So you're sitting around with your friends and a deck of cards, looking for a game to play. You're sick of such retreads as hearts and spades, and you're not the gambling type. You consult among yourselves only to find that each of you knows only a couple of card games, and there isn't a single card game that you all know. Well, we have the answer! In this compact book we present the ten best card games you've never heard of, for you and your friends to learn together and laugh together. While these games are long, they are by no means grueling; rather, they are fun games with an emphasis on quick thinking and intuitive evaluation. Most importantly, all of the games here have one or more unique features; this is not a book containing a hundred variants of poker, but rather a book with ten very different novel games, each of which has the potential to provide you with countless hours of fun.

If you just want a game to play and don't want to read through all of this... well, games in general fall into three types. There are the fun games, which are fun to play and a good time, not requiring much strategy or deep thought. There are the games for inveterate card-players, that involve lots of counting and knowing how suits typically split and which have significant cardplay components; call these "veteran games." Then there are the games which are simply unorthodox, that aren't frivolous games but which don't have a big advantage to veteran cardplayers; call these "abstract games." The last two categories are, of course, fun, but in a way that is perhaps more appealing to people who like playing cards than to the general population.

Of the games in this book, Creights and Spielen are definitely fun games. Barbu, 99, and Oh Hell are definitely veteran games. Psychological Jujitsu and Canadian Fish are definitely abstract games. Of the remaining three, Ninety-nine is somewhere between a fun game and an abstract game, Napoleon is somewhere between a fun game and a veteran game, and Liars' Poker is a mix of all three. So if you're just looking for one game, that's the all-purpose preview.

Jump to:

-1. What is a Deck of Cards?
Whenever we refer to a deck of cards in this book, we mean a standard 52-card deck. This contains four suits (clubs, diamonds, hearts, and spades) and thirteen ranks (2, 3, 4, 5, 6, 7, 8, 9, 10, jack (often denoted "J"), queen ("Q"), king ("K"), and ace ("A")). The 52-card deck contains one card for each suit-rank combination. Jacks, queens, and kings are referred to as "face cards", as the cards usually contain pictures of people, while the number cards (2 through 10) are occasionally referred to as "spot cards." The order of the ranks is usually as stated, although sometimes the ace is low instead of high (or worth 1, when spot cards are worth their number and jacks, queens, and kings are either all worth 10 or worth 11, 12, and 13 respectively.) In some games, there is an ordering of the suits, although in most games all suits are created equal.
0. The Basic Idea of a Trick-Taking Game
If you are familiar with trick-taking games such as hearts, spades, or bridge, you do not need to read this section.
One of the more common types of card games is trick-taking games. The basic concept of a trick is as follows: one player leads a card, and then each player in turn (usually clockwise) must play a card of the suit led. If they do not have any cards of the suit led, they may play any card. Some trick-taking games have what is known as a trump suit; this suit is higher than all the others, and if there are any cards of this suit in a trick, the person who played the highest card of this suit wins the trick. Playing a trump to a trick that is not of the trump suit (which necessitates being void of the suit led) is known as ruffing or trumping; trumping a trick that has already been trumped with a higher trump is called overruffing or overtrumping. Playing a lower trump in a trick that has already been trumped is rare, and is known as underruffing or undertrumping. If there is no trump suit, or no trump is played in the trick, then the highest card of the suit led wins the trick. The effect of winning the trick is twofold; first of all, you collect the cards of the trick, and second of all, you lead the first card of the next trick.

Trick-taking games have various goals, but the basic idea is that at the end of the hand, points are awarded based on the cards that people have collected in the tricks they have won. The two main subtypes types of trick-taking games are those where the results are based on who has taken certain cards in their tricks (as in hearts) and those where the results are based on how many tricks players have taken (as in spades or bridge.) A complete description of any trick-taking game should also contain instructions on who makes the first lead; this can be determined by actions prior to the beginning of the play (as in bridge), who has a given card (as in hearts), or may simply be the dealer or player to the left of the dealer.

1. Creights
Creights is a speed-based variant of Crazy Eights. Perhaps "variant" is not strong enough of a word; consider the classic game of Crazy Eights, and ascribe an additional rule for virtually every card, and you end up with creights, a game which is guaranteed to liven up every party. The one drawback of the game is that people occasionally start thinking, and creights is really no fun as a strategy game.
Players: 3 to 5, 5 to 8 with double-deck rules.
Playing time: Depends almost entirely on the speed of the players. 1-3 hours for the long version, half that for the abbreviated game with four players.
The Deal: Creights consists of fifteen deals. On the first hand, each player receives eight cards; on the second hand, seven, and so forth. On the eighth hand each player receives only a single card, and the number of cards goes back up to eight, increasing by one on each ensuing deal. In the short version of creights, only the first eight hands are played. Deal rotates to the left. If the dealer deals too many cards, he or she must keep all extra cards; if the dealer deals too few, he or she receives a dumbass card (see later.)
The Play: At the beginning of the game, the dealer turns up the top card of the deck, and all remaining cards are placed face down in a draw pile. The player to the left of the dealer ordinarily starts, with play proceeding clockwise. In general, on his or her turn, a player must make a legal play if possible on top of the play pile; if not, he or she must draw a card. Legal plays are usually cards of the same rank or suit as the top card of the play pile; see below for exceptions. The deal typically ends when one player gets rid of all their cards.

The main feature of creights is that nearly every card does something to the game when played, as follows.

So, as you can see, creights is not your garden-variety backyard game of ringtoss. It will probably take you a few deals to get comfortable with the rules; until then, we recommend you keep the rules for what the cards do on a piece of paper nearby.

In addition to the play itself, there are a few other rules in creights. First of all, whatever card the dealer turns up counts as if the dealer had played it (with play going clockwise.) Thus, for instance, if a six is played, the dealer starts play; if a ten is played, the player to the right of the dealer starts with play going counterclockwise; if a five is played, each other player must draw an additional card starting with left-of-dealer; if a two is played, the deal begins in a crank with value two, and so forth. If an eight or nine is played, dealer must call a suit before looking at his or her hand. In the case of a nine, dealer may call any suit.

Now, with all these rules, you might wonder what the penalty for erring is. The answer is appropriately named dumbass cards. When a player does something stupid, he or she is eligible to receive a dumbass card (given by any other player.) Such events include but are not limited to: calling an inappropriate suit when playing a nine, making an illegal play (such as the jack of spades when the suit is not spades and the card showing is not a jack; you'd be surprised at how often this can happen), being completely unaware that it is your turn, playing out of turn, not calling the suit before looking at your hand on a wild as dealer, and just generally doing anything stupid. As you gain experience, you probably also want to institute the five-second rule, giving a dumbass card for taking too long. We recommend that on a player's first game of creights, he or she not be given dumbass cards, or at least for the first few deals. You do not get dumbass cards for drawing when it is not your turn; instead, you get to keep these bonus cards. This can in fact (if you're clever enough) be a way to trick others into thinking it's their turn, giving them a dumbass card (a useful ploy if someone is about to go out.)

Another miscellaneous rule is the one-card rule. If at any point during your turn other than the beginning you will have exactly one card in your hand, you must say "one card" (or "last card") before making any motion to play cards out of your hand. Failure to do so will result in a very harsh penalty; at your next non-crank turn, you are skipped and must instead draw two cards from the top of the deck. Situations in which this applies include the following.

The one-card rule applies as usual during a crank. The final situation brings up another rule of creights; when a person plays his or her last card, all of the effects of that card are carried through before the game ends. If, as in the situation with the six, this results in the person no longer having no cards, the game continues. Other situations in which this can happen is if the person plays the nine and calls an inappropriate suit (receiving a dumbass card), or if a person goes out during a crank and the crank comes back around to them, at which point (having no crank cards left) they are back in the game with a lot of cards. Similarly, if a person plays a five, the other players each draw a card before the hand is scored, and if a person plays a seven to go out, the person across from them gets a card. If in addition the last card played is the seven of diamonds, that person not only receives an extra card, but owes a beer to the person going out. This is a very important rule, and should not be dismissed. The last, and possibly most important, part of creights, is shuffle pressures. If at any time you have to draw a card and there are no cards left in the draw pile, you receive a shuffle pressure. This has three consequences:
  1. You must shuffle the deck. Take the top card, leave it face up, and shuffle the rest of the play pile, placing it face down as the new draw pile.
  2. Everyone laughs at you.
  3. For your first shuffle pressure, you get five points. Each additional shuffle pressure is worth twice as much as the previous one, so the second is worth 10, the third 20, and so on. Shuffle pressures are cumulative across rounds. If you get two shuffle pressures in the first round of eight, and then another in the round of seven, this one shuffle pressure is worth 20 points, not 5. Shuffle pressures in a round are denoted by a dot (one for each shuffle pressure) next to the player's score for that round.
If you need to draw, but there are no cards in the draw pile, and in addition there is only one card in the play pile, you get a shuffle pressure and the round ends immediately. Note that for this to happen 51 of the 52 cards must be in someone's hand.

