A game for the year you first lead a project: a supervisor with no time, a senior collaborator with many projects, a junior you are responsible for, and a deadline. Every decision moves six hidden dials; the classic problems (time pressure, authorship fights, silent collaborators, dropped balls, burnout) fire when the dials say so, and most can be prevented or repaired. The debrief traces every problem back to the decision that seeded it.
The situations are drawn from lab-culture discussions and from the Path to PhD newsletter; which move is “sound” is one researcher's judgment, and every reveal gives its reasons.
It assumes no experience of running anything, and no vocabulary either: the principal investigator (PI), author order, what counts as noise. Every term the game uses has a definition on hover and a chip under the card it appeared on, and the field manual carries the whole glossary and the reading behind each problem.
You are a second-year PhD student, and for the first time the project is yours to run. The idea came out of a workshop conversation: a method your lab has been circling for a while, with a theory part and an experimental part. Your supervisor likes it, a professor at another institute wants in, and the target is a conference deadline in month 8. Nobody has written down who does what.
Nine months, two or three decisions a month. Each decision moves six dials you cannot see directly (alignment, trust, buffer, credit, energy, feedback loop), and the classic problems fire when the dials say so: a collaborator goes quiet, an authorship question comes up late, a deadline lands on a week with no slack. Most can be prevented, and most can be repaired if you notice in time. At the end, a debrief traces every problem back to the decision that seeded it.
Four people matter. Their personalities are drawn fresh each run, and they change which move is sound: a supervisor who wants to be asked and one who trusts you to decide need different messages; a student who hides being stuck needs a different kind of attention from one who asks. When a decision depends on the person, the reveal says so, and says what would change with the other kind.
The game carries two versions of each person; real people are more than a trait, and reading them is part of the job.
Nothing here assumes you already know the vocabulary. A term with a dotted underline has a definition on hover, and the chips under each card repeat the ones that card used; the field manual, in the top bar, carries all of them.
Every decision moved them; the red marks are the months a problem fired. Hover a line for its name.
Problems that fired, problems you prevented, and the decisions behind each.
Descriptive, not graded. Where your choices sat on five spectra, this run (hollow) and across all runs on this device (filled). Each pole has a virtue and a vice.
The hard moments of the same project the run puts you in: the collaborator who has gone quiet, the authorship conversation nobody opens, the reminder to a supervisor who has not replied, the result that died the week before the deadline. Each one gives you two messages and asks which you would send.
Neither is a caricature. The weaker message is polite and well meant and leaves out something the other person needs: the reason, the date, the ask, or a way to answer in one line. Length does not predict the answer, so read for what the message carries rather than for how long it is. After you pick, the reveal names the feature that made the difference and the principle behind it.
A round is eight messages, the ones you have not seen first. You are the same person you are in the run: a second-year PhD student running a project for the first time, with a deadline in month eight.
A few messages go to people outside the project: a postdoc down the hall, a professor visiting for two days. Each of those says who they are.
Two messages for the same moment. Which one would you send?
Most management problems are communication problems. The difference between the two messages was rarely tone. It was whether the message carried context, a concrete ask, a date, and a way for the other person to say no.
Across every run on this device. Play twice before reading much into it.
Every problem in the game. Open one for what it looks like, what causes it, how to prevent it, how to repair it once it has fired, and what to read. The manual uses roles rather than names: your supervisor is the principal investigator (PI), the professor who runs the lab. A run gives the same four people names. Every term with a dotted underline has a definition on hover, and the glossary at the foot of this page carries all of them.
Terms the game uses, one line each.
Everything linked in the game. P2P is The Path to PhD, the newsletter this game grew out of.