Every order of arrival for three partners, and what each player adds
coalition is one function. Everything below came out of it during this
build, at parameters taken from the essays rather than invented for this page — so a figure
here is the same figure a reader meets in an essay, and if the generator changes, this page
changes with it.
With nothing chosen
An exchange box, and a core that shrinks as the market grows
A coalition of three that the two-trader market did not have
Treating two traders of one kind differently invites a block
The core narrows to the price as the traders multiply
The core shrinks like one over the number of traders
What it checks while it draws
Collected by running the family and recording what it asserted, not written here. The count is how many separate times the claim was put to the test while these drawings were made.
- at a quota of 9 Luxembourg is powerless exactly when the quota is even ×9
- the layer formula and the average over all 3! orders agree for A ×2
- the layer formula and the average over all 3! orders agree for B ×2
- the layer formula and the average over all 3! orders agree for C ×2
- the quota is a whole number between 1 and 120 ×2
- the weights is a list of 3 to 3 numbers ×2
- A and B are interchangeable ×1
- a division outside the core is refused by some group ×1
- a family demanding more than there is rules the core out ×1
- a game that is not convex has an arrival-order vector the core refuses ×1
- a member who can never swing a vote has no power by either count ×1
- a merger helps under majority and hurts under unanimity and under a quota short of it ×1
- a network view's game is one of its own ×1
- a rational is a whole numerator over a non-zero whole denominator ×1
- a rational is never divided by zero ×1
- a satisfied set of certificates finds splits, unless the core is a single point off the lattice ×1
- a split some coalition can beat is objected to ×1
- a stable split is one every objection to which can be answered ×1
- a whole-number seed ×1
- an even number of players from 6 to 60 ×1
- an input the model never reads gets nothing when the unknowns are filled independently ×1
- and at least one of the other rules fails one ×1
- and D adds nothing ×1
- and every corner of the core is one of them ×1
- and in 1981 more than in 1973 ×1
- and it falls at about the square-root rate over the whole sweep ×1
- and neither end gains from losing it ×1
- and satisfies the other three ×1
- and the additivity column's rule gives a different answer here ×1
- and the efficiency column's rule gives a different answer here ×1
- and the symmetry column's rule gives a different answer here ×1
- and the the null player column's rule gives a different answer here ×1
- and their heights add to it ×1
- at least one assembly drawn has power and weight ordered differently ×1
- at least one family is satisfied, so the table is separating them ×1
- averaging what each adds satisfies all four on this game ×1
- averaging what each player adds passes every condition on every game tested ×1
- below one half the line beats every split off it ×1
- between one and four assemblies ×1
- between three and eight increasing sample sizes ×1
- Bird's division is one no group refuses ×1
- Bird's payments add to the cost of the whole tree ×1
- both cheapest trees cost the same ×1
- C is not interchangeable with them, so symmetry is a real condition here ×1
- cut is read only by the linkcut view ×1
- cutting a link changes both of its ends by the same amount ×1
- donations that raised the donor's share of orders ×1
- each entry of the capacities the users need is a whole number between 1 and 200 ×1
- each entry of the split argued about, as three weights is a whole number between 0 and 100 ×1
- each entry of the weights is a whole number between 1 and 40 ×1
- each share lands on the grid the search used ×1
- every arrival-order vector shares out exactly the whole ×1
- every coalition's value is a whole number ×1
- every estimate is a share ×1
- every family in the table really is balanced ×1
- every held share below one half gives a stable set ×1
- every marked split really attains the smallest imbalance ×1
- every merger lost, gained or broke even ×1
- every order adds exactly the number of pairs, one arrival at a time ×1
- every order has exactly one pivotal member ×1
- every ordering ends with everybody in ×1
- every other split on the lattice is beaten by one of the three ×1
- every price is in the core when the sides are equal ×1
- every split hands out the whole ×1
- every split outside the core is beaten from inside it ×1
- fill is read only by the attribtable view ×1
- filled from matching houses, the unread input takes credit ×1
- fills is read only by the attrib view ×1
- graph is read only by the network and linkcut views ×1
- graphs is read only by the networks view ×1
- in 1958 Luxembourg is never pivotal ×1
- in 1973, with more members to share with, Luxembourg has power ×1