And that's it for creights. This may seem like a lot of rules, but they can be internalized fairly quickly.
Scoring: As previously mentioned, the round ends when one player is bereft of cards. Each player then receives points for each card they have in their hand (the player going out receives 0); the object is to accumulate the fewest total points in the fifteen hands. In general, the better a card is during the play, the worse it is to have in your hand at the end of the game. The scoring is as follows:

Strategy: Creights is not inherently a game of strategy, and what little strategy there is you will figure out very quickly. In general, 3's and aces are good cards to keep in one's hand (by not playing them if at all avoidable); 3's because they can drastically reduce your score, and aces because they protect you during a crank and are low-point cards. One of the beautiful things about creights is that the game is balanced, as mentioned before, so that the better a card is to play, the worse it is for you during scoring. In particular, keeping 8's and 9's will ensure that you are always able to play, but keeping too many on hand is a recipe for disaster if someone goes out. One of the most important elements of creights is the beer rule; this happens just enough (about once per 15-hand game) that it doesn't have a huge effect, but isn't entirely frivolous. For that reason, playing the seven of diamonds not as your last card "ought" to be avoided.
1a. Double-deck rules
If you're going to play hearts with more than five people, you ought to use two decks (for exactly five people, the more adventurous can try playing with just one deck, but two is standard.) It is possible to play with identical rules, but some people play with the following modifications.

First, one set of fives is removed from the deck to eliminate excess blood (however, both sets of crank cards are left in, so cranks can now go up to 24 cards.) The other two modifications concern playing identical copies of the same card. If it is your turn, you may play two copies of the same card (provided, of course, that the card is legally playable.) The first copy acts naturally; the second copy, on the other hand, counts as if the next person to play had played it; thus, if it is a seven, the player across from them gets a card, if it is a five, everyone other than them gets a card, and so forth. Their turn is then skipped. Note that if you play a pair of identical tens, the first ten played reverses order, sending it to the player who just played; the second ten reverses order again, though, making it your turn again.

The second case of this occurs when you play a card identical to the card just played. It then remains your turn, and your play (which may be drawing if you have no legal plays) takes the place of the next person's turn. This may be combined with the previous rule, so that you may actually end up taking the turns of the two players after you.

These rules do not apply to sixes; sixes in double-deck creights function just as in single-deck creights.

2. Barbu

Barbu is a very involved strategy game. Hailing from an Old French card game of the same name, its main distinguishing feature is that it is composed of eight simple but different card games; the complexity in the game lies not only in the play of the cards, but in judging how good one's hand is for each game. As such, Barbu has an element of strategy absent from most other games, while at the same time remaining simple in structure. It is intellectually engaging, but nowhere near as taxing as other such games, like bridge.
Players: 4.
Playing time: 2-3 hours.
The Deal: Barbu consists of thirty-two deals. Deal rotates to the left. On each deal, the dealer, after examining his or her cards, chooses which of the eight subgames will be played. Each dealer must choose each of the eight games exactly once in his or her eight deals.
Doubling: After the dealer deals out thirteen cards to each player, there is a round of doubling. Each player in turn, starting with the player on dealer's left and proceeding clockwise, can choose to double any subset of the other three players (or pass, doubling no one.) If you are doubled by another player, you may choose to redouble them. Dealer may not double anyone, but may redouble anyone who doubles him. After the round of doubling, there is a round of redoubling, in which players may choose to redouble players who have redoubled them (but no new doubling can take place.) Doubling (and redoubling) is symmetric; if I double you, it is the same as if you double me. The effect of doubling is as follows: if you and I are doubled, then the difference between my score and yours in that hand is added to my score and subtracted from yours. So, for example, if we are doubled and I score -2 in the round, and you score -10, then I get in addition to my -2 eight points, and you lose eight points; if neither of us is doubled with anyone else, I would thus get +6 and you would get -18. If we are redoubled, double the difference is added to my score and subtracted from yours. Note that doubling does not change the total points awarded in the hand. Now, you might ask why you would want to double the dealer; having chosen the hand, dealer in all likelihood has a better hand for it than you do. Well, first of all, in a few games only one player gets negative points, and if you are sure it won't be you, you may as well double the dealer. The other, more important reason, is that during the game, each player must double every other player twice out of the eight hands that they deal. It is desirable to get these "dealer doubles" out of the way quickly, as otherwise you will be forced to double the dealer on his last two deals.
The Play: Barbu is composed of eight subgames. After the deal, the dealer chooses the game, and the doubling and redoubling, if any, takes place. Dealer then makes the first play, and play continues counterclockwise. All but one of the eight games are trick-taking games, with scoring as explained below. Note that in many of the games, only a few cards matter; the hand can be thrown in as soon as these cards have been collected.

Barbu: This is the namesake subgame of the overall game, and is quite simple. It is a trick-taking game where the player who takes the king of hearts gets -15, and all other players get 0. In this game, no player can ever lead hearts unless they have only hearts left.
Hearts: This game is quite similar to the traditional game of hearts. It is a trick-taking game, where each heart taken is worth -2, and the ace of hearts is worth -6. Unlike the game of hearts, there is no queen of spades or shooting the moon. No heart can be led until a heart has been played or the player on lead has only hearts left.
Queens: In this trick-taking game, all queens are worth -6 points to whoever takes them. The other 48 cards are worth nothing.
Nullo: The object of this trick-taking game is simply not to take tricks; each trick is worth -2 to its winner.
Last Two: In this trick-taking game, the first 11 tricks are worth nothing. The twelfth trick, however, is worth -10 to its winner, and the thirteenth trick is worth -20. While this may seem simple, Last Two is probably the most complicated of the eight subgames.
Ravage: Also called Ravage City, this is the most unique of the eight subgames. In this trick-taking game, the player that takes the most cards of any one suit gets -36, and all other players get 0. For example, if I have nine diamonds, and no one else has nine or more of any suit (in the tricks we've taken), I lose ("get ravaged"), and take the -36. If two or more players tie, they split the -36 evenly, regardless of how many suits each has been ravaged in.
Trumps: In this trick-taking game, there is a trump suit, which dealer names when he chooses the game, before the doubling round. The object here is simply to take take tricks; each trick is worth +5. Unlike in some other games, if you can trump you must. If you are playing to a trick that has already been trumped, and you can overtrump, you must, but you are not required to undertrump.
Dominoes: Dominoes is the only non-trick-taking game of the eight subgames. In dominoes, when dealer calls the game, he also calls a "pivot card," which is a rank (so, for instance, "Dominoes around the nine.") On each turn (starting with dealer and moving clockwise), a player may either play any pivot card (card of the specified rank), or a card which is of the same suit and either one higher or one lower than a card already played. Aces are always high and can be played only on kings of the same suit (unless they are the pivot card); similarly, deuces can be played only on threes. All cards played are in open view, in four rows (corresponding to the suits) face up on the table. If you cannot play a card on your turn, you must pass. The object of the game is to get rid of all of one's cards first. The first player to go out gets +30 points; the second player, +20, the third player, +10, and the fourth player, 0.

