polynomial-time algorithms for special games like weighted voting
You hold 30% of the votes on a PE-Backed company's board. Three other partners hold 28%, 22%, and 20%. Decisions require a 51% majority. You assume your influence is roughly proportional to your 30% - but when you map out which groups of voters can pass a resolution without you versus with you, your real power to tip outcomes is wildly different from 30%. The gap between voting weight and actual influence is where Operators get blindsided in governance, cost sharing agreements, and consortium decisions.
A weighted voting game assigns unequal votes to players and sets a passing threshold. Computing the Shapley value for these games is fast for any real governance structure, letting you quantify each player's real influence - which often diverges sharply from their raw vote share.
A weighted voting game is a special structure inside Game Theory where:
You write it as [threshold; w1, w2, ..., wn] where the first number is the votes needed to pass and each wi is player i's weight.
Example: [51; 30, 28, 22, 20] means 51 votes needed to pass, and the four players have 30, 28, 22, and 20 votes respectively.
The critical insight: weight is not power. A player with 1% of the weight might have 0% of the power (they can never be the deciding vote in any group), or a player with 49% might have exactly the same power as one with 2% if they each serve as the deciding marginal contribution in the same number of winning combinations.
Because weighted voting games have a clean threshold structure, computing the Shapley value for each player is fast for governance-sized problems (5-20 players). A spreadsheet or short script handles it in seconds. This means you can run the numbers on real board structures, committee rules, and consortium agreements before you commit to them.
If you run a P&L inside a larger structure - a PE-Backed company, a consortium, a Holding Company with multiple business units - your formal authority is governed by voting rules. Those rules determine:
The P&L impact is indirect but high-stakes: the player who controls the vote controls the Budget, the Capital Structure changes, and the strategic direction. Misjudging your actual influence means either overinvesting in fights you can't win, or failing to exercise Leverage you actually have.
List all players with their weights and the passing threshold.
[51; 35, 25, 25, 15] - four partners, 51 votes needed.
A group wins if the sum of its members' weights meets or exceeds the threshold.
| Group | Total Weight | Wins? |
|---|---|---|
| {35, 25a} | 60 | Yes |
| {35, 25b} | 60 | Yes |
| {35, 15} | 50 | No |
| {25a, 25b} | 50 | No |
| {25a, 15} | 40 | No |
| {25b, 15} | 40 | No |
| {35, 25a, 25b} | 85 | Yes |
| {35, 25a, 15} | 75 | Yes |
| {35, 25b, 15} | 75 | Yes |
| {25a, 25b, 15} | 65 | Yes |
| {all four} | 100 | Yes |
For the Shapley value, you list every possible ordering of players joining one at a time and ask: when this player arrives, does the group's running total first cross the threshold? The player who tips the total past the threshold in a given ordering gets credit for that ordering.
The 35-weight player tips the outcome in more orderings than their 35% weight would suggest. The 15-weight player tips it in fewer.
For this game across all 4! = 24 orderings, the Shapley values are:
The 35-weight player has 5x the influence of the 15-weight player, despite having only 2.3x the votes.
For governance-sized problems - 5 to 20 players - listing all orderings or building a weight-sum table runs in seconds on a spreadsheet or a short script. This is what makes weighted voting practically useful: you can compute Shapley values as a live decision rule tool before you ratify any governance structure.
Use weighted voting analysis when:
Don't bother when:
A PE firm acquires a company. Post-deal board has 5 seats: PE fund holds 3 seats (60% weight), the founder/Operator holds 1 seat (20%), and an independent holds 1 seat (20%). Passing threshold is simple majority: 51%. The Operator wants to understand their real influence over Capital Allocation decisions.
Define the game: [51; 60, 20, 20] with players PE, Operator, Independent.
Enumerate winning groups: {PE} = 60 >= 51 (wins alone). {PE, Operator} = 80 (wins). {PE, Independent} = 80 (wins). {Operator, Independent} = 40 (loses). {All} = 100 (wins).
Find which players are the deciding marginal contribution in each winning group. In {PE}: only PE is deciding. In {PE, Operator}: PE is deciding (removing PE drops to 20), but Operator is NOT (removing Operator leaves 60, still wins). In {PE, Independent}: same logic - Independent is NOT deciding. In {All}: only PE is deciding (removing PE gives 40 < 51; removing either other player still leaves >= 80).
Compute Shapley values by checking all 3! = 6 orderings. PE tips the total past 51 in ALL six orderings. Operator and Independent tip it in zero orderings - there is no sequence where either non-PE player's arrival first pushes the running total to 51. Result: PE Shapley = 1.000, Operator Shapley = 0.000, Independent Shapley = 0.000.
The Operator has zero real power despite holding 20% of the board seats. PE can pass any resolution alone.
