The Wrong Call: What MBA Students Keep Getting Wrong Inside Business Simulations

Research spanning decades of classroom data reveals patterns of error that mirror real executive failures — and why that may be exactly the point.

The teams had all the data in front of them. Market share reports, cost structures, competitor pricing, demand forecasts. They had studied strategy for months. And yet, as researchers at a Taiwanese university watched 245 working MBA students navigate the People Express business simulation game, the same sequence unfolded again and again: students defaulted to intuition over analysis, doubled down on failing strategies, and allowed the loudest voice in the room to steer decisions that the evidence did not support.

That study, published in 2025 in the journal Studies in Educational Evaluation, offered one of the clearest recent portraits of what actually happens when graduate business students are placed inside a simulation and forced to make real choices under pressure. The conclusion was not that simulations fail students. It was something more instructive: that simulations reveal, with uncomfortable precision, the gaps between what students know and what they do.

These tools have become one of the most widely deployed instruments in graduate management education. Harvard Business Publishing alone distributes more than 200,000 subscriptions annually to institutions around the world. A McKinsey study found that students enrolled in digitally centered experiential programs achieve up to 35 percent higher retention and skill application rates compared to traditional teaching methods. A 2022 systematic literature review in Human Behavior and Emerging Technologies, which followed PRISMA guidelines and analyzed 57 empirical papers published between 2015 and 2022, concluded that business simulation games improve knowledge acquisition, cognitive skills, and interactive competencies across higher education settings.

Yet widespread adoption has not eliminated a persistent set of behavioral tendencies that scholars have documented across institutions, disciplines, and geographies. These are not random errors. They are systematic ones, shaped by cognitive habits that students carry into simulations from their prior experience, their personality, and the structure of group decision making itself. Understanding what those errors are — and why they are so consistent — is the first step toward designing learning experiences that actually correct them. If you want to understand how business simulations work and why they matter in higher education, the patterns documented below are essential reading.

The Illusion of Competence

Among the most studied phenomena in business simulation scholarship is overconfidence: the tendency of students to rate their own judgment more highly than the evidence warrants.

This dynamic appears early and consistently. A paper on cognitive bias in executive decision making published in the journal Electronics in 2025 describes overconfidence as a condition in which an individual's subjective confidence surpasses their objective accuracy. Inside simulations, this typically manifests in the opening rounds: teams that perform reasonably well early on often attribute their results to strategic insight rather than to the structure of the game, which is deliberately designed to reward conservative, foundational decisions at the start.

The reckoning arrives later. When market conditions shift, overconfident teams hold their original positions longer than the evidence recommends. They interpret early success as proof that their framework is correct, rather than as a feature of a controlled environment built to let every team find its footing.

Harvard Business School professor Max Bazerman has found that firms implementing systematic debiasing strategies saw an average increase of 7 percent in return on assets — a figure that suggests the cost of overconfidence is not academic. It is measurable, and it compounds.

The Trap of Sunk Costs

If overconfidence shapes the opening behavior of MBA simulation teams, the sunk cost fallacy tends to define their middle game. The phenomenon is well documented in behavioral economics and appears with striking regularity in educational settings.

A study published in Experimental Economics and cited in a 2024 proceedings paper from the SHS Web of Conferences found that 70 percent of individuals continued to invest resources in a failing proposition simply because they had already committed to it. The money had already been spent. It could not be recovered. And yet the prior commitment reshaped how participants assessed what to do next.

Inside the simulation, this bias takes a recognizable form. A team allocates a significant share of its budget to a marketing campaign in round two. By round four, the data shows the campaign is underperforming. The rational action is to reallocate. The common one is not: teams that have publicly committed to a strategy — particularly when that commitment was contested and then defended — are significantly more likely to increase rather than reduce their exposure to a failing position.

A review of cognitive biases in organizational decision making published by the UK government's Infrastructure and Projects Authority noted that optimism bias, which causes decision makers to overweight their odds of success, functions as a causal component of sunk cost behavior. The two reinforce each other: students who are confident their strategy will eventually perform are more likely to hold it as contrary evidence accumulates.

A 2016 laboratory study published in the Journal of Economic Psychology added a counterintuitive dimension to this picture: cognitive ability did not reliably reduce susceptibility to the sunk cost fallacy. Researchers found strong evidence of the bias among high ability subjects, suggesting that intelligence alone does not protect against this class of error.

When the Room Decides

Business simulations are almost universally conducted in teams. The pedagogical reasoning is sound: real business decisions are rarely made by individuals in isolation, and learning to function inside a group under pressure is a core competency of management education.

But collective settings introduce their own liabilities — ones that scholars have traced with precision.

A longitudinal field study published in the journal Small Group Research examined 18 business process reengineering projects conducted by 71 first-year MBA students over a three-month period. The study found that groupthink concurrence-seeking behavior had a measurable negative impact on group performance. The effect was not incidental. It was structural: when teams developed strong internal cohesion early on, they became less likely to surface dissenting information and more likely to accelerate toward consensus regardless of whether that consensus was well grounded.