Scoring: Scoring for each of the subgames is as above. After scores for each subgame are determined, the effects of doubling are imposed (simultaneously.) In general, points from doubling outweigh points from the games (which tend to even out over the course of 32 deals), and thus a lot of the strategy lies in knowing when to double. These scores are then added to a running total. As there are a lot of things to keep track of in a game of Barbu, we suggest the following method of keeping score: first, have one column for each player. Mark off eight lines corresponding to the eight subgames, and cross out the appropriate box when the dealer calls the game; this keeps track of which dealers have called (and therefore cannot call again) which games. In each column, also put two copies of the names of the other three players; these stand for dealer doubles, and are crossed out whenever a dealer double is made. Below, simply add the score of each subgame to a running total after the subgame is completed. After calculating each score (modified for doubling), you may want to check to make sure you have it right, as it is very frustrating to complete a game of Barbu only to discover a scorekeeping error which may be unrecoverable. A handy way to do this is to sum the players' scores; this is unaffected by doubling, and thus is only dependent on what games have been played. The total score for each game should be as follows:

When you are done with a game of Barbu, each game has been played four times, so the players' scores should sum to -144.
Strategy: The strategy in Barbu lies almost entirely in hand evaluation. If you know which hands are good for which games, you will be greatly aided in the processes of game choosing and doubling. Fortunately, as all of the games are fairly simple, it is in general not too difficult to determine which game is best for your hand as dealer. However, there are a couple of pitfalls.

First, it is worth noting that only one of the eight games favors having high cards; in six of the games, one wants to avoid winning tricks. The eighth is dominoes, and because of this imbalance one generally calls dominoes with a high hand. While it may seem that trumps is more an escape route for a generally high hand, in these situations it is often better to play a low-risk game such as Barbu; with judicious use of trumps, dealer ought to win seven or eight tricks. An ideal hand for Last Two can also contain many high cards; as long as they're complemented by twos or threes to relinquish the lead, having aces and kings allows you to get control of the hand and play out your problem suits to ensure they are not led in the twelfth or thirteenth trick. Similarly, you can get away with high cards in Hearts; the ideal hearts hand is one with lots of hearts, including some low ones so that you take no heart tricks. Often, you will find that no game is particularly good for your hand; in this case, put on a poker face and call something that is unlikely to hurt you, such as Barbu, Nullo, or Queens. Getting -6 as dealer playing Queens is not nearly as bad as getting -36 as dealer playing Ravage - or probably more, since it is likely that if you have a mediocre hand, one or more of the other players will have a good hand and double you.

The second part of Barbu strategy lies in the doubling. In general, doubling the dealer is best done in games where only one player loses (Barbu, Ravage, typically Last Two) and you are sure it won't be you. But beware on Ravage; you should have in each suit either two low cards (7 or below) or only one high card (9 or above) and one low card. It often happens that each player is confident about his or her hand for Ravage, with the end result that all players end up redoubled with each other. In this case, the player that gets ravaged gets -252, with each other player getting +72. This 324-point swing is the largest possible in any subgame, and the player getting -252 is virtually eliminated from contention. This is also the case in Last Two, with the loser getting -210. So, if you are doubling in Ravage or Last Two, you should be as close to sure that you will not lose as possible. If there is even a hint of a doubt, do not redouble.

Other good games on which to double the dealer are Nullo and Queens. Because these games do not provide large swings, if you think you will take at most three tricks in Nullo or at most one queen, it is probably worthwhile to use a dealer double (and double all of the other players as well.) Never double the dealer on Trumps; even if you have a very good hand, the dealer probably has a better one. Be careful about doubling the dealer on Dominoes; in all likelihood he or she will control the hand, and your one two may come back to haunt you.

The game is largely determined by the doubling. Do not be afraid to double on instinct. Better than instinct, however, is table knowledge; if a player passes, he or she likely has a bad hand, and even if you have a hand that is below average you should consider doubling them. Likewise, if a player doubles you, figure that player for an above-average hand and redouble only if you have a very good hand. Towards the end of the game, if you are losing by a lot you may need to double to get back in the game. Conversely, if you are winning, you don't want to be forced to double the other players when they're dealing, so it is a good idea to get those dealer doubles out of the way as quickly as possible. You should not confine your dealer doubling to hands where you have a great hand for what the dealer has called; if you get no points in the hand, the dealer double cannot hurt you. Otherwise, in the eighth round, you must be forced to double the dealer with a horrible Ravage hand; if the dealer redoubles, you will go down at least -108, even without the other players also doubling. This is quite undesirable.

3. Ninety-nine
Ninety-nine starts out as a very simple game, but grows in complexity with each deal that is played. The beauty of this game is that the players control how the game will change, making it a game of creativity rather than analytic thinking, something very rare among card games. The game is not to be confused with 99, also in this book; we have made the distinction between the two here by representing one by letters and one by numbers, but in practice both games have the same name.
Players: 2 and up. Probably best with 3-5.
Playing time: Ninety-nine is a game of indefinite length; it lasts as long as the players want it to. It becomes rather untenable after about 10-15 hands, which should take about an hour.
The Deal: At the beginning of each hand, each player receives three cards. The actual identity of the dealer is irrelevant; the player that won the last hand goes first (and typically deals also.) For the first hand, you can use your favorite metric (oldest, youngest, longest beard, person whose birthday is coming up soonest, etc.) to determine who goes first.
The Play: As stated before, ninety-nine starts out as an incredibly simple game. Play starts with the player that won the last hand, and continues clockwise. On your turn, you play a card face up and draw a card. The card played adds to a running total, which starts at 0; you announce the total as you play the card. Aces through tens add their numeric value. Jacks and queens add 10 each. Black kings can add either +10 or -10, as determined by the player who plays them; red kings are -10. The only restriction on playing a card is that you cannot play a card that makes the total greater than 99; if you have no legal plays, you are out of the hand. The last player remaining in the hand wins. Note that you draw a card immediately after playing one; if you forget to draw this card before the next player plays, you do not get a replacement card and are left with one fewer than you started. As more cards gives you more flexibility, it is desirable to preserve your original quota of three, so do not forget to draw.
After the Hand: The defining part of ninety-nine comes after the hand. After each hand, the winner of the hand invents a new rule which then becomes a rule all subsequent hands. This rule can be anything at all; the only constraint is that the rule must be symmetric to all players (e.g. the rule cannot be of the form "All players whose names start with M start with four cards instead of three.") Also, the rule should not give a sizable advantage to the player who plays first, lest the game degenerate into a trivial win for the player who won the last game. Some common types of rules are those that do something to the order of play (e.g. "Playing a seven reverses the direction of play"), things that do something to the total (e.g. "Playing a two can, at the player's discretion, add 12 to the total instead of two"), things that do something to the number of cards (e.g. "If you play a nine, you don't get to draw a replacement card"), and constraints on the total (e.g. "The total can never be a multiple of ten.") But these are just examples; the scope of possible rules is limited only by your imagination. If at any time a play is made which violates a rule (initially, only plays that put the total above 99, but the number of illegal plays often increases with the number of rules), the illegal play is discarded, the player does not get a replacement card, and it is still that player's turn. (This typically happens when a player forgets some rule or another.) If a player has no cards, he or she a fortiori has no legal plays, and is out of the hand.
Scoring: Ninety-nine does not have actual scoring, just as it has no predetermined ending point (a logical point to stop the game is when the rules become sufficiently complicated that players have forgotten most of them.) If you're a very competitive person, you probably shouldn't be playing this game, but if you must, you can score by giving each player one point for each hand won.
Strategy: There is no scoring in ninety-nine, and thus there really is no strategy also. Try to avoid playing kings, as having the ability to decrease the total as it approaches 99 is invaluable; keeping low cards is also good for this. Always remember to draw a card immediately after you play one, as you will feel quite silly (and in all likelihood lose the hand) if your opponents have a stash of two cards on hand, while you are forced to play the card that you get. It goes without saying that remembering the rules is tremendously helpful.