Insight: One player's weight alone meets the threshold - the Operator's 20% board representation is purely ceremonial. If you're negotiating a deal like this, the only moves that matter are: (a) negotiate the threshold up to 67% or 75%, which forces PE to need at least one ally, or (b) negotiate for blocking rights on specific decision types like Capital Structure changes. Weight without the ability to be the deciding vote is worth exactly nothing.
Three business units share a data platform. They vote on the annual infrastructure Budget. BU-Alpha (the largest, $50M Revenue) gets 5 votes, BU-Beta ($30M Revenue) gets 3 votes, BU-Gamma ($10M Revenue) gets 2 votes. Passing threshold: 6 votes needed (simple majority of 10). Each BU's cost share is currently proportional to votes: 50%, 30%, 20%. Is this fair?
Define the game: [6; 5, 3, 2].
Winning groups: {Alpha, Beta} = 8 (wins). {Alpha, Gamma} = 7 (wins). {Beta, Gamma} = 5 (loses). {Alpha, Beta, Gamma} = 10 (wins). No player can win alone (Alpha alone = 5 < 6).
Deciding marginal contributions: In {Alpha, Beta}: Alpha is deciding (removing drops to 3), Beta is deciding (removing drops to 5 < 6). In {Alpha, Gamma}: Alpha is deciding (drops to 2), Gamma is deciding (drops to 5 < 6). In {All three}: only Alpha is deciding (removing Alpha gives 5 < 6; removing Beta gives 7 >= 6; removing Gamma gives 8 >= 6).
Compute Shapley values across all 3! = 6 orderings. Alpha tips the total past 6 in 4 of 6 orderings. Beta in 1 of 6. Gamma in 1 of 6. Shapley values: Alpha = 4/6 = 0.667, Beta = 1/6 = 0.167, Gamma = 1/6 = 0.167.
Alpha has 50% of the votes but 66.7% of the power. Beta and Gamma each have equal power (16.7%) despite Beta having 50% more votes (3 vs 2). If cost sharing tracks real influence, Alpha should pay 66.7% of the Budget (not 50%), and Beta and Gamma should each pay 16.7%.
Insight: Beta and Gamma have identical power despite different weights - neither can win without Alpha, and each is equally useful as Alpha's partner. The proportional-to-votes cost split overcharges Gamma and undercharges Alpha relative to actual governance influence. When negotiating consortium agreements, compute Shapley power before accepting weight-proportional cost sharing.
You're the CFO designing a Capital Investment approval committee. Five stakeholders: CEO (wants speed), CTO (wants technical quality), COO/Operator (wants P&L impact), VP Sales (wants customer features), VP Finance (wants ROI discipline). You need to set weights and a passing threshold. Current proposal: CEO gets 4 votes, VP Finance gets 1, others get 2 each, and 7 of 11 total votes are needed to pass.
Define the game: [7; 4, 2, 2, 2, 1]. CEO can't pass alone (4 < 7). Check whether the four non-CEO members can pass together: 2+2+2+1 = 7, which meets the threshold. So the non-CEO group CAN approve investments without the CEO.
Compute Shapley values across all 5! = 120 orderings. For each ordering, track the running total and mark which player's arrival first pushes it to 7 or above.
Results: CEO = 0.400, CTO = 0.150, COO = 0.150, VP Sales = 0.150, VP Finance = 0.150. The CEO holds 36% of the votes but 40% of the influence - a mild concentration. But the real surprise: all four non-CEO players have identical Shapley values, despite VP Finance having half the votes of the others.
Why VP Finance's single vote buys the same influence as each 2-vote player: VP Finance tips the outcome in exactly the same fraction of orderings. When the CEO and one weight-2 player have already joined (total = 6), VP Finance's 1 vote pushes to 7 - just like a weight-2 player would. And in the four-player non-CEO group {CTO, COO, VP Sales, VP Finance} totaling exactly 7, every member is the deciding marginal contribution since removing any one drops the total below 7. The positions where each player tips the outcome turn out to be perfectly symmetric.
Now compare: raise the threshold to 8. Same weights [8; 4, 2, 2, 2, 1]. The non-CEO group totals 7 < 8, so they can NEVER pass without the CEO. New Shapley values: CEO = 0.500, each weight-2 player = 0.167, VP Finance = 0.000. VP Finance is completely powerless - a phantom seat at the table.
Why VP Finance drops to zero: for VP Finance's 1 vote to matter, some group of other players would need to total exactly 7 (so that adding 1 reaches 8). But no combination of {4, 2, 2, 2} sums to 7. Check: 4+2+2+2 = 10, 4+2+2 = 8, 4+2 = 6, 2+2+2 = 6, 2+2 = 4. None equal 7. VP Finance can never be the deciding vote under this threshold.
One threshold point - from 7 to 8 - wipes out VP Finance entirely and shifts the CEO from 40% to 50% influence.
Insight: Small changes to the passing threshold create large, nonlinear shifts in the distribution of real influence. Under threshold 7, VP Finance's single vote is worth exactly as much as each weight-2 player's two votes - a result no one would guess from the raw numbers. Under threshold 8, that same vote is worth nothing. Always compute Shapley values for proposed governance rules before ratifying them. The influence structure you're actually creating may differ dramatically from what the weights suggest.
