01Round one is an experiment, not a move
Move one lever and hold everything else still. Price, or advertising, or capacity, but only one. Then read what moved in response and by roughly how much. You are not trying to win round one. You are estimating how the model behaves, and a single clean round teaches you more than three ambitious ones that changed six things at once and produced a result nobody can attribute.
Say the cost out loud to your team before you do it, because somebody will object. You may finish round one behind, and being behind in round one is worth very little. What it buys is a model you can predict for the rest of the game, which is worth a great deal both on the leaderboard and in the report you eventually have to write.
02The movement between two statements is the lesson
Every round you get a fresh income statement and balance sheet, and the useful object is not either one on its own. It is the difference between this round's and last round's, read in a fixed order so that nothing gets skipped in the rush to enter decisions.
- Revenue against units sold, which separates a price effect from a volume effect
- Gross margin against unit cost, which tells you whether volume was bought with margin
- Operating expense lines against the decisions that funded them, one line at a time
- Inventory or unused capacity against demand, which is where simulated firms quietly bleed
- Cash last, because it is the consequence of everything above and explains none of it
03The variance note, written before the result
Before decisions lock, write down what you expect to happen and why, in three or four sentences. After the round closes, write what actually happened and where the two parted company. That gap is the only genuinely original thing a simulation produces, because everyone gets the same statements and nobody else has your prediction.
It is also, conveniently, most of the final report. Rubrics on these capstones score reasoning far more heavily than ranking, so a report that reconstructs decisions from memory in the last week reads thin no matter how well the firm performed. A report assembled from eight variance notes written while the memory was fresh reads like analysis, because it is.
Keep the notes short and keep them dull. Three sentences of expectation, three of outcome, one naming what you will change. A note nobody can face writing after a bad round is a note that stops existing in round five, which is exactly the round you will most want to explain later.
04Two failures that repeat every term
The first is chasing the leaderboard. Teams optimise for a rank that carries modest weight and neglect rubric rows worth more than the rank, then write a report explaining a performance nobody was grading that closely. Read the assessment document before round one and work out which rows the simulation actually feeds, because some of them it does not feed at all.
The second is one member holding the decision screen. It is efficient for four rounds and produces a report the rest of the team cannot write, because they never saw the statements move. Rotate who enters decisions, keep the variance notes in a shared file that everybody writes into, and the report ends up with four authors who each understand the same firm.
Where a capstone carries a written component alongside the simulation and the two have to agree with each other, that is a project rather than an assignment, and it runs under capstone work.
FAQQuestions to control
Is there a strategy that generally works?
Consistency beats cleverness in most of these models, because they reward decisions that fit together over rounds rather than a single strong move. Pick a position you can hold, fund it properly, and avoid reversing course in consecutive rounds, which usually costs more than either direction would have earned. The model punishes contradiction, and contradiction is what panic produces.
What if my team is losing badly by round four?
Read the statements before changing anything. A firm losing on margin needs a different correction from a firm losing on volume, and the two look similar on a leaderboard. Then change one thing, not everything, so the recovery is attributable in your report. A documented recovery from a bad position often reads better in an assessment than an untroubled run nobody had to explain.
How much of the grade is the simulation result?
Usually less than teams assume, and your assessment document says. Most of these capstones weight the analysis, the written report and sometimes a presentation above the outcome itself, precisely because outcomes depend partly on what other teams do. Read the weighting in week one, because it determines how much of your effort belongs in the model and how much in the writing.
Can a specialist run the simulation for me?
The decisions and the reasoning have to be yours, since the write-up is an examination of your understanding and a viva or presentation will find any gap immediately. What genuinely helps is the accounting layer alongside you: reading each round's statements, naming which decision produced which movement, and building the variance notes into a report that holds together across eight rounds.