4. 99
99 is a delicious medium between a card-counting game where all 52 cards are in play, such as bridge, and a game of imperfect information where there is a lot concealed, such as poker. In this trick-taking game, each player strives for an individual goal unknown to the other players, with asymmetry between suits providing complex nuances of strategy. This game is not to be confused with ninety-nine, also in this book; we have made the distinction between the two here by representing one by letters and one by numbers, but in practice both games have the same name.
Players: 4. Can be adapted to five, but loses much of its charm; if this is done, remove the twos of diamonds and clubs from the deck beforehand.
Playing time: 45-90 minutes. Like most card games, the playing time of 99 is largely dependent on the speed of the players.
The deal: Deal starts with an arbitrary player, and rotates to the left. All 52 cards are dealt out, with each player receiving thirteen cards.
Bidding: Each player sets aside three cards, which represent the number of tricks the player proposes to take in the play of the hand. These three cards per player are out of the game, and are not revealed until the end of the hand. The bidding is done simultaneously and independently, with this round being over when each player has decided on his or her bid. The cards encode the bid as follows: a diamond is worth 0 tricks, a spade 1, a heart 2, and a club 3, with the bid being the total of the suits of the three cards. Note that there are 10 tricks in total (as each player plays with a ten-card hand); three diamonds can represent either 0 or 10 tricks bid. A handy mnemonic for this is that spades have one "point," hearts two, clubs three, and diamonds ... are mostly round.
The Play: Each hand has a trump suit, which is diamonds for the first hand, then spades, hearts, and clubs. The fifth hand is no-trump, and then it wraps around back to diamonds. (This order corresponds to the order of the suits in the bidding.) Player to the left of the dealer makes the opening lead, and play proceeds clockwise. During the play, each player aims to get exactly the number of tricks he or she bid before the play.
Scoring: At the end of each hand, each player gets a number of points equal to the number of tricks they have taken. More importantly, each player who made his or her bid exactly gets a bonus of 10 points. If a player has reached 99 or more by making their bid, and they have the most points, they win. In practice, this will be the first player who reaches or exceeds 99 points, except that a player cannot win unless they made their bid on that hand.
Strategy: 99 is a simple game, but the strategy is quite complex. Unlike most trick-taking card games, the suits will in general not split evenly; typically, spades are the least numerous suit, followed by diamonds, hearts, and clubs. Indeed, it is quite common for a player to be void in spades at the beginning, or at the very least have a singleton; contrariwise, it is not out of the question for all 13 clubs to be in play. Make note of this when determining your bid, as if you have the ace-king of spades (and spades are not trump), you should definitely not count on scoring both of them.

Putting aside the three cards serves a dual purpose. Not only do these cards determine your bid, but they also provide you with a way to modify your hand. This interplay occasionally results in a sticky situation, especially when diamonds are trumps. In this case, often a player will want to bid low, but will be unable to do so without discarding diamonds that make their hand even worse. A similar situation arises when spades are trumps.

More often, however, you can make use of this by discarding cards of the appropriate rank. It is a naive strategy to always determine how many tricks your hand can take, and discarding the lowest cards allowing you to make that bid; frequently, you can be more sure of making your bid by discarding jacks and queens, or even aces and kings, and making a lower bid than one might otherwise. Discarding high cards also has the effect of disrupting the other players' planning; the vast majority of the points scored come from making one's bid exactly, and if the other players end up taking more tricks than they bid, that's good news for you. Similarly, you want to try to void yourself in a suit if possible (if you have two spades and want to bid 2, discarding both of them and a diamond is preferable to discarding a heart and two diamonds), for this will both provide a trick source and nullify opponents' aces that they are no doubt counting on. Do not be afraid to discard trumps if necessary; this, again, will disrupt opponents' logic.

5. Oh Hell (also known as Scotch Bridge)
Oh Hell is probably the best-known of the games mentioned in this book. However, it is still rarely played, and seems to be dying today; a shame, as it is one of the more intriguing trick-taking games in existence. Furthermore, like Barbu, it is in and of itself a diverse card game, with strategy for card play in each hand quite different, which makes playing a game of Oh Hell far from a monotonous drone.
Players: 3-5, best with four.
Playing time: About 2 hours for the long game, half that for the short game.
The Deal: With four players, Oh Hell consists of 25 deals. First dealer is determined however one desires, and deal rotates to the left. On the first deal, each player gets one card; on the second deal, each player gets two cards, and so on up to the thirteenth deal, where each player gets 13 cards. In the remaining twelve deals (which are not played in the short version), the number of cards goes back down from 13 to 1. (In the three-player version, the game lasts 33 deals, with the number of cards going from 1 to 17 and back to 1; with five players, the game is 19 deals, with number of cards per player going from 1 to 10 and back to 1.) In all but the thirteenth deal, the top card of the remaining pile (which is otherwise unused) is turned up, and its suit becomes the trump suit for the hand. In the thirteenth deal, there is no trump.
Bidding: Each player, starting with the player to the left of the dealer and continuing clockwise, announces the exact number of tricks he or she proposes to take in the hand; this is noted on the scoresheet. Each player is free to bid anywhere between 0 and the number of tricks in the hand, except dealer; dealer is not allowed to bid the number which will make the total number of tricks proposed equal to the number of tricks in the hand. This rule, known as screw the dealer, is not used in some variants, but adds an element of spice to the game by ensuring that at least one player will fail to make their bid.
The Play: Once the bidding is completed, the hand is played. Player to the left of dealer leads to the opening trick, and play proceeds as a trick-taking game, with as many tricks being played as there are cards in each player's hand.
Scoring: As in 99, making one's bid exactly is the goal; in fact, this is more important than in 99, as each player gets points only if they bid the exact number of tricks they take. There are many scoring systems used in Oh Hell. The one which is most well-balanced is as follows: for a bid of zero, a player receives the number of tricks in the hand (which reflects the added difficulty of making this bid as the number of tricks in the hand increases.) For any other bid, a player receives twice the number of tricks bid, plus a three-point bonus. Of course, these points are only obtained if the player makes the bid exactly; otherwise, the player receives nothing. The game ends after all 25 hands have been played, and the player with the most points wins.
Strategy: Strategy in Oh Hell is very different in the low-card (1-5), medium-card (6-9), and high-card (10-13) hands. When there are few cards in the hand, one has very imperfect information; an added element here is that the lead becomes incredibly important. In general, having trumps (even low trumps) is very important, as it is unlikely that the other players have many; furthermore, if they do have one or two trumps, it is quite likely that they have a void in some suit or another, and will use their trumps to take those tricks. Contrariwise, if you have high cards in other suits, the odds are high that they will be ruffed if the other players choose to. In the round of one, the player leading determines the suit in play; if you are not the first person to play, and you do not have a trump, you should bid zero as it is unlikely, even if you have an ace, that the player on lead will lead a card of your suit.

In the medium-card hands, it becomes likely that a suit will go around at least once before it is ruffed. However, it is still unclear what a high card is; in particular, it is quite difficult to know whether or not you can count on a king for a trick. Similarly, low trumps also become hard to value, as some other player may have high trumps and choose to lead them out. The medium-card hands are probably the most difficult to play, which explains why we offer the least advice for them.

In the high-card hands, you should generally count on any given card being in play. You can also bank on each suit going around twice before players can ruff; consequently, it becomes safe to count kings as tricks most of the time. Trump length presents the biggest problem for bidding; if you have five small trumps, you don't really know how many of them will win tricks. In general, a good rule of thumb is to assume that each trump after the second will win a trick, more if you have high trumps. This is a slight undervaluing of your hand, but if necessary it is much easier to ditch low trumps under high trumps than to be forced to win tricks with trumps you do not have. Obviously, this should be adjusted depending on the number of cards in the hand. At this point, you should also start paying attention to what's been played, as most cards will be in play.

In each case, you should consider how many tricks you will win on a card-by-card basis. Remember, though, that you may end up winning tricks by virtue of other players giving them to you; you should pay attention to bids before yours to see how aggressive the other players will be during the play.

Most of the strategy during the play applies to all types of hands. A good strategy at the beginning of a hand is to play the cards whose value you are uncertain of. Consider the following situation: in a hand of four, you have the ace and four of diamonds, the queen of spades, and the three of hearts, with diamonds as trump, and you have bid two. If you are on lead, you want to lead the queen of spades; if it wins, you do not want to win your four of diamonds, and so you should play the four of diamonds on the next trick (which it will presumably lose.) If it loses, you need to win the four of diamonds in order to make the contract, which will affect your play accordingly.

In general, there are good cards such as the four of diamonds in this example, for which you have a lot of choice over whether or not they will win a trick, and bad cards such as the queen of spades, whose trick-taking value is uncertain. As illustrated above, you want to play your bad cards first so you know what to do with your good cards. In the high-card hands, bad cards are holdings such as Q-10, where you have little choice over how many tricks you will win (with there being significant variance depending on how the other players choose to play the trick); good cards are holdings such as A-K-5-2, which can win zero, one, or two tricks depending on what you need from them to make your bid exactly.

During the play, it helps to be aware of how many tricks the other players have taken, and what their bids are. Let's say you are on the sixth trick of a ten-trick hand, and the first player leads the two of spades; this is the first spade that has been played. The second player follows with the ten, and you are in the third position with Q-4. Will the queen of spades win a trick for you? If the player in fourth position has bid one and already taken one, he is unlikely to play the ace or king if you play the queen. Conversely, if that player has bid three and so far has shown no high cards and taken no tricks, it is likely that they have either the ace or king of spades, and will take the trick (indeed, they may have both.) In this case, if you want to take a spade trick, you should probably play the four and hope to score a trick with the queen on the second spade trick.