Kayes, Kayes, and Kolb, writing in Simulation and Gaming in 2005 in a study drawing on nearly four decades of scholarship on experiential learning in teams, described the mechanism directly: smaller teams with members who share similar attitudes tend toward high cohesion, and teams with too much cohesion are liable to suffer from groupthink — a flawed decision process in which members move to action without adequately considering the full range of available evidence.

A separate study of 61 MBA project teams at a leading North American university, published in Organizational Behavior and Human Decision Processes, produced a parallel result: when teams attempted to pursue exploration and exploitation simultaneously in the same working session, outcomes deteriorated. The activities themselves were not harmful in isolation. Compressed into a single episode rather than distributed across time, they generated dissonant effects that damaged performance.

Short Rounds, Short Thinking

A defining structural feature of most business simulations is the round: teams make decisions for a simulated quarter, receive feedback, and advance to the next period. The design is intentional, creating natural checkpoints for reflection and allowing instructors to observe behavioral shifts over time.

Evidence from the field suggests that students frequently respond to this structure in ways the design did not anticipate. Because each round has a discrete end point and immediate consequences, many teams optimize for period over period results rather than for strategic positioning across the full arc of the game. They price aggressively to capture market share in the current round without modeling the margin implications two periods forward. They cut research and development spending to improve near term profitability, without accounting for the competitive position that spending was meant to protect.

The 2025 Electronics paper noted that executives in real organizations similarly shift strategies impulsively in response to immediate market signals, resulting in erratic strategic trajectories. The implication for classrooms is pointed: this is not simply a corporate habit that experience corrects. It is a cognitive orientation that students carry into the room, and that structured simulations surface without necessarily dismantling.

A paper examining Executive MBA participation in business strategy simulation, published on ResearchGate, revealed a consistent thread: although students' knowledge and strategic competencies expanded through the exercise, their decision making styles were not significantly influenced by practice. The analytical component showed improvement. But the broader orientation toward how decisions are made — including the time horizon over which consequences are evaluated — showed little movement.

The Data Left Unread

Simulations generate substantial quantities of information. Scorecards, market reports, competitor analyses, financial summaries. The information is available. It is often not engaged with carefully, or at all.

The 2022 systematic review in Human Behavior and Emerging Technologies found that student reflection reports consistently identified the same source of regret in postgame evaluations: decisions made without adequate engagement with available analytical materials. Students noted, after the fact, the specific choices they wished to revisit, and underuse of provided data appeared repeatedly across institutions and course formats.

The study of 100 business administration teams referenced in research published on ResearchGate approached the same problem from a different angle. Comparing the number and type of mistakes in the first half of a simulation with those in the second half, researchers found that simple decision making skills improved across rounds. Complex, dynamic decision making did not improve at the same rate. The implication was significant: the kinds of errors that simulations are designed to surface — those involving interdependent variables, long feedback loops, and compound tradeoffs — are precisely the ones that students find most resistant to correction.

What Failure Is For

The scholarly consensus on business simulations is not that they fail to teach. It is that their teaching depends heavily on what happens after the mistake is made.

Kayes and colleagues, writing in Simulation and Gaming, described the mechanism by which teams that intentionally focus on learning rather than on winning can overcome the social loafing, groupthink, and diffusion of responsibility that group settings tend to produce. The structural key was not the error itself but the reflection that followed: teams that created deliberate space to examine what had happened and why outperformed those that moved directly from one decision period to the next.

Harvard Business Publishing's distribution of more than 200,000 subscriptions annually reflects an institutional conviction that this model works. The 2022 review conducted under PRISMA protocols, covering 57 empirical studies, found consistent evidence that it does — at least when the exercise is paired with structured debriefing and formative assessment. The McKinsey figure of 35 percent higher retention in experiential programs points in the same direction.

But the behavioral tendencies documented across this body of literature are also a clear signal for students who approach simulations as competitions to be won rather than as processes to be understood. The teams that finish in the top quartile of most business simulations are not the ones that avoided all mistakes. They are the ones that read the data, surfaced the dissent, named the sunk cost for what it was, and adjusted.

The most consequential lesson embedded in a business simulation may not be strategic. It may be epistemic: understanding how you are likely to be wrong, and building the habits that make it possible to find out before the cost is real.

For MBA students and business faculty looking to put these dynamics to the test in a rigorous, structured environment, EXSIM by Eureka Simulations offers one of the most complete platforms available for that kind of experiential learning. Designed for higher education and corporate training, EXSIM places participants inside realistic competitive markets where every decision — from pricing and production to team coordination and resource allocation — carries visible consequences across subsequent rounds. The platform is built precisely around the gap the research identifies: the distance between knowing a concept and performing it under pressure. Whether inside a classroom or a corporate training program, it is the kind of environment where the errors described in this article do not stay theoretical for long.

Sources: Studies in Educational Evaluation (2025); Human Behavior and Emerging Technologies (2022); Simulation and Gaming (2005); SHS Web of Conferences (2024); Journal of Economic Psychology (2016); Electronics (2025); UK Infrastructure and Projects Authority literature review; ResearchGate / MBA strategy simulation research; Harvard Business Publishing; McKinsey & Company.