Liar's Poker: Rules, Strategy, and Wall Street's High-Stakes Game

Search Fintrev

Search our guides, reviews and glossary.

Liar’s Poker is a bluffing game played with ordinary dollar bills. Each player holds a bill and looks only at the serial number on it. Players then take turns bidding on how many times a particular digit appears across all the bills in play. A bid of “three sevens” claims there are at least three 7s among everyone’s serial numbers combined. Each new bid has to be higher than the last, and bidding continues until someone calls the previous bid a lie. The bills are then revealed and counted to settle who pays.

The game became part of Wall Street lore through bond trading floors, and it gave its name to Michael Lewis’s 1989 book about his years at Salomon Brothers.

How to Play Liar’s Poker

What you need

Two or more players and one bill per player. Any denomination works, but the classic version uses $1 bills, and that bill often doubles as the stake. A U.S. bill’s serial number contains eight digits (plus letters, which are ignored), so each bill contributes eight digits to the pool. With four players, the table holds 32 digits in total.

Players should draw bills at random, for example by taking them from a register or a stack, so nobody picks a bill with a serial number they already know.

Bidding

1. Each player checks their own serial number without showing it.
2. The first player opens with a bid made of a count and a digit, such as “two fives.” This means at least two 5s exist across every bill at the table.
3. The next player either challenges or makes a higher bid. A bid is higher if it names:
   • a larger count of any digit (“three twos” beats “two fives”), or
   • the same count of a higher-ranked digit (“two sixes” beats “two fives”).
4. Play continues around the group, with bids climbing, until someone believes the latest bid is false and challenges it.

A player’s own bill counts toward the total. If you hold three 8s, then “three eights” is a guaranteed bid for you, whatever anyone else holds.

Digit ranking

In the commonly described version, 0 is the lowest digit and 9 is the highest. Some groups count 0 as “ten” and rank it above 9 instead. Because the ranking decides which same-count bids are legal, the table should settle it before the first hand.

Challenges and payouts

When a bid is challenged, everyone reveals their bills and counts the named digit.

• If the bid is true (the count meets or exceeds the claim), the challenger was wrong and pays the bidder.
• If the bid is false, the bidder pays each player who challenged.

In multi-player games, a challenge often does not end the round immediately. The remaining players get a chance to either raise or also challenge, so one bad bid can cost the bidder several payments at once. The usual stake is one dollar per payment, typically the bill being played, but groups can agree on higher amounts before starting.

Examples

These hands are hypothetical and only illustrate the mechanics.

Example 1: A safe opening

Three players, so 24 digits on the table. Your serial number is 41275547, which holds three 4s. You open with “three fours.” This bid cannot lose, because your own bill already satisfies it. If anyone challenges, they pay you.

Example 2: Raising on the digit

The current bid is “four sixes.” You hold one 6 and two 9s. You could raise the count to “five sixes,” which depends heavily on other players’ bills. Or you could bid “four nines,” which keeps the count the same but moves to a higher digit you partly hold. The second bid relies on two other 9s turning up among the remaining 16 digits, a more comfortable position than needing four more 6s.

Example 3: Calling a bluff

Four players, 32 digits. The bid reaches “eight twos.” You hold no 2s. If digits are spread evenly, the table should contain about three 2s, so a claim of eight is far above expectation. You challenge. The bills show four 2s in total, so the bid fails and the bidder pays you (and anyone else who joined the challenge).

Liar’s Poker Strategy

Start with expected counts

If you treat each serial digit as roughly random, each digit from 0 to 9 has about a one-in-ten chance of appearing in any given position. That gives a simple baseline:

Expected count of a digit ≈ (number of bills × 8) ÷ 10

With five bills (40 digits), you would expect about four of any given digit. This is an average, not a guarantee. Actual counts vary from hand to hand, and serial numbers are not designed for games. Still, the formula shows roughly where bids stop being reasonable and turn into bluffs.

To adjust, combine what you know with what you don’t. Your own eight digits are certain. For the other players’ digits, apply the one-in-ten estimate. If you hold three 5s in a five-player game, a reasonable estimate for 5s is your three plus about 3.2 from the other 32 digits, so around six.

Use the bid ladder against opponents

The game rewards more than accurate estimates. Because every bid must go higher, the main skill is pushing the next player into a position where both raising and challenging are uncomfortable.

• Bid where you are strong, then switch. A bid on a digit you hold several of is both safe and informative to others. Sometimes you may want to bid on a digit you don’t hold so opponents misjudge your hand.
• Watch what others bid. A player who keeps returning to one digit probably holds some of it. Some players do this deliberately to bait challenges.
• Raise in larger groups. With more than two players, a frequent strategy is to keep raising rather than challenge. Challenging means betting that a specific claim is false, and a wrong challenge costs money. Raising passes the problem along, with the hope that someone else either challenges a weaker bid or overreaches.
• Mind the count of challengers. In versions where a failed bid pays every challenger, a bold bid on a crowded table carries more downside than the same bid between two players.

Psychology

Players are not simply trying to guess the true count. They are trying to get rivals to make mistakes. Confident tone, fast bids and patterns that are hard to read matter as much as arithmetic. This mix of probability, pressure and bluffing is why it was popular on trading desks: traders were already pricing uncertainty, reading counterparties and deciding when a quoted price looked wrong.

Tip: Played for money, Liar’s Poker is gambling. If you or someone you know has a gambling problem, the National Problem Gambling Helpline is 1-800-522-4700.

Michael Lewis’s “Liar’s Poker”

The book

Michael Lewis published Liar’s Poker in 1989. It recounts his time as a bond salesman at Salomon Brothers in the late 1980s. The book is nonfiction and semi-autobiographical: Lewis describes events he witnessed and people he worked with, written from his own point of view.

It portrays a trading floor driven by large bonuses, aggressive risk-taking and a culture that rewarded audacity. Salomon was a leading bond house, and its mortgage department turned home loans into tradable securities. That business, mortgage-backed securities, would later play a central role in the 2008 financial crisis, which Lewis covered in The Big Short.

The book’s most famous scene involves the game itself. Lewis writes that Salomon chairman John Gutfreund challenged bond trader John Meriwether to a single hand for $1 million. Meriwether instead proposed $10 million, and Gutfreund declined. The story has become shorthand for the risk appetite of the era. Meriwether later founded Long-Term Capital Management, the hedge fund whose 1998 collapse led to a Federal Reserve-organized rescue.

What Lewis intended

Lewis, who left Salomon to write, has said he meant the book as a cautionary tale about questionable and deceptive behavior at his firm. He has also acknowledged that some readers did the opposite and used it as a career guide, applying to Wall Street in hopes of the paydays it described.

Does Salomon Brothers Still Exist?

Not under its own name. In 1997, Salomon Brothers merged with Smith Barney to form Salomon Smith Barney. That firm then became part of Citigroup, serving as its investment banking arm. In 2003, Citigroup dropped the Salomon name entirely, partly because of its association with past financial scandals.

The Bottom Line

Liar's Poker matters to anyone curious about Wall Street culture or looking for a quick bluffing game that rewards both arithmetic and nerve. Players do best by anchoring bids to the expected count of about eight digits per bill divided by ten, then using the rising bid ladder to pressure opponents. Readers of Lewis's book should treat it as he intended, a cautionary account of 1980s Salomon Brothers rather than a career guide.