Being aware of where the other players stand in relation to their bid also allows you to better interfere with their plans. Suppose there are two tricks remaining, and you have the ace of trumps and the two of another suit, needing one more to make your bid. A trick comes around to you, with the jack of a third suit the highest card played thus far. If the player of the jack needs one trick for their bid, you can prevent them from making their bid by playing the ace of trumps and leading the two (unless their remaining card happens to be either a trump or of the same suit as your two.) If, on the other hand, that player has already made their bid, you should ditch the two, forcing them to take one more trick than they would otherwise want. In general, if you are sure of your bid and just wondering when to take the tricks, you should try to take the tricks that would otherwise be won by a player that wants them.

You should also realize that other players will be trying to do the same to you. This is why bidding zero is difficult in many-card hands; you may have a hand which is unlikely to take any tricks if every player is trying to win every trick, but if you bid zero, the other players will be attempting to force you to win tricks. Perhaps on the fourth trick of clubs, another player leads the three, and you play your last remaining club, the five (having already played the six, eight, and jack.) If the other players were trying to win as many tricks as possible, they would trump here, but if you have bid zero they will gladly discard cards of other suits and give you the trick. Consequently, with holdings such as 8-J, bidding zero is unadvisable as the other players can easily foist a trick off on you.

On each hand, you should also know how many tricks have been bid in all; in particular, whether the total number of contracted tricks is greater than or less than the number of tricks in the hand. If it is larger, players will be more aggressive in trying to take tricks; if it is smaller, they will be less aggressive. In particular, if you are third or fourth to bid and the number of tricks already bid is high, you can bid zero with more impunity, as the other players will doubtlessly focus more on taking their allotment in a time of shortage than on forcing you to take a trick.

6. Napoleon
Napoleon is a card game for five or six players, which in itself is rare. In addition to this, it also has the unique feature of secret (and changing from hand to hand) partnerships, only known by one player. This intrigue in turn leads to a type of strategy which is essentially unique, as players try to guess what the partnerships are, and avoid putting all of their eggs in one basket.
Players: 5-6 players. Rules described below are for six players; see the five-player addendum at the end.
Playing time: 60-90 minutes; this time can be changed by playing to a different ending point.
The Deal: Each player receives eight cards and the remaining four are set aside in the kitty; deal starts at an arbitrary point and rotates to the left. Each player, starting with dealer, either bids a number higher than the previous high or passes. The object of Napoleon is to take face cards in tricks; the number bid is the number of face cards (between 1 and 16) that the bidder contracts to take. (Aces count as face cards in Napoleon.) You must bid higher than the previous bid; you cannot bid equal to it. The bidding ends when all players pass; the highest bidder then becomes Napoleon for the hand. If you pass, you may re-enter the bidding at a later time; passing is not the same as folding in poker or other such games.
Pre-hand Procedures: After the bidding, Napoleon names the trump suit. He or she then looks at the four cards in the kitty, and can exchange up to four cards from his or her hand with cards in the kitty. If Napoleon discards face cards in the kitty, they do not count towards his quota; so if you have bid 16, you will not want to do this. At this point, Napoleon has the option of changing the trump suit; if he does so, his effective bid is increased by one for the play (but not for scoring afterwards.)

Napoleon then chooses one of the fifty-two cards in the deck to be the secretary card. The player who holds this card becomes secretary for the round; this player is on Napoleon's side, and all face cards he collects in tricks count towards Napoleon's total. In addition, the secretary card becomes the highest card in the deck, and is stripped of its suit. The secretary card may be played at any time, even if the secretary has cards remaining in the suit led, and automatically wins the trick. If the secretary card is led, it counts as a trump for purposes of following suit. Note that the identity of the secretary is unknown to both Napoleon and the other players until the secretary card is played. If Napoleon names a card in his own hand as the secretary card, he has no partner (but the card retains all of its value as above.)
The Play: Napoleon makes the opening lead, and play proceeds as an ordinary trick-taking game. If at the end, Napoleon and the secretary together have taken a number of face cards equal to or greater than the number bid, they win; otherwise, the other players win. Extraneous talking is of course allowed, but Napoleon is only allowed to say two things relevant to the game: "Be careful, Secretary," and "Take it, Secretary." The latter comment communicates to the secretary that Napoleon wants him to take the trick, and typically he will play the secretary card. Of course, the secretary may have already played to the trick, in which case the command is useless. As a variant, you can make the command require the secretary to play the secretary card at his or her next turn. The former comment is just a general warning.
Scoring: If Napoleon wins, he gets a number of points equal to the following: 2 if the bid is eleven or below, the bid minus nine if the bid is between 12 and 15, and 10 if the bid is 16. (This is the original bid, and is not increased by 1 if Napoleon changes trump suit after looking at the kitty.) The secretary receives a number of points equal to half Napoleon's total (round down.) If Napoleon names the secretary card to be in his own hand, he gets the secretary's points in addition to his own.

If Napoleon is defeated, each player other than Napoleon and the secretary receives two points. Like Oh Hell, Napoleon scoring systems vary; if you meet someone who knows how to play Napoleon and who did not learn it from this book, they are likely to have a different scoring system.

The game ends when one player amasses a total of 21 points. This number can be changed up or down depending on how long you want to the game to be.
Strategy: In the bidding, it pays to bid aggressively. Having a long trump suit is more important than having exceptionally high cards; long trumps allow you to control the hand. Remeber that as Napoleon, you have lots of things that can go right for you: you may find high cards or trumps in the kitty, defenders may be scared to drop high cards early, or you may end up with a secretary who also has other high cards and can win tricks for you.

As a defender, your task is very difficult, as you do not know who the secretary is. In general, it may very well be the case that your only hope is to discard face cards early in the play, as Napoleon will typically lose his tricks early (as opposed to at the end, when players have probably saved face cards in order to have a chance at winning tricks.) However, if you drop them on the wrong tricks, you may give them to the secretary and doom your efforts; discarding face cards that may win tricks, in particular, is difficult to do, as you're giving the card to someone whose loyalties you are unsure of as opposed to winning it yourself (and you know which side you're on.) Nevertheless, often you must bite the bullet. Especially with face cards which will not win tricks, such as jacks, you may be best served by giving them to another player in a trick, as opposed to possibly having to discard it on a trick of Napoleon's later. One good strategy as a defender is to realize that if a certain player (i.e. one who has already collected numerous face cards) is the secretary, you have no chance at defeating Napoleon, and then proceed under the assumption that that player is not the secretary. Obviously, it behooves you to figure out which player is the secretary as soon as possible; if Napoleon names a high card (as is usually the case), the other players who have been bidding are more likely to be the secretary.

As Napoleon, your first task is to exchange cards with the kitty and then choose a secretary. Generally, it is a good idea to void yourself in as many suits as possible; you should have at least four trumps, including two of the top three. A secondary long suit is better than a balanced hand; after you draw trumps, you can then run the cards in that suit. Don't be afraid to discard singleton jacks and queens, as these will almost certainly end up in the hands of the opponents if you keep them in your hand (and prevent you from trumping that trick.) The secretary card has two purposes; first of all, it gives you an ally, and second of all, it allows you to eliminate a card from the game (remember, the secretary card becomes suitless.) Thus, if you have the king and queen of trumps, and call the ace as the secretary card, your king and queen become the two highest trumps. Generally, you should call the highest missing trump or, if you have no problem with trumps (having the top two and length, or the top three), an ace in a side suit where you have a king (thus making your king high.)

During the play, you want to milk the confusion among the other players as much as possible. Generally, this involves losing your tricks early in the game, where the opponents are suspicious of one another, not knowing which one of them is the secretary. Also, remember that if everyone follows suit, you cannot lose more than four face cards in a given trick; losing these tricks early increases the chance of everyone in fact following suit (and being unable to drop jacks and queens of other suits on the trick.) Save the secretary card as long as possible; invoke it only if you think that the trick being taken contains (or will contain) enough face cards for you to lose your contract, or if the current trick is the last trick you cannot win in your hand. It is far more important for the defenders to know who the secretary is than it is for you to know, and as such this should be concealed as long as possible.