- in a convex game every arrival-order vector is a corner of the core ×1
- in the drawn case Banzhaf's share rises and the share of orders does not ×1
- mergers go both ways ×1
- model is read only by the attribution views ×1
- net is read only by the network-bill views ×1
- no coalition objects to a split none of them can beat ×1
- no coalition of the other members totals exactly one short of the quota ×1
- no group pays more under Bird's rule than its own cheapest tree would cost ×1
- no half-and-half split beats another ×1
- no other split on the table has a quieter list of complaints ×1
- no point of the line beats another ×1
- no split in the core beats another ×1
- no split survives when some balanced family over-demands ×1
- nothing both beats the split and is beaten by it ×1
- on this game the nucleolus and the average split are different points ×1
- one A and two B's block A's share 2.15, and one of each does not ×1
- one more vote on the quota and there is such a coalition ×1
- one to three networks ×1
- relaxing by less does not ×1
- relaxing by the computed amount lets a split through ×1
- short in proportion, the scarce side's share rises towards the whole pair ×1
- so their average lies in the core ×1
- some coalition holds one player and not the other ×1
- some coalition objects to every split drawn here ×1
- some donation raises the donor's normalised Banzhaf share ×1
- some family attains the maximum ×1
- some house in the data matches the known inputs ×1
- some split is individually rational ×1
- some split on the lattice pays every player at least what they earn alone ×1
- somebody can swing something, or the assembly decides nothing ×1
- the allocation uses exactly the market's goods ×1
- the average split is the average of the arrival-order vectors ×1
- the B's gain and the A is held exactly level ×1
- the balance verdict is computed for every split ×1
- the bill is drawn on a triangle for three users ×1
- the capacities the users need is a list of 2 to 5 numbers ×1
- the census finds enlargements in which an old member gains ×1
- the certificate is drawn for three players ×1
- the closed-form shares pay the whole bill ×1
- the coalitions together demand more than there is to give ×1
- the core is drawn for three players ×1
- the core of the convex game is not empty ×1
- the core, nucleolus and excess views are written for a game whose value is shared out, and this game's value is a bill ×1
- the core's corners are found for three players ×1
- the credits add up to the prediction less the starting value ×1
- the cut is one of the network's links ×1
- the denominator the splits are searched over is a whole number between 12 and 180 ×1
- the empty coalition is worth nothing ×1
- the error at the largest sample is a third of the error at the smallest ×1
- the error is a distance ×1
- the exact arithmetic stays inside the safe integer range ×1
- the exact shares add to one ×1
- the excess over a half lies between 0 and 0.45/√n ×1
- the family chosen is a whole number between 1 and 5 ×1
- the fill is one of baseline, marginal, conditional ×1
- the fills are baseline, marginal or conditional ×1
- the game is big enough that listing the orders is the point ×1
- the game is one of majority, pairs ×1
- the game is one of partnership, majority, gloves, shops, capacity, dummy, square ×1
- the game names a value for every coalition ×1
- the go-between takes the largest share ×1
- the grid divides the whole group's value evenly ×1
- the group A pays no more together than it would alone ×1
- the group AB pays no more together than it would alone ×1
- the group ABC pays no more together than it would alone ×1
- the group ABCD pays no more together than it would alone ×1
- the group ABD pays no more together than it would alone ×1
- the group AC pays no more together than it would alone ×1
- the group ACD pays no more together than it would alone ×1
- the group AD pays no more together than it would alone ×1
- the group B pays no more together than it would alone ×1
- the group BC pays no more together than it would alone ×1
- the group BCD pays no more together than it would alone ×1
- the group BD pays no more together than it would alone ×1
- the group C pays no more together than it would alone ×1
- the group CD pays no more together than it would alone ×1
- the group D pays no more together than it would alone ×1
- the held share lies on the lattice ×1
- the imbalance separates the splits, so the condition bites ×1
- the kernel is drawn for three players ×1
- the lattice holds every value exactly ×1
- the layer formula and the average over all 4! orders agree for D ×1
- the layers stack up to the largest requirement ×1
- the least core is drawn for three players ×1
- the marginal contributions along one ordering add up to the whole ×1
- the marginal view is drawn for three players ×1
- the marginal view is written for a game whose value is shared out ×1
- the model is one of proxy, and ×1
- the network is one of line3, triangle, vee, star4, bridge ×1
- the network is one of remote, far, tie ×1
- the network's whole earning is shared out ×1
- the nucleolus survives every objection exactly when some split does ×1
- the number of users is a whole number between 4 and 11 ×1
- the objection is drawn for three players ×1
- the other tree's Bird division is in the core too ×1
- the others fail ×1
- the pairwise surpluses are drawn for three players ×1
- the picture is symmetric between B and C ×1
- the power shares add to one ×1
- the proportional rule needs somebody to be worth something alone ×1
- the proposed split is beaten ×1
- the proposed split shares out the three pairs ×1
- the quota is more than half the weight and no more than all of it ×1
- the rarer input of the product takes the larger share ×1
- the requirements are given smallest first ×1
- the resolution of the exhaustive sweep is a whole number between 12 and 120 ×1
- the resolution of the relaxed sweep is a whole number between 12 and 120 ×1
- the resolution of the sweep the verdict is checked against is a whole number between 12 and 120 ×1
- the resolution the proposals are searched over is a whole number between 8 and 48 ×1
- the rule in the additivity column fails additive ×1
- the rule in the efficiency column fails efficient ×1
- the rule in the symmetry column fails symmetric ×1
- the rule in the the null player column fails null player ×1
- the scarce side takes everything ×1
- the scarce side's share falls as the market grows ×1
- the search finds no balanced family of at most three coalitions outside the table ×1
- the search is written for three players ×1
- the second game names a whole value for each of the eight coalitions of three, starting at nothing ×1
- the seed is a whole number between 0 and 10000 ×1
- the shares add to what the whole group is worth ×1
- the shares add up to what the whole group is worth ×1
- the split argued about, as three weights is a list of 3 to 3 numbers ×1
- the split hands out the whole ×1
- the split is three shares adding to one ×1
- the split lies on the lattice ×1
- the splits achieving the minimum are one point to within a grid step ×1
- the splits give different surpluses, so the table is comparing them ×1
- the subset formula and the average over orders agree ×1
- the sweep and the balancedness condition agree about whether any split survives ×1
- the symmetric set and the low line are stable ×1
- the table is drawn for three players ×1
- the test is drawn for three players ×1
- the three lefts and three of the rights earn more alone than they are paid ×1
- the users do not all need the same thing, or there is nothing to divide unevenly ×1
- the view is one of axioms, core, nucleolus, excess, unique, power, sample, layers, balanced, certificate, leastcore, surplus, kernel, counter, attrib, attribtable, attribdata, network, networks, linkcut, marginals, convexity, mst, mstcore, msttable, mstbig, stabledom, stablesym, stablediscrim, stablecheck, stableconvex, eec, parity, quotasweep, newmembers, merge, donate, glovestep, glovepath, glovereplica, gloveblock, gloveequal, repbox, repshrink, reprate, repwho, repcoalition, repequal ×1
- the weights are not all nought ×1
- the weights are positive ×1
- the whole group is worth a whole number ×1
- the worst-treated A and B both prefer splitting their joint endowment evenly ×1
- there is more than one layer, or the figure has nothing to show ×1
- two games on the same number of players ×1
- two hundred lefts against three hundred rights take nearly everything ×1
- two to four games this family knows ×1
- users is read only by the large network view ×1
- what all four players are worth together is a whole number between 4 and 40 ×1
- when a stable split exists, every balanced split is one ×1
- width times r settles from twenty replicas ×1
- with equal sides the average over orders splits each pair evenly ×1
- with every input known the value is the prediction itself ×1
- with fewer links than some of the others ×1
Where it is called
Every figure on this list is drawn by the same rule, so a change to the rule changes all of them at once. That is why the list is published.
A market too large to bargain in
Two traders swapping goods can settle anywhere along a stretch of possible deals; nothing forces a price. Bring in a second pair of the same traders and some of those deals can be refused by a group of three. With thirty of each, the only deals no group can refuse lie within a few thousandths of the one a price would produce. Edgeworth guessed in 1881 that competition shrinks bargaining to prices; Debreu and Scarf proved it, and the shrinking can be computed.
AppliedA share of the votes is not a share of the power
Give three members four, four and one vote, with five needed to pass. Every winning coalition needs exactly two of them, so all three have equal power — and one of them holds a ninth of the votes.
AppliedA split nobody can walk away from
Every way of dividing what a group earns is a point of a triangle, and every coalition's threat to leave cuts a straight line across it. What survives all the cuts is the set of stable divisions — and for one three-player game there is nothing left.
AppliedA vote worth nothing until it was outnumbered
From 1958 to 1973 Luxembourg held one vote of seventeen in the Council of the European Communities and could never once change an outcome. When the Council grew, Luxembourg's share of the votes fell and its share of the power rose from nothing. Power is not a quantity a member holds; it is a property of the whole assembly, and changing the assembly moves it in directions nobody would guess.
AppliedAn objection one player makes to another
The core lets a coalition object to everybody at once. Narrow it to one player objecting to one other, require every such objection to be met by an equally loud one coming back, and exactly one split survives — with no dictionary order anywhere in the argument.
AppliedCutting a link costs both of its ends the same
Three players, any two of whom can earn 1 together — but only if they are linked. Link all three and each is due a third. Remove one link and the player holding both of the others is due two thirds. Averaging over orders on the game the network allows is the one rule under which breaking any link costs the two players it joined exactly the same, and it pays go-betweens more than their links.
AppliedEach user pays for its own last link
Several users must be connected to a source, and the cheapest network that does it is a tree. Dividing its cost so that no group of users would rather build its own looks like a hard search, and it has a one-line answer: each user pays for the link that joins it to the tree on its way to the source. No group is ever overcharged — while the average over orders of arrival, the rule that settles so much else, can charge a pair more than its own connection costs.
AppliedFive weighings and the question is closed
Searching the triangle of splits can only ever fail to find a stable one, which is not the same as there being none. Weighing five families of coalitions against the whole settles the question outright — and the family that fails is the proof that nothing survives.
AppliedNone of the four conditions is spare
Four conditions pick out one sharing rule. The half that is usually shown is that they are enough; the other half is that each is needed — drop any one and a different rule satisfies the rest, so the list cannot be shortened.
AppliedOne glove too many
Give some people left gloves and others right ones, and let any group sell the pairs it can make. If the two sides are equal, the core — the splits no group can beat by walking out — is every price at once. If one side has a single glove more, the core is one split: the scarce side takes the whole of every pair and the other side gets nothing, however large the market. The average over orders of arrival barely notices the difference, and the two rules disagree about almost everything a market is.
AppliedSharing a cost that is not the sum of its parts
Three users need capacities three, six and twelve of one shared thing, and serving any group costs the largest of them. Averaging what each adds over every order of arrival divides the bill — and for this family the average collapses to a rule anybody could apply by hand.
AppliedThe corners are the orders of arrival
Line the players up, let each join in turn, and pay each what it adds on arrival: every order gives a split. When a newcomer always adds at least as much to a bigger group, those splits are exactly the corners of the core — so the core is never empty, it is the outline of the orders, and the average over all of them lies inside it. For a group worth the square of its size the outline is a hexagon whose corners are the six orderings of 1, 3 and 5.
AppliedThe objection nobody can make louder
When no split of the winnings survives every group's objection, the core is empty and the question changes — which split makes the loudest objection as quiet as it can be? Sorting the complaints and minimising them in dictionary order picks exactly one split, always, whether or not the core exists.
AppliedThe order everybody arrives in
Three people jointly earn nine, and the question is what each is owed. Ask instead what each adds on walking into a room the others are already in, average that over every order they could have arrived in, and four modest conditions leave no other answer.
AppliedThree splits that answer every other
When any two of three people can take a pound, no division of it is safe from a pair walking off. Von Neumann and Morgenstern's answer was not a division but a set of them: three half-and-half splits that never beat one another and between them beat everything else. It is a solution — and so is the line on which one player is held at any fixed share below a half, so the same game has infinitely many, each describing a different settled way of treating the third player.
AppliedToo many orders to list
The rule is an average over every order the players could have arrived in. At seven players that is five thousand orders and at twenty it is more than there are seconds in the age of the universe — so the average is sampled, and the error falls at a rate that can be measured.
AppliedWhat a missing input is worth
A model prices a house at 180 from its size, its garden and its bedrooms, and the question is how much of the price each input is responsible for. Make the inputs the players and the average over orders answers it — once somebody decides what the model says when an input is not known. Three reasonable decisions give bedrooms nothing, nothing, and sixteen, for a model that never reads them.