What we do
Simulation
Model the likely outcome of a decision before the institution commits to it, across resource allocation, portfolio choices and standing against peers. You see the consequence of a move on paper, before it costs anything in practice.
The evidence
Priced before it is promised.
A point rank cannot be defended once the edition lands. A corridor can: it carries the width the point hides, and a floor the institution holds even when the field runs against it.
The problem
Institutions commit serious money to ranking outcomes: a hiring round, an international push, a research investment, each justified by where it should move the institution’s standing. Almost none of it is priced first. Each agency runs its own machinery and no two run the same one: ratios capped in one scheme and uncapped in another, sharp changes damped here and credited in full there, weight regimes revised on their own schedules and repricing every position overnight. That machinery is public, but it is rarely modelled. So boards are promised point ranks nobody can stand behind, budgets land on indicators already at their ceiling, and when the edition is published no one can say how much of the movement was the institution’s own doing. The cost is not only the wasted spend. It is the credibility of the next ask.
The machinery
What actually decides the value of a move.
Each is published by the agency that runs it, and none is shared — the same decision prices differently in each. So we model every system on its own terms and never carry a signal from one into another. Modelled today: QS and THE.
Cap
QS
A ratio is allowed to count only so far. Past the ceiling, another appointment adds nothing to the score however well the money is spent.
QS caps international ratios at 0.5 and faculty-student at 0.3. THE publishes no ratio cap — its only cap is on reputation self-votes.
Damping
QS
A sharp change is held back and released over several editions, so the year a move is paid for is rarely the year it shows.
QS damps six of its nine indicators. THE publishes no damping rule, so the same move prices differently in each.
Regime
QS · THE
The weights themselves are revised. When they are, every institution reprices at once and a position held under the old weights may not survive the new.
Both revised in 2024, independently and to different ends.
Inside the simulator
How a move gets priced.
- 01
Rebuild
Every indicator through its own caps and damping, the overall composed exactly as the agency composes it. Nothing is fitted to the headline rank.
- 02
Move the field
Rank is read from thousands of joint futures in which every institution moves at once. A position is lost as easily to a rival’s good year as to your own bad one.
- 03
Price the lever
A lever is priced only when its path to the score survives validation. What fails is declined, with the reason stated.
- 04
Return a corridor
A range with a floor, never a point, and provisional until the next edition scores it against reality.
Twelve months on, someone can say how much of the movement was the institution’s own doing. That is the whole point.
Also in What We Do
Talk to us about simulation.
Start a conversation