Skip to content

Multiple-Award or Portfolio Selection

Selection method — instantiates Bounded Rivalry Governance

Splits the prize across several winners instead of crowning one, so rivalry still sharpens performance without collapsing into winner-take-all lock-in.

Most contest designs assume a single winner. Multiple-Award or Portfolio Selection refuses that premise. Instead of asking "who is best?" it asks "what set of winners best serves the purpose?" — dividing the scarce prize (funding, contracts, slots) into a designed portfolio of awards and deliberately keeping more than one rival alive past the finish line. The point is not fairness for its own sake; it is that a lone winner concentrates power, kills off the diversity of live alternatives, and inherits durable leverage over the next round. Portfolio selection trades a little peak performance for resilience, a standing bench of options, and a field that stays contestable.

Example

A national science agency has a fixed pot — say ≈$60M — to advance a hard problem such as grid-scale energy storage. The obvious design is one flagship award to the top-ranked proposal. Multiple-Award/Portfolio Selection instead funds five teams at roughly $12M each, chosen not by rank alone but to span distinct technical bets: two established favorites, two long shots on different chemistries, one wildcard. When two of the approaches stall in year two, the program still has three live shots on goal — and no single laboratory ends up owning the only viable design, and the standard that would grow around it.

How it works

  • Set a portfolio objective, not just a ranking — define the coverage or diversity the set should achieve, not only the top score.
  • Choose the split — how many awards, how large each, equal or tiered.
  • Select for complementarity — the marginal award is picked for what it adds to the set, so a lower-ranked-but-different entrant can beat a higher-ranked-but-redundant one.
  • Check concentration — confirm no single winner still walks away with outsized downstream advantage.

The distinguishing move is that selection scores the portfolio, not each entrant in isolation.

Tuning parameters

  • Number of awards / prize granularity — more winners buy resilience and diversity but thin the funding each receives and blunt the selective signal.
  • Diversity vs. pure-merit weighting — how much you select for spanning different approaches versus taking the top N by score; over-weighting diversity funds weak-but-different bets.
  • Award-size distribution — equal splits versus a large lead award plus smaller hedges.
  • Down-select cadence — a one-shot portfolio versus funding many now and cutting to a few at a later evidence gate.
  • Concentration ceiling — an explicit cap on how large a share of the prize any one winner may hold.

When it helps, and when it misleads

Its strength is that it keeps the field alive and contestable, hedges the irreducible uncertainty about which approach will win, and starves winner lock-in at the source — you cannot capture a standard you only partly won. It is the direct antidote to concentration risk.[1]

Its failure mode is dilution: split too far and every bet drops below critical mass ("peanut-buttering"), so nothing gets enough to succeed. It can also decay into a spread-the-money spoils system where "diversity" is cover for declining to choose. Run backwards, it rationalizes funding a favorite plus a few tokens. The discipline that guards against this is to tie every award to the portfolio's purpose and to keep a real down-select gate, so multiple awards are a hedge, not a permanent entitlement.

How it implements the components

Multiple-Award realizes the prize-shape side of the archetype — the components that decide how the scarce prize is divided:

  • prize_decomposition_or_multiple_winner_design — its core act: dividing one prize into a designed set of awards rather than a single victory.
  • scarce_prize_or_selection_constraint — it operates directly on the scarce prize, redefining "the prize" from a single slot into a portfolio.

It does not vet who may enter (that's Tender or RFP Process) or score the entrants (that's Ranked Leaderboard With Audit); and while splitting the prize is the front-line defense against winner lock-in, the after-the-fact review of whether a win created durable power belongs to Post-Contest Impact Review and Antitrust or Competition Review.

Notes

Multiple-Award is a selection-shape decision: it presumes a ranking already exists and simply refuses to collapse it to one winner. Reach for it when uncertainty about the best approach is high or when concentration is dangerous — not as a default that quietly dodges the work of choosing.

References

[1] Concentration risk — the exposure created when a single winner holds an outsized share of a resource, standard, or capability, so its failure or capture endangers the whole system. Spreading awards is the standard hedge, the same logic as diversifying a portfolio. ↩