Montaa // Decision intelligencePre-launch · design partners wanted
For the rare, multi-party calls you only get once — a funding round, an acquisition, a negotiation. Montaa models the other side's incentives and unknowns, plays out how it could go, and shows which move holds up best and what would change the answer.
01The problem
Every big decision involves other actors with their own goals, private information and moves still to come. Intuition can't hold all of that at once, and one wrong call can cost a deal, a company or a decade.
Montaa doesn't predict the future. It puts your assumptions on the table, plays them out against the other side, and shows which moves still hold up when you're wrong.
02How it works
01
Write the decision in plain language: who is involved, what each party wants, the options on the table, what's at stake, what you already know.
02
Montaa turns your brief into a formal game — players, information, preferences, moves, hidden unknowns — and lists every assumption it made so you can correct it.
03
Tens of thousands of rollouts play out every branch across your uncertainty ranges — including counterparts who adapt, bluff, and misjudge.
04
Get the move that holds up best across scenarios, shown as ranges with the trade-offs, the risks, and what would change the answer.
Scenario
Weighing an acquisition. Target board is split, an activist investor holds a stake, a rival bidder may emerge, and the regulator timeline is uncertain.
Model
players: 4 + 1 inferred · info: incomplete · stages: 5
rollouts: 50,000 · solver: Bayesian NE + CFR
input confidence: low–medium · 6 of 9 inputs are your estimates
Recommended next move
Take a minority stake now; keep the option to acquire.
| Option | Outcome range | Bad case |
|---|---|---|
| Stage in via minority stake | 0.50 – 0.75 | −0.25 – −0.05 |
| Acquire now | 0.40 – 0.75 | −0.60 – −0.25 |
| Wait six months | 0.25 – 0.60 | −0.35 – −0.10 |
| Pass | 0.10 – 0.30 | −0.10 – 0.00 |
Ranked by how well each option holds up across assumptions, not by a point estimate.
03Capabilities
01 / 04
Who really wins when anyone moves.
Big decisions are rarely two-sided. Montaa models every party's utility and maps how value flows between them as positions shift — so you see the decision as a system, not a standoff.
02 / 04
Every concession has a price. See it.
Instead of one 'best' answer, Montaa charts the full frontier of achievable outcomes and shows what you give up — in gain, speed and durability — for each step along it.
03 / 04
Know the bad day before it arrives.
Thousands of simulated futures turn gut-feel risk into a range of outcomes, not a single number. See the median, the tail, and which of your moves fatten or thin it.
04 / 04
The party not at the table still votes.
Boards, regulators, investors, rival bidders: influence that never appears on the invite. Montaa estimates how likely hidden actors are from observed behaviour and tests every scenario with and without them.
04Under the hood
Montaa combines classical solution concepts with modern simulation and self-play methods — the same families of techniques used in economics, auction design, and superhuman game-playing AI.
uᵢ(sᵢ*, s₋ᵢ*) ≥ uᵢ(sᵢ, s₋ᵢ*)
Find stable outcomes where no party gains by unilaterally changing course.
σᵢ(θᵢ) ∈ argmax E[uᵢ | θᵢ]
Reason about what others privately know, believe, and are hiding.
backward induction on the move tree
Discard empty threats; keep only commitments that stay credible at every stage.
max Πᵢ (uᵢ − dᵢ)
The principled split of surplus relative to each side's walk-away point.
uᵢ − dᵢ ∝ maxᵢ − dᵢ
A fairness benchmark proportional to what each side could ideally achieve.
φᵢ = Σ |S|!(n−|S|−1)!/n! · [v(S∪i) − v(S)]
Attribute value to each party by their average marginal contribution to every coalition.
Σᵢ∈S xᵢ ≥ v(S) ∀ S ⊆ N
Test whether any sub-group could profitably walk away and break the deal apart.
E[U] ≈ (1/N) Σₖ U(ωₖ)
Sample thousands of futures to estimate expected value, variance and tail risk.
UCT = X̄ᵢ + c√(ln N / nᵢ)
Efficiently explore deep move sequences, focusing compute where it matters.
σ ← regret-matching(Rᵀ)
Self-play to converge on robust strategies in games with hidden information.
P(a) ∝ exp(λ·u(a))
Model real counterparts as noisy and boundedly rational, not perfect calculators.
min_a max_ω [ u*(ω) − u(a, ω) ]
Choose the action you'll least regret across the scenarios you can't rule out.
Grounded in the literature
Nash (1950), The Bargaining Problem · Harsanyi (1967–68), Games with Incomplete Information Played by “Bayesian” Players · Shapley (1953), A Value for n-Person Games · Kalai & Smorodinsky (1975), Other Solutions to Nash's Bargaining Problem · McKelvey & Palfrey (1995), Quantal Response Equilibria for Normal Form Games · Kocsis & Szepesvári (2006), Bandit Based Monte-Carlo Planning · Zinkevich et al. (2007), Regret Minimization in Games with Incomplete Information
05Honest by design
Every model of a live deal rests on estimates — yours and ours — and Montaa won't dress them up as certainty. Outputs are ranges. Every assumption is visible and editable. And you're told which assumption would flip the recommendation, so you can challenge it before you act on it.
Recommendation
Take a minority stake now; keep the option to acquire.
Because
In most rollouts where a rival bidder emerges, buying outright at today's price overpays; a staged entry holds up better.
Assuming
Board stays split · regulator review ≥ 90 days · rival has not yet valued the asset.
Breaks if
The target accelerates a competing process. Re-run recommended.
Flips if
The activist's exit price is ~15% lower than assumed: 'Acquire now' then ranks first.
Confidence
Low–medium. 6 of 9 inputs are your estimates. Treat the ranking as a stress test, not a forecast.
06Who it's for
Lead and follow-on dynamics, terms versus dilution, and who signals whom.
Bidder behaviour, board and activist dynamics, and the sequencing of your next move.
Competitive processes, syndicate dynamics and when to push or walk.
Pressure-test a client's recommendation against how rivals and stakeholders will react.
Concession sequencing and walk-away points across several parties.
An acquisition, a market entry or a pricing move where rivals and boards respond.
Montaa is in development. We're inviting a small group of founders, deal teams and advisors facing a live decision to shape it and get first access.
This form asks for no deal details, so please don't include confidential information. See our privacy note. Montaa is decision support, not financial or legal advice.
Pre-launch · all figures on this page are illustrative · no customer results yet