As secretary, you are the only player in the game who must keep a poker face. Try to encourage other players to drop face cards on your tricks early; as defenders will be doing this as well, you should be able to blend in somewhat (don't be too blatant.) In general, you should play the secretary card only when Napoleon instructs you to; he knows what he's doing, and knows where his weakness that needs to be covered is. Of course, if there's a trick with a lot of face cards, you may want to go ahead and take it - especially if you have strength in other suits that Napoleon does not know about (perhaps a side suit ace you want to lead.) A good rule of thumb is that you should reveal the secretary card on your own if and only if you are sure that you and Napoleon will be able to take the rest of the tricks.

6a. Rules for Five Players
Five-player Napoleon is similar to six-player Napoleon. Introduce a joker into the deck, making 53 cards in all. The joker is similar to the secretary card; it has no suit, and can be played at any time, even if its owner is not void in that suit. It is higher than any other card, and thus automatically wins the trick, except that it is outranked by the secretary card (unless it is the secretary card itself.) Deal ten cards to each player, with the remaining three cards in the kitty. Everything else is the same. In the five-player game, expect bids to be lower; Napoleon has a larger percentage of the other players on his side, but this is offset by the fact that he gets to exchange a much smaller portion of his hand with the kitty, severely limiting his ability to void himself of suits and otherwise tailor his hand.
7. Liars' Poker
Liars' Poker is a rather unique game, in that there is no proscription against looking at your opponents' cards. Indeed, this makes the game perfect for playing in cramped spaces, such as car trips or lines. The game itself is very simple, but the element of espionage and coalitions makes Liars' Poker a primarily social game. Like poker, the game involves lots of bluffing; unlike poker, wagering is not involved.
Players: At least two, and in theory unlimited. In practice, the game tops out at around eight or nine. Multiple decks may be desired for games with many players.
Playing time: Can be adjusted based on the number of cards in the hand. A good rule of thumb is that each round will take two minutes; the number of rounds is approximately equal to the total number of cards.
The Deal: On the first deal, each player gets some number of cards; this number can be adjusted depending on how long you want the game to take. We recommend the following: with two players, nine cards; with three players, seven; with four players, six; with five players, five; with six players, four; with seven or more players, three. In each subsequent deal, the player who lost the previous deal gets one fewer card than in the previous round; everyone else gets the same number of cards. Thus, in each successive deal, one fewer total card is in play than in the previous deal.
The Play: Starting with the player who lost the previous deal (in the first deal, the dealer starts), each player going around to the left names a poker hand or challenge the previous person. This poker hand must be one of the following, in ascending order: a single card, a pair, two pairs, three of a kind, a straight, a full house, four of a kind, a straight flush, five of a kind, six of a kind, etc.. Deuces are wild. When you name a hand, you must provide all of the relevant information; for instance, "three queens," "a ten through ace straight," or "the six through ten of hearts." You do not need to name every card; "a seven" rather than "the seven of hearts" suffices. Note that a flush is not a valid Liars' Poker hand. Within each category, ordering is as in poker; for instance, a full house of eights and fives is higher than one of fives and eights, and a six-through-ten straight beats a four-through-eight straight.

The hand ends when someone elects to challenge the previous person rather than name a higher hand. When a challenge occurs, everyone lays down their cards. If, among the cards of all the players at the table, the poker hand named exists, the challenger loses the deal; if the hand does not exist, the player who named it loses the deal. Note that the person on the other side of the challenge gains nothing. The exact poker hand named must exist; for instance, if the hand challenged is "three queens," and there happen to be four kings among the collective cards of the table, this is irrelevant; the challenge fails if and only if there are three queens (including deuces) among the cards in play.

Unlike most games, Liars' Poker specifically encourages cheating. Any trick is acceptable for gaining a glimpse of other players' cards; for instance, you may look over their shoulder or off reflective surfaces. The only thing which is verboten is physically touching their cards.
Scoring: As mentioned before, after each hand the player who has lost gets one fewer card in the next hand. If a player has no cards remaining, he or she is out of the game; as a courtesy, a player who has just been winnowed down to one card is allowed to pick their card out of the fanned (face-down) deck. The goal of the game is to be the last player with cards remaining.
Strategy: Don't be afraid to look at people's cards! The skill in Liars' Poker comes in estimating the probabilities of various scenarios of cards in play based on other people's bids, and the more hard information you have, the better off you'll be. To take an extreme example, if you've managed to peek at everyone else's cards, you know precisely what the cards in play are, and when it comes around to you you know whether a challenge will succeed or not.

If no one looks at anyone else's cards, the amount of information you have is directly proportional to the number of cards in your hand. This means that a player with more cards is less likely to lose a given round, making Liars' Poker a positive-feedback game of sorts. As such, if you and another player have few cards left, it may be quite advantageous for you to combine forces and share your cards with each other; this doesn't affect the procedure of the game (i.e. you still take turns whenever your turn would normally come up), but it gives both of you more information. Coalitions are common, especially when the game is down to three players, at which point the two players who are behind will typically gang up on the player who is ahead. Remember, though, that at the point when your hands are mutually visible, you are potentially quite vulnerable to the sneak attacks of other players' roving eyes. In extreme cases, it may be impossible for you to pick up your cards without anyone else seeing them. Sometimes, it is better to not look at your cards at all, or at least until you have to; low calls will typically not get challenged, since there is little benefit to challenging someone unless you feel that you yourself will get challenged on any higher bid.

It is very important to take other players' bids into consideration. If a player bids three queens, he likely has queens; if you are staring at a queen and a deuce, four queens or even five is a sound bid. (Remember that, at least in theory, the other player made their bid ignorant of your hand.) Conversely, though, it is often worthwhile to bid something you don't have at low levels where you're relatively certain you won't get challenged, in order to trick others. At the beginning of the game, about half the deck will be in play; with the four deuces, a bid of three of a kind will typically not get challenged, so use these bids to your advantage.

It is very difficult to be certain that the player before you doesn't have their bid, and because of the system it tends to be unprofitable to challenge players unless either the odds are extremely against them or you think you will get challenged and lose on any higher bid. In particular, if you bid something that the player in front of you has bid (typically bids go around the circle about twice, so they are likely to have bid something), they will be stuck between a rock and a hard place. In this vein, the player in front of you is the most important one as far as looking at their cards goes; if you can look at their cards without them noticing, in particular, when you bid something they have (especially if they haven't bid it yet), they are likely to think that you have some and not challenge you. Remember, you can't lose if you aren't challenged and don't challenge anyone.

Liars' Poker is essentially a game of psychology. Try to figure out the tendencies of the other players; if someone seems to never have what he bids (everyone's cards are revealed at the end of every hand, so you can actually track this), don't take his bids into consideration. If someone is reluctant to challenge, try to sit behind them. Make deals with people; if the guy in front of you wants to bid something, let him. Try to gang up on the player in front, since his position will only get better as time goes on. Show your cards to other players when it's advantageous to you; if someone is hesitating about challenging, and you don't have any of the cards in question, you can often convince them to challenge by showing them your useless cards, especially if you once claimed to have something useful for the current call. (A better strategy is to get them to commit to challenging if you don't have any before showing them your cards.) Similarly, you can often avoid being challenged (the guy in front of you doesn't want to challenge you since he might lose, but he may not know what to bid) by showing your hand to the person in a position to challenge you; this is a win-win situation, since it gives him more information as to what to bid. In a game with lots of players, team up with people early, and try to bid high enough that the bidding will be challenged before it gets around to you (but not so high that you'll be challenged.) Bid confidently, as if you know what you're doing, even when you don't. Last but not least, don't get suckered in by other people's psychological ploys, like the ones mentioned here.

