Challenges of Innovation Pt. 3 – Soft Requirements

Alex Weldon : April 7th, 2015


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.

Last week we talked about the hard limits on the design space for a poker variant – that is, the characteristics that a game would have to have to be recognizable as a form of poker, rather than a different game entirely.

In a nutshell, these were: the use of a standard deck of cards; the mixture of public and private information about hand strength; betting rounds using the standard set of actions; and the goal of winning the pot either by making the others fold or having the best hand at showdown.

These requirements leave considerable room for experimentation, with the most obvious variables being:

  1. What cards players may use to form their hand, and how and when these cards are distributed,
  2. What information players have access to (including which cards they can see), and how and when they get it,
  3. The number of betting rounds,
  4. Possible actions for players between betting rounds (e.g. drawing), and
  5. The rules for determining the winning hand(s) at showdown

Having five possible axes along which to modify the game would seem to suggest a huge number of possibilities, and indeed, no list of variants covers the full breadth of what’s been tried in casual dealer’s choice home games over the years. Despite