Weight is not power. A player's share of the votes can be completely unrelated to their ability to tip outcomes. A player with half the votes of another can have identical Shapley influence - or a player with positive weight can have exactly zero power. Always compute the Shapley value to find real influence in any weighted voting structure.
Weighted voting is computationally practical. For any governance-sized problem (5-20 players), you can compute Shapley values with a spreadsheet or a short script in seconds. This means you can (and should) run the numbers on real board structures, committee rules, and consortium agreements before you ratify them.
Threshold design is the real lever. The passing threshold (how many votes it takes to win) often matters more than the weight distribution. Shifting a threshold by a single vote can turn a balanced structure into one where a player has unilateral control, or make another player's weight completely ceremonial.
Assuming proportional weight means proportional influence. This is the central trap. In the consortium example, Beta has 50% more votes than Gamma but identical power. In the committee example, VP Finance has half the votes of the other non-CEO players but identical power under one threshold - and zero power under another. Never negotiate governance terms based on weight alone - always compute the Shapley value.
Ignoring the threshold when evaluating a deal. Operators in M&A due diligence focus on how many board seats they get but skip the threshold analysis. Two seats on a seven-member board with simple majority is very different from two seats with a 75% threshold requirement. The threshold determines whether your seats have real power or are decorative.
A startup has three co-founders. They set up voting with weights: CEO = 5, CTO = 3, COO = 2. Passing threshold is 6 (simple majority of 10). Compute each founder's Shapley value. Does any founder have disproportionate power relative to their weight?
Hint: List all 3! = 6 orderings. For each ordering, track the running total and mark which player's arrival first pushes the total to 6 or above. That player gets credit for that ordering. Each player's Shapley value = (number of orderings where they tip the total) / 6.
Orderings and the player who tips the total past 6:
Shapley values: CEO = 4/6 = 0.667, CTO = 1/6 = 0.167, COO = 1/6 = 0.167.
CEO has 50% of the weight but 66.7% of the power. CTO and COO have equal power (16.7% each) despite CTO having 50% more weight (3 vs 2). The CTO's extra vote over the COO buys zero additional influence because neither can form a winning group without the CEO, and each is equally sufficient as the CEO's partner.
You're negotiating post-acquisition board terms. The PE fund proposes: PE gets 4 seats, you (Operator) get 2 seats, one independent gets 1 seat. Simple majority rules (4 of 7 seats needed to pass). What is your Shapley value? What single change to the structure would most increase your real power?
Hint: Model as [4; 4, 2, 1]. Compute all winning groups, then Shapley values across 3! orderings. For the second part, consider: what happens if you raise the passing threshold from 4 to 5?
Game: [4; 4, 2, 1]. PE can win alone (4 >= 4). One player can pass anything unilaterally.
Shapley values: PE = 1.000, Operator = 0.000, Independent = 0.000. You have zero power.
Most impactful change: raise the threshold to 5. New game: [5; 4, 2, 1]. Now PE can't win alone (4 < 5). Winning groups: {PE, Operator} = 6, {PE, Independent} = 5, {PE, Operator, Independent} = 7. Note {Operator, Independent} = 3 < 5, so the non-PE group still can't win.
New Shapley values: PE = 4/6 = 0.667, Operator = 1/6 = 0.167, Independent = 1/6 = 0.167.
Raising the threshold by just 1 vote moves you from 0% power to 16.7% power and strips PE of unilateral control. This is why higher threshold requirements matter more than seat counts in M&A due diligence. An alternative approach: negotiate blocking rights on Capital Structure changes, which effectively creates a higher threshold for those specific decisions.
Weighted voting sits at the intersection of Game Theory and Shapley value - it's the most practically relevant game structure for governance because real organizations everywhere use threshold voting. Your prerequisite knowledge of the Shapley value (averaging marginal contributions across all orderings) applies directly here, and the clean threshold structure of weighted voting makes the computation fast for any real governance problem - which means you can use it as a live decision rule tool, not just a theoretical concept. Downstream, weighted voting connects to Efficient Allocation (Shapley-based cost sharing in consortium structures), public good provision (how groups collectively fund shared infrastructure), and Capital Allocation decisions (who really controls where the money goes). For PE operators specifically, weighted voting analysis is the quantitative backbone of board governance - it turns vague notions of 'influence' and 'control' into computable numbers that directly inform deal negotiation, Vendor Negotiations, and post-merger integration.
Disclaimer: This content is for educational and informational purposes only and does not constitute financial, investment, tax, or legal advice. It is not a recommendation to buy, sell, or hold any security or financial product. You should consult a qualified financial advisor, tax professional, or attorney before making financial decisions. Past performance is not indicative of future results. The author is not a registered investment advisor, broker-dealer, or financial planner.