Challenges of Innovation is a multi-part essay on the current state of poker from the perspective of a commercial game designer. It seeks to answer the questions of whether No-Limit Hold’em is destined to remain the predominant variant, why it’s so hard to come up with good new variants, and what avenues of innovation appear the most promising. To read from the beginning, click here.

Our final example variant, and probably the final instalment of this series, is a game I call Snake. Although none of its individual mechanics are as exotic as the ones we saw in Bullseye, last week, they are combined to produce a game that plays rather differently than any existing variant.

Its closest relative Badugi, in that it is a four-card triple draw game where hand strength is primarily a function of how many cards the player can legally use together. The resemblance ends there, however, as the addition of community cards radically changes the game’s information structure, and the rules for forming and comparing hands are likewise entirely different.


  • The game is played with blinds and a dealer button. Order of play is the same as for other flop and draw games, for both the betting and drawing rounds.
  • Players are dealt four cards face down to begin with.
  • There are four betting rounds. Any betting structure – Limit, Pot-Limit or No-Limit – is viable. If played as a Limit game, big bets kick in for the third and fourth betting rounds, as usual.
  • After each of the first three betting rounds, one community card is dealt face up.
  • After each community card is revealed, there is a drawing round in which each player may choose to discard any number of cards and be dealt replacements, or stand pat.
  • After the fourth and final betting round, play proceeds immediately to showdown, with no additional community card or draw.
  • At showdown, each player forms the largest and best hand possible using his or her four personal cards and the three community cards. Hands can thus be up to seven cards in size, if all personal and community cards can be used.
  • A legal hand consists of cards of any number of consecutive ranks, with multiple cards of a single rank permitted. In other words, 4-4-5-5-5 and 7-8-9-T-J are both legal five-card hands, but 3-3-6-6-7 is not, as 3 and 6 are not consecutive without a 4 and a 5 as well.
  • Aces can be consecutive to either a 2 or a K, but not both at once (i.e. no “round-the-corner” straights). A-2-3 and Q-K-A are both legal, but not K-A-2.
  • If one player can form a larger hand (consisting of a greater number of cards) than all others, that hand wins.
  • Between hands of equal size, the hand containing the largest set of cards of the same rank wins: quads beats trips, which beats a pair, which beats a hand of all singletons.
  • If the largest sets are of identical size, the second largest is compared: Q-Q-Q-J-J beats A-K-K-K-Q, for instance, because it has a pair to go with the trips. If the hands are tied for number and size of sets, then the rank of the largest sets are compared, followed by the next sets down and, if necessary, finally the singleton kickers from highest to lowest. This is done in exactly the same manner as in conventional highball poker variants, so Q-J-J-T-T beats J-J-T-T-9 due to its kicker, and so forth.

Hand ranking list

The showdown rules for determining the winner between hands of equal size may sound confusing, but in fact they are almost identical to the standard poker hand rankings, but with straights and flushes removed and the scheme extended to handle hands of up to seven cards. Here is the full ranking scheme, along with proposed names for the new hands:

  • 4+3 (“Carrier”) – e.g. K-K-K-K-Q-Q-Q
  • 4+2 (“Battleship”) – e.g. K-K-Q-Q-Q-Q-J
  • 4 (“Quads”) – e.g. Q-J-J-J-J-T-9
  • 3+3 (“Six-pack”) – e.g. Q-Q-Q-J-T-T-T
  • 3+2 (“Boat”) – e.g. K-K-Q-J-T-T-T
  • 3 (“Trips”) – e.g. K-Q-Q-Q-J-T-9
  • 2+2+2 (“Three Pair”) – e.g. K-K-Q-J-J-T-T
  • 2+2 (“Two Pair”) – e.g. K-K-Q-J-T-9-9
  • 2 (“Pair”) – e.g. K-Q-J-T-T-9-8
  • Singletons – e.g. K-Q-J-T-9-8-7

It’s important to remember, however, that hand size is the most important factor and these hand rankings only serve to break ties, just as in Badugi. Thus, even K-K-K-K-Q-Q loses to K-Q-J-T-9-8-7 because a seven-card hand always beats a six-card hand.


Obviously, good starting hands are ones which contain some combination of connected or paired cards. The actual ranks are of secondary importance, but Jacks and Tens strike a fairly ideal balance between having kicker value and being able to extend in both directions, whereas Kings and Aces will tend to create hands that are drawing a little bit more thinly on average. Starting hands can be “dense” or “wide” depending on how tightly packed the cards are. J-J-T-T is a dense hand, for instance, while 9-7-5-3 is a wide hand. A wider hand is more likely to connect with the community cards, but weakly, whereas a dense hand is more speculative. Any card from 2 through Ten gives 9-7-5-3 a reasonable first draw, whereas J-J-T-T will either have a pat monster on a 9 through Queen, or else miss entirely.

The addition of community cards means that considerably more information can be gleaned from the number of cards each player draws. In particular, it creates many situations in which a player may want to draw more cards on a later street than an earlier one. For instance, a player holding T-9-8-7 and catching a T for the first community card will want to stand pat. If the second community card is a Queen, however, that player will likely want to discard the 7 or perhaps both the 8 and 7 in the hopes of getting a Jack to connect with the Queen. This pattern of standing pat and then drawing is a giveaway that the second community card was “on the wrong side” for that player (and therefore that they don’t have a Jack), while a player who drew one card in both the first and second rounds is actually much more likely to have a strong hand.

