each game will yield a different result, with a proportion close to two-thirds. The game parameters have been pre-selected by us, and it is possible to vary the number of mushrooms or the desired proportion.
Then the main adventure begins!
Initially, the student plays solely based on their intuition. They explore the proposed universe and make their first measurements. The number of available measurements is limited, and it is important not to waste them in order to test enough species. Based on their own estimation, they can then recommend mushrooms that seem to be good. But be careful, this is all about probabilities, and the student will discover that even if they think they have a good chance of success, the outcome may not be as expected!
This is where we introduce a measurement tool that provides an estimation of the likelihood of the property "being a good mushroom." It is the Turing cursor, inspired by the method Turing used to decrypt the Enigma machine codes during World War II (see The Turing Cursor for more details) !
This tool will be extremely valuable in helping the student make the most appropriate decision and maximize their score. The decision-making time will also be significantly improved on average.
At the end of the workshop, the student is encouraged to question what a measurement is and exercise their critical thinking skills, having gained some experience in a probabilistic world.
For future workshops, we have added an additional subtlety to create more stakes in terms of the score. The player will be able to dissuade the authorities from choosing a mushroom for treatment if they believe there is a sufficient probability for it to be bad. If their estimation is correct, they receive a reward. However, if that category of mushroom was actually good for treatment, the player will lose points.
The visual part of this activity was created using artificial intelligence that generates an image from text. The possibility of generating this type of graphical representation was added to our platform during the summer of 2022 through the Stable Diffusion algorithm. But this story will be the subject of a future article ;-)