8. Psychological Jujitsu
While Liars' Poker is mostly a game of psychology, Psychological Jujitsu is entirely a game of psychology. There is very little luck involved in the game, and it's not a game where excessive thought or speed will help you. Simply put, in order to win, you must figure out what your opponent is thinking.
Players: At least two; best with two or three. If more than three are playing, Psychological Jujitsu requires an additional deck or additional decks of cards.
Playing time: 5-10 minutes per hand.
The Deal: One suit of cards is set aside, shuffled, and placed face down on the table. Each player receives a full set of cards, comprising A, 2, 3, ..., K; these represent the numbers 1 through 13 in the obvious way. The game can in fact be played without a deck of cards, simply giving each player tokens labeled 1 through 13 (or any other number), and having an extra set of tokens bearing these values.
The Play: Psychological Jujitsu consists of thirteen rounds. On each round, the top card of the extra suit is turned up and auctioned off. Players choose secretly which of the cards in their hand they are playing, and simultaneously reveal their choices. These are bids for the current card; the highest-ranked card played wins the auctioned item, which is worth a number of points equal to its rank (ace being 1, jack 11, queen 12, king 13, and all other cards their numeric value.) If multiple players tie, the card is thrown out. In any case, all cards played to that round are discarded. Each player starts with thirteen cards, and there are thirteen rounds, so each player will submit each bid exactly once over the course of a hand. When a player wins a card, that card is placed face-up in front of the player. In the case of a tie on a given round, the way we have presented is only one variant. Another variant is to throw out all tied cards and give the card to the highest card that is not tied; another variant is to have all players take back the cards they have played and put the card up for auction on the bottom of the deck.
Scoring: At the end of a hand, each player's score is the total of the ranks of the cards they have won. There are 91 points in all. The game can either be played by the hand or cumulatively; in the latter case, players add their scores from successive hands until a player reaches a predetermined limit (200, for instance.) In this case, one can play to either end the game immediately once a player reaches 200, or finish the game at the conclusion of any hand which puts a player over 200.
Strategy: In the two-player game, the best strategy is to always play one rank higher than your opponent, except that when he plays 13, you play 1. This will allow you to win 12 of the 13 cards, and even if your opponent takes the king, you will win by the potent margin of 78 to 13.

Unfortunately, employing this strategy requires figuring out what your opponent is going to play. Clearly it makes more sense to play high cards on the cards that are worth more, but there is no specific strategy that can work, since your opponent can always play the above strategy to dominate you. Playing a card one higher than the card up for bids is good, since the natural correlation is to play a card about equal to the one up for bids, but by the same token playing a card two higher is also good, since your opponent is likely to play a card one higher. And so forth. There's obviously a breaking point to this argument, since by playing a card two higher you are essentially conceding the two highest cards, and this will eventually result in a loss.

Don't be afraid to play very low cards when very high ones are up for bid. You'll lose, but your opponent will have expended a much higher card in the process, leaving you with a significant advantage for the remainder of the game. In general, you want to try to lose by a lot when you lose and win by only a little when you win. With more players, the strategy is murkier; assuming you play the main variant described here, it is a reasonable strategy to play your highest cards on mid-range cards (8, 9, 10) and hope that the other players all play their highest cards on higher cards, cancelling each other out. But in general the optimal strategy is obviously highly dependent on what your opponents are doing.

Finally, keep track of what your opponents have played; it's important when you're trying to figure out their decision process.

9. Canadian Fish
Canadian Fish, like Creights, can loosely be called a variant of a simple children's card game. In this case the game in question is Go Fish, and, like Creights, the variant is far more complicated than the original. However, unlike Creights, the rules of the game are very simple; the complexity comes in the actual play. This provides an excellent combination of a game that is easily explained but complicated enough to provide a superb form of mental exercise. Unlike most of the games here, memory is an extremely important part of Canadian Fish, and not just memory for what cards have been played, as is usually the case in card games.
Players: Exactly six. I suppose in theory the game could be played with a different number, but it wouldn't work nearly as well.
Playing time: 15-30 minutes per hand. The length can vary wildly from hand to hand, and also depends heavily on how analytical the people you're playing with are.
The Deal: Remove the eights from the deck, leaving 48 cards. This splits the deck into eight suits -- the two through seven of each suit, and the nine through ace of each suit. Whenever we say the word suit in this section, we refer to one of these eight six-card suits. Deal eight cards to each of the six players.
The Play: In a hand of Canadian Fish, there are two teams of three; the teams alternate as you go around the circle, so your teammates are the players two to the left and two to the right of you. You may wish to assign teams randomly or mix things up after each hand. By convention, the eldest Canadian at the table goes first. If there are no Canadians, decide who goes first however you want, such as picking the person who was in Canada most recently, who has been to Canada the most times, etc.. On your turn, you do the following.
  1. Ask any opponent for a specific card. You cannot ask for a card that is in your hand, and you cannot ask for a card unless you have a card in that suit. For example, to ask for the seven of clubs, you must have the two, three, four, five, or six of clubs (you can of course have multiple low clubs), and you cannot have the seven.
  2. If the player asked has that card, they must give it to you. If they do not, it becomes their turn.
At any time, any player, regardless of whose turn it is, may declare a suit by saying "declare." Once a player does this, silence reigns until the declaration is finished. The player must now assign all six cards of a suit among the three members of that team; if I am on a team with Bob and Rachel, I could, for instance, say, "Declare: I have the two and five of clubs, Bob has the three, four, and seven, and Rachel has the six." Regardless of correctness, all of the cards of that suit are removed from play; if I am correct, our team "wins" the suit, and if I am incorrect in any way the other team "wins" the suit. You can only declare cards to be in your teammates' hands; if I have five of the low clubs and I know which of my opponents has the last one, I cannot declare this. Similarly, if my opponents have all six of the cards, I cannot scoop them by declaring the suit. You are allowed to declare suits even if you don't have any of them, though this is rare. The declaration, whether correct or incorrect, does not affect whose turn it is; this remains the same, and play continues as usual. The only exception occurs when the declaration removes all remaining cards from the hand of the player whose turn it is; in that case, that player must (with no input from anyone) transfer control to one of their teammates with remaining cards. Players with no cards are out of the game, and are not allowed to declare suits. If at any point one team has all of the remaining cards, nothing further can be said. They must declare all of the remaining suits; anyone can declare these suits, but the first person to say "declare" must declare the suit as usual. The game ends when all eight suits have been declared. If at any point you screw up by asking for a card you have, or a card in a suit you don't have, the suit containing that card is immediately removed from play and won by the other team. Conventions are explicitly forbidden in Canadian Fish. You are of course allowed to make inferences, but you cannot, for instance, agree beforehand that if you ask an opponent for the jack of a suit, you have the queen or king, either in secret or openly.
Scoring: One point for each suit that your team has won. The team with more suits wins; score is typically not kept, but if it is, each player gets a number of points equal to the number of suits their team has won. The game should consist of a fixed number of deals; six is a good number. Teams can either be kept the same throughout or mixed up. In any case, the goal is simply to win as many suits as possible.
Strategy: Obviously, if one could remember everything that has been asked, one would be doing well. In practice, most humans cannot remember this much, so you must decide what to remember. Since you can never acquire cards in a suit that you don't have, it is obviously best to focus your attention on the suits you do have. It is usually better to retain perfect information about a few suits than imperfect information about all of the suits, though not always.

Something that is extremely important is to keep control away from dangerous opponents. Your team can conspire to blackball one or two of the opponents so that it is never their turn, and this is often a good strategy. For example, suppose that everyone knows (because they've seen the cards being exchanged) that one of your opponents has the two through six of clubs. He has asked both you and your right-hand teammate about the seven, and neither of you have it. If one of your opponents had the seven, they would declare the suit now, since they know the location of the other low clubs; therefore, your left-hand teammate must have the seven. If the opponent with the low clubs gets a turn, he will take the seven from your left-hand teammate and they will win the suit. If, however, you never ask him a question, it will never be his turn, and he cannot take the suit.

Therefore, you should never ask him a question. It is not permitted to say things like this explicitly; you can't tell your teammates not to ask him, but hopefully they've been paying enough attention to realize what's going on. Of course, your opponents, if they've been paying attention, should know not to ask your teammate with the seven any questions. This is fairly frequent, and is one reason why you shouldn't totally ignore suits even if you don't have any cards in them. Each game is different, and much of the skill in the game lies in figuring out what information is important, since you usually can't remember everything.

If someone asks you about a card in a suit, and you have a different card or multiple cards in that suit, it is often a good idea to ask them about a card in that suit, although this does run the risk of giving away information (namely that you have another card in that suit.) Sometimes the best strategy if you have multiple cards in a suit is to lie low while other people ask about it; you end up with a lot of information, sometimes enough to place the entire suit. For purposes of memory load, you should generally try to clean out suits sequentially, so you have less to remember (since once a suit is out of play you can forget everything you know about it.