For this same reason, players should consider occasionally being deceptive with their draws. For instance, the player holding T-9-8-7 in the above example may opt to stand pat again and hope that the third community card is a Jack or, perhaps, something like a 6. If it is not, then they can draw in the third round while giving away slightly less information, but of course, the downside to this plan is giving up one opportunity to draw.

Many interesting situations come up in which one player has a powerful but hard-to-extend (i.e. “dense”) five- or six-card hand, while the another has few pairs, but much better chances to make a full seven-card hand on the final community card and/or draw. This, in turn, creates interesting scenarios on the final street, where the holder of the smaller made hand needs to try to guess whether his opponent has a full seven or not. For instance, let’s say a player holds J-J-T-T on a T-9 board, a 5 comes as the final community card, and both players stand pat. He has the near-nuts for 6 cards (only losing to Q-Q-Q-J or 9-9-9-x), but now has to figure out whether or not his opponent is likely to hold something like 6-7-8-9, having made a seven-card hand with the 5.

Position is extremely important, perhaps even more so than in other draw games due to the higher amount of information. There are often spots where it’s unclear how many cards one should draw, or even whether to draw. Seeing how many cards others have drawn and thus getting a sense of their strength makes those decisions easier.

A player who completely misses the first community card has an interesting choice between drawing a large number of cards (possibly even all four), or standing pat and hoping that the second community card hits them; assuming the player’s hand is sufficiently removed from the first community card, then if the second card hits them, it will likely miss everyone who connected with the first. This is another reason that position can be so important; if the player is last to act when put in this position, standing pat becomes a more attractive option. It’s likely that doing so will intimidate the other players into checking, whereupon the player can check as well and see the second community card for free.


  • The information provided by the community cards and the draw is greater than the sum of its parts, as mentioned above. The depth of the hand reading skills involved in Texas Hold’em is one of the reasons for that game’s popularity, and one of the things that is lacking from existing draw games. Hand reading is definitely both more difficult and more important in Snake than in any standard draw variant, and I suspect that it could turn out to rival or exceed Hold’em’s in some situations.
  • Some starting hands are clearly better than others, but no hand is ever way ahead until the community cards have come out. A hand like pat quads looks excellent, but needs a perfect runout to stay ahead, and even something like 6-7-8-9 will sometimes miss the board. The most uncoordinated hands, such as 3-7-T-K are also the most likely to connect in some small way, which means that the more gambling-oriented players will find it hard to fold in the first betting round, making it a game that’s likely to generate a lot of action.
  • Drawing decisions are often a lot more complicated than in other Triple Draw games, where at most the decision is usually between keeping a mediocre pat hand or drawing one card. In Snake, there can often be several options worth considering, particularly on the first and second community cards.
  • The showdown mechanics produce a nice distribution of hand strengths. Players with a reasonable hand after the first community card will usually have something between a very strong 5-card hand and a weak 7-card hand by showdown, leading to a lot of marginal situations. At the same time, strong 7-card hands are entirely possible, although the very best (7-card quads and up) are extremely hard to make, as they should be. The drawing mechanic makes exact odds calculations impossible, but a Carrier (7-card quads plus trips) is probably at least as hard to make as a Royal Flush in Hold’em, and thus an event similarly worthy of celebration.


  • Although the hand ranking scheme is actually very logical, simply being an extension of the way pairs, full houses and so forth are compared in other poker games, it gives the impression of being complex because there’s no way to explain it which is both concise and complete. After a few hands, it should seem intuitive to most players, but convincing casual players to try it in the first place might be challenging.
  • The hand-size-first rule may upset some players who get excited when they are dealt what they believe to be an extremely powerful starting hand like pat trips and quads. These hands bear a cosmetic resemblance pocket pairs or rolled-up trips in Hold’em and Stud respectively, but in fact are more like speculative hands in that they need good community cards in order to have showdown value. In terms of game balance, this is just fine, but players may feel cheated the first few times their “monster” fails to improve and leaves them with a pitiful 4-card hand at showdown.
  • Although monster showdown hands are possible, there’s no single “best” starting hand, which is likely to be a downside for casual players. This is again a problem with perception, rather than with reality, but it’s clear that casual players love getting dealt their “pocket rockets” in Hold’em and equivalent hands in other games. In Snake, there are a great many excellent starting hands with considerably better equity than the average, but the differences between them are situational. While this adds strategic depth, it will likely create more limping and less raising in the first betting round, and feel both confusing and unexciting for first-time players.
  • The fact that the community cards can lead a formerly pat player to start drawing again means that there are likely to be more cards drawn overall than in Badugi or Lowball Triple Draw. This, in turn, means that the deck is more likely to run out when playing at a large table. Discards can always be reshuffled when needed as in other draw games, but this should be a rare occurrence; although it feels like the game should work well with up to eight, I suspect that in practice, a loose table of that size might make reshuffling annoyingly common. It may turn out that the game is best played 6-Max or smaller, which would limit its appeal somewhat. Of course, since suits have no relevance, there’s no reason one couldn’t create a two-deck variation for large tables, although this would change the drawing odds considerably and add a number of additional possibilities for extremely rare and powerful hands, all the way up to seven-of-a-kind.

Alex Weldon (@benefactumgames) is a freelance writer, game designer and semipro poker player from Montreal, Quebec, Canada.