Please enter the passkey to continue to the game instructions.
Incorrect passkey. Please try again.
In this game, you act as a retailer within a beer supply chain.
Each week, customers place orders for beer. Based on this demand, you must decide how many units to order from your supplier.
You only have access to local information, such as your current inventory, outstanding backorders, and incoming deliveries.
You do not know future customer demand, and all orders and deliveries are subject to delays.
For example the units you ordered from your supplier (the Wholesaler) will not always arrive immediately the week after,
but could arrive in 2-3 weeks instead or even partially distributed over several weeks.
Your decisions therefore affect the system with a delay, and their full impact may only become visible several weeks later.
Your objective is to minimize total costs over the course of the game. These consist of:
Total Cost = $200 + (Inventory x $0.50) + (Backlog x $1)
Example:
You have 300 beer units in your inventory and a backlog of 100 beer units.
$200 + (300 x $0.50) + (100 x $1) = $450 Total Costs
Balancing these two cost drivers is the central challenge of the game.
The game illustrates how coordination problems and delayed feedback can lead to inefficiencies in supply chains.
The dashboard provides an overview of your start-of-week situation (results from last week) to support this week's ordering decisions.
It displays, among other information:
The dashboard reflects only local information available to you as a retailer. It does not provide insight into future demand or the internal states of other supply chain members.
If not, please review the instructions once more before starting the game.
Thank you for taking part in this study.
We would now like to briefly explain the specific objectives of the research project.
This study investigates how people make decisions when they receive different levels of support from an AI system. We are particularly interested in whether and how various forms of assistance influence decision performance. Depending on the condition, participants either receive no support, a brief AI-generated summary of the current situation, or a more direct AI recommendation. Each participant experiences only one of these versions, but everyone performs the same underlying task.
Your participation helps us better understand how humans interact with AI in complex decision-making contexts. Thank you again for your contribution.
Please click the link below to complete the study:
Continue to Questionnaire