If you have all six cards of a suit, or if you know where a suit is, it can on occasion be useful to not declare it. By saving the suit, you increase the chance that other people will ask you questions since you appear to have more useful cards than you actually do; you can also declare the suit at an advantageous moment when it's your turn in order to transfer control to an appropriate teammate. If you have all the cards, there is no risk in holding onto them, since you're going to get the suit eventually. If you don't have all the cards yourself, the risk in holding onto them is that your teammates may ask unnecessary questions about that suit, so it is usually best to declare these as soon as possible. Of course, if everyone knows you have all six cards in a suit, there is no benefit to holding onto them.

If your teammate asks about a suit that you have, be aware! You know that they also have a card in that suit, and so by teaming up often you can tease all the cards from the opponents. In particular, if you don't have the card that they asked about, you can do well to ask for it yourself, since you have a better chance of being correct.

In general, you shouldn't declare a suit unless you are absolutely certain. If your team has all six cards of a suit, they're not going anywhere; the only way your opponents can win the suit is if you declare it wrong. The benefit of declaring earlier is that your teammates won't waste questions asking about the suit, but this is usually small compared to the potential loss of the suit to your opponents. Asking a question is not as big an advantage as you might think, because your opponents get just as much information from a failed question as your teammates do. Note that if you ask a question about a suit that your team has all six cards of, your opponents in fact get no useful information.

There are a couple of exceptions to this rule; the most obvious is in the end-game situation, which occurs surprisingly frequently, where only one team has cards remaining and they need to declare a suit or suits. In this case, you're not getting any more information, so the person with the most information (usually, but not always, the person with the most cards in the suit) should declare the suit. Often this comes down to a 50-50 guess.

A more subtle reason why you might want to declare a suit that you're not sure of is for transferring control. Let's say there are two suits left, the low clubs and the high diamonds. You have the four, five, and six of clubs, and you know that your left-hand teammate has the three. You know that one of your teammates has the two and the other one has the seven, but you don't know which. Your right-hand teammate also has the king of diamonds; your opponent has the other five diamonds, and has recently asked both you and your left-hand opponent for it, so everyone knows where the last diamond is.

In this case, you should guess the clubs. Maybe you could improve the club knowledge if you asked the opponent for the three (and, incidentally, this is in general good strategy when you have all six cards of a suit; you know the questions will fail, but it's the only way to convey information about the suit), allowing one of your teammates to declare. However, if you did this, you would lose the diamonds for certain. A better strategy is to guess the clubs, which you will get right 50% of the time; regardless of success, though, you are out of cards and you can transfer control to your right-hand teammate, who will acquire all the diamonds from the remaining opponent. This gives you the diamonds always and the clubs 50% of the time, which is preferable to the other scenario, where you always lose the diamonds and get the clubs some percentage of the time. (It would have to be 150% for this to be the preferred choice.)

In general, stay alert and you will do well.

10. Spielen
Spielen is a fast-paced game which is played almost entirely outside the box. It is unique among all the games here for two reasons: first, it is the only one of the ten games which is played in real time (as opposed to turn-based), and secondly, it is the only game where you plan strategy with a teammate beforehand. (Along with Canadian Fish, it is one of only two team games.) Spielen breaks down into two phases; the phase before the game is devious planning, while the game itself is frenetic.
Players: Players play in teams of two. Two teams of two (four players) is ideal; the game can certainly support three teams, and perhaps four, although that's stretching things a bit.
Playing time: One hand takes around two minutes. The game consists of a number of hands, usually around ten; before a game, there is a planning session which can take as long as you want. Five to ten minutes is probably about right.
The Planning Session: During the planning session, the two members on each team get together in secret to make up codes for the game. See the later sections for what the point of these codes are, as well as some sample codes.
The Deal: Deal four cards to each player.
The Play and Scoring: The play goes as follows: the dealer turns up, simultaneously, four cards from the top of the deck. Players now simultaneously may opt to exchange cards from their hand for an equal number of cards from the board, simply throwing them into the middle (face up) and taking other cards. You may never at any point have more than four cards in your hand, although you may have less than four cards for an instant or even for a significant amount of time. (There is little benefit to doing so, although it might certainly confuse people.) When no one wants any of the cards on the board, the dealer sweeps them into a discard pile and draws four new cards from the top of the deck. If the deck runs out of cards, the discard pile is reshuffled.

The goal of the game is to get four of a kind. Well, not really; the goal of the game is to get four of a kind and convey this information to your partner without letting the opponents catch on. This is where the codes come in. During a game of Spielen, people should be talking. Generally this is just harmless banter, but you're allowed to talk about the game (unlike most games, where table talk is discouraged.) The only things that aren't allowed are "under-the-table" things; you can't kick your partner under the table, or say things in a foreign language that only they understand, or anything like that. All communication must be observable by everyone, although nonverbal communication such as placing your hands on the table, rearranging your cards, and winking is all fair game.

The game usually ends when one person has four of a kind and their partner knows it (because it's been communicated somehow.) At this point, the partner calls "Games" (English for "Spielen.") If their partner does in fact have four of a kind, that team gets a point (and the round ends); if they don't, the other team gets a point.

However, if you think that one of the opponents has four of a kind, you can scoop them by calling "Anti-Games." If you are correct -- if either opponent has four of a kind -- your team gets two points, while if you're wrong the other team gets two points. So the goal is not just to tell your partner that you have four of a kind, but to do so without letting the other people crack the code. You must set all codes beforehand; you cannot confer between deals if your opponents have cracked your code.

But wait, there's more. If you maneuver it so that both you and your partner have four of a kind, either of you may call "Double Games," which is worth two points; again, your opponents get the same number of points for an incorrect call. If you think both of your opponents have four of a kind, you may call "Anti-Double Games," which is worth four points for you if you're right and four for your opponents if you're wrong. This is, of course, exceedingly daring.

In many cases, especially after the code has been cracked, two people may declare at the same time. Whichever team started speaking first has the declaration that is evaluated; if it's simultaneous, then no one gets any points and the round is declared over. The first team to accumulate 6 points wins (for a longer game, play to 11.)

Strategy: Almost all of the strategy lies in concocting codes, although if you're clever you can keep track of cards. In practice, because the game is so chaotic, this is pretty difficult; many players may be exchanging various numbers of cards at once, and certain cards may be picked up off the table before you even really have a chance to see them. Another non-code method is simply to observe whether people are exchanging cards; obviously, if someone has four of a kind they're unlikely to be exchanging any cards, although proceeding from this to a sure inference is tricky.

Codes come in many forms. The more complicated a code, the harder it is for opponents to crack, but as the code complexity grows so too does the difficulty of remembering it. Similarly, you want codes that can innocuously be slipped into everyday life, but the more natural they are the greater the chance that you will slip up and give the code by accident. For example, I once had a code with a partner consisting of all seven-letter words; we must have given the code a million times by accident. In general, things that require a lot of parsing time are bad codes due to the pace of the game. You're not allowed to stop and think during the play; if you're the dealer, you're expected to swipe the cards within a couple of seconds after people have stopped exchanging, and if you're not the dealer, the cards will be swiped quickly so you don't really have time to stop.

Word codes are generally decent: for example, if you say the name of a state. This is, of course, easy to crack, so you might change it to saying the name of a state beginning with a consonant, or one west of the Mississippi, or something. A decent meta-strategy is to wait for some amount of time after the code is given before saying Games; the downside of this is that the other team might say Games in the meantime.

Tapping your fingers on the table is reasonable, but the problem with this and other physical cues is that your partner might not be paying attention to you, and so you may have to give the code repeatedly before they pick up on it. The opponents are likely to pick up on it in the meantime; after all, there are twice as many of them, so if everyone knows the code the chance of your teammate getting it first is only 1/3. And also, it's worth half as much to call Games as it is to call Anti-Games, so having a code that's hard to crack is pretty important, even if your partner doesn't pick up on it immediately.

You should, of course, have several codes, ideally in different media (in other words, a couple of verbal codes and a couple of physical codes.) But the more codes you have, the harder they are to remember; four is probably a reasonable number. Since you can't change the codes during the game, you'll want multiple codes in case some of them get cracked, as well as in case your partner is oblivious to one of them. The game is fast enough that you'll forget things from time to time.

(c) Mike Develin, 2002. Thanks to everyone who taught me these games and everyone who I've played them with, especially Joon Pahk, whose writeups of Creights and Barbu were the basis for the ones appearing here.

Back to Mike's homepage