How it works

What is measured — and the honest answer is nothing

Twenty-four real public people, chosen by hand. Their standing curves are modelled, not measured — a smooth interpolation through a handful of hand-entered figures, flat before the first and after the last, with everything in between a modelling choice. No number here is a fact about anyone's finances, and none of the figures was read from its source: they are approximate recollections standing in for citations until a human checks each one.

Every figure behind these curves was typed in from recollection and rounded to about two significant figures. None was read off a publisher’s page. The roster is a model — closer in kind to a set of assumptions than to a citation list — and the code records that on every single anchor rather than in a disclaimer nobody reads.

The scale

Standing is unitless. 100 means the roster median at the first quarter of the window, and nothing else. It is not dollars, it is not a rank, and it does not convert to either. A person’s curve is a monotone cubic through the logarithm of their anchors, flat before the first and after the last, deflated to one currency and one basis so that two curves can be compared at all.

What the game actually uses is not the curve but the rank: your position among all 24 people, recomputed every quarter against an identical field. A hand is won by climbing more places than the other card over a drawn stretch. That is why the roster matters so much — a rank is a position among the people we chose, and a different roster gives a different answer.

The formula

• The deal. 24 people × 17 entry years = 408 cards, of which 372 can be dealt. The rest are refused because their person has too few anchors before that year, or because too much of the stretch would fall in a flat extrapolated tail.
• The stretch. 172 pairs of (settle quarter, judging window), enumerated exhaustively. Your card is graded against the other over one of them, drawn after you decide.
• The equity. Ahead scores 2, level 1, behind 0, summed over all 172 stretches — so equity is an integer count out of 344 and two opposing counts always add to exactly that. There is no sampling anywhere in this product.
• The price. The other seat calls when its equity beats (S−1)/2S, which depends on the stack and on nothing else. That is the only number here that would read identically in a real poker room.

The certified charts

Built from 83,436 exhaustively enumerated cells. The solve is exact; the inputs are modelled.

Who is on the roster, and why

Everyone here is a public figure, chosen for having a trajectory shape the game needs — an early peak, a late bloom, a collapse, a steady climb. Nobody was included because of a scandal, nobody was a minor at any point in the window, and nobody whose circumstances are a live legal matter is here.

• Andrew Lloyd WebberComposer of stage musicals · 7 anchors

  Was already near the top of the roster when the window opens and barely moves after that, so his standing falls simply by being passed. The clearest case of a flat arc losing places without going anywhere.

• Bill GatesCo-founder of Microsoft · 8 anchors

  Peaks early in the window and then gives much of it away while later fortunes climb past, so the arc rises in level and slides in the standings at the same time. The roster needs at least one card where those two things disagree.

• Damien HirstVisual artist · 7 anchors

  A sharp peak in 2008 and a fall straight after it, which is the shape the mechanic bites hardest on: back him a year early and the short judging windows are kind, back him a year late and none of them are.

• Dolly PartonCountry singer, songwriter and businesswoman · 6 anchors

  A long, even climb from the bottom half of the roster, which is the arc every short judging window reads as nothing happening. The deck needs cards that are quiet by construction and not by accident.

• George LucasFilmmaker and studio founder · 7 anchors

  One step change in the middle of the window — the sale of his studio — with flat stretches either side. A card whose whole strength is which side of that step your entry year lands on.

• Jay-ZRapper and music-business executive · 6 anchors

  Starts near the bottom of the roster and accelerates late, so his best decades are the ones bought in the middle of the entry range rather than at either end.

• Jeff BezosFounder of Amazon · 8 anchors

  A spike in 1999, a collapse in 2001 and then the largest sustained climb on the roster. Three different arcs in one card, which is why the entry year matters more here than anywhere.

• Jeff KoonsVisual artist · 6 anchors

  Climbs steeply through the middle of the window and then flattens, which is the mirror image of the early-peak cards and gives the long judging windows something to disagree with the short ones about.

