Skip to content

Post-Contest Impact Review

Retrospective review — instantiates Bounded Rivalry Governance

Looks back after the contest to check whether the winner actually delivered the intended value and what harms leaked out, then feeds the redesign of the next round.

Post-Contest Impact Review is the one backward-looking mechanism in the set. After the winner is chosen and the dust settles, it asks the questions the contest design could not answer in advance: did winning actually track the purpose, or did the metric reward the wrong thing? What harms landed on people outside the arena? Did victory hand the winner durable power beyond the prize itself? It then turns those answers into changes to the rules, metric, or eligibility for next time. Its output is not a score but a verdict-plus-revision — and the unit under review is the contest, not the entrants.

Example

A government ran a reverse auction to award rural-broadband subsidies to the lowest bidders meeting a coverage spec. Two years on, a Post-Contest Impact Review checks what actually got built: did winners deliver the promised coverage, or underbuild and pocket the difference? Did the low-bid designs leave towns half-served or push costs onto subscribers (stakeholder harm)? Did the largest winner parlay its award into a regional lock on the infrastructure (winner power)? The review finds a pattern of over-optimistic bids, and recommends tighter build-out verification and clawback clauses for the next round. This is ex-post evaluation applied to a contest.

How it works

  • Compare realized outcomes to the stated purpose — did winning mean what it was supposed to mean?
  • Trace harms and spillovers to non-participants and third parties.
  • Review winner power — did the victory create monopoly, standard capture, or lock-in unrelated to the contest's purpose?
  • Feed concrete revisions — metric, eligibility, arena boundary — into the next contest's design.

The distinguishing feature is that it judges the governance design itself, after the fact, using outcomes the up-front rules could only guess at.

Tuning parameters

  • Lag — how long after the contest to review; too soon and impacts haven't landed, too late and the next round is already locked.
  • Scope of impact — winners only, the whole field, or non-participant stakeholders too — the wider scope guards against survivorship bias.
  • Counterfactual rigor — a casual before/after versus a real comparison to what would have happened otherwise.
  • Bindingness — advisory findings versus mandated design changes and clawbacks.
  • Cadence — a one-off post-mortem versus a standing loop that runs every cycle.

When it helps, and when it misleads

Its strength is that it closes the loop competition otherwise leaves open: it is the only step that can catch a metric that rewarded the wrong thing, a winner that delivered less than promised, or a victory that quietly became a lock-in — and it prices in harms the in-arena design missed. It is anchored in the discipline of ex-post evaluation.[1]

Its limits are inherent to being retrospective: it can only fix the next contest, never the one that already misfired. Reviewing only the winners invites survivorship bias, and it is the easiest step to skip once the prize is awarded and attention has moved on. Run backwards, the review is written to vindicate the original design rather than test it. The discipline is to pre-commit to the review before the contest ends, include losers and non-participants in scope, and give the findings teeth — real revisions, not a filed report.

How it implements the components

Post-Contest Impact Review realizes the retrospective, learning side of the archetype — the components that judge and recalibrate the contest after it runs:

  • learning_and_recalibration_loop — its core: it converts realized outcomes into revisions of the contest's rules, metric, and eligibility.
  • stakeholder_harm_monitor — it traces and records the harms and spillovers that landed on people outside the arena.
  • winner_power_and_lock_in_review — it checks, after the fact, whether the victory created monopoly, standard capture, or durable lock-in unrelated to the contest's purpose.

It measures and learns after the fact but does not design the up-front prize split (that's Multiple-Award or Portfolio Selection) or price externalities into the contest in advance (that's Externality Bond or Liability Rule); it reviews what the contest actually did.

References

[1] Ex-post evaluation (or post-implementation review) assesses a decision after its effects have materialized, comparing realized outcomes against intended ones. Its entire value is in feeding the next decision — and reviewing only the survivors is the classic way it misleads. ↩