• Jerry SeinfeldComedian and television writer · 7 anchors

  A step up as his series went into syndication, then a long plateau. Backing him one entry year apart is close to the cleanest before-and-after pair in the deck.

• Larry EllisonCo-founder of Oracle · 8 anchors

  Climbs across the whole window but does it in violent steps, so the same card is a rise or a fall depending on where the drawn stretch happens to start.

• MadonnaSinger and songwriter · 7 anchors

  Rises through the first half of the window and then holds station while the roster around her moves, so her rank falls without her level ever falling much. The disclosure about roster-relative ranks needs a card that demonstrates it.

• Masayoshi SonFounder of SoftBank · 8 anchors

  The sharpest collapse on the roster: a very large 2000 figure and a very small 2001 one, then a slow recovery and a late slide. Any deck whose thesis is that timing decides everything needs one card this violent.

• Michael DellFounder of Dell Technologies · 8 anchors

  An early peak, a decade of going sideways, then a late climb. Backing him from 1998 and from 2010 are almost opposite bets on the same person, which is the point of the entry year existing at all.

• Michael JordanProfessional basketball player, later a team owner · 7 anchors

  The roster's late bloomer: most of the climb arrives long after the career that made him public, so the arc a player reads before deciding says very little about the decade they are backing.

• Oprah WinfreyBroadcaster and media proprietor · 7 anchors

  A steady climb through the first two thirds of the window and then a plateau, which puts her at the top of the performers early and holds her there while the founders pass her. A good card for showing that a plateau is not a fall.

• Paul McCartneySinger, songwriter and bass player · 7 anchors

  High at the window's start and almost level throughout, so every judging window over him is close to a no-contest. The runout table needs cards that produce level results honestly rather than for want of data.

• Phil KnightCo-founder of Nike · 7 anchors

  A slow compounder for fifteen years and then a fast one, which is the arc the long judging windows reward and the short ones cannot see. Half of the runout table exists to tell those two apart.

• Ralph LaurenFashion designer and founder of his company · 8 anchors

  Climbs, plateaus, dips and recovers, all within the same order of magnitude — the arc that is hardest to read from the panel, and therefore the fairest test of whether a player is pricing the card or the story.

• Reed HastingsCo-founder of Netflix · 7 anchors

  His figures only begin once his company was public, so his early entry years are simply not dealable — and that is correct rather than a gap to be papered over by inventing an anchor. The deck needs one card that demonstrates the rule.

• Richard BransonFounder of the Virgin group of companies · 8 anchors

  Rises for twenty years, flattens, then falls back when air travel stopped. One of only two cards on the roster whose last decade is a decline, which the shove/fold decision needs in order to be a decision.

• Stephen KingNovelist · 6 anchors

  Nearly flat for thirty years in the lower middle of the roster, so his rank is decided almost entirely by what everyone else did. The plainest illustration of why a rank is partly a property of the people it was measured against.

• Steven SpielbergFilm director and producer · 7 anchors

  A long, even rise with no reversal anywhere in the window — the control card. If a player is losing big blinds on this one, the leak is in how they price rather than in what they read.

• Tom CruiseFilm actor and producer · 7 anchors

  Doubles early and then grows slowly for two decades, which puts most of his judging windows in the narrow band where the break-even equity actually bites. Cards near the threshold are where dEV is earned or given away.

• Warren BuffettInvestor and chairman of Berkshire Hathaway · 7 anchors

  The compounder: near the top of the roster at every quarter in the window, with small dips and no collapse. Almost every stretch over him is a hold rather than a climb, which is exactly the card the short judging windows go silent on.

Corrections and removal

If you are on this roster and would rather not be, or a figure behind your curve is wrong, we will take you off. Removing a person is a one-line change to the seed list — but it is not cosmetic and we will not pretend otherwise: every rank is a position among everyone else, so taking one person out changes every other person’s curve, every climb, and every historical score in the game. Those hands are re-scored against the new roster or discarded, never quietly left standing against a roster that no longer exists.

Raise it on the repository this is built in, or through whoever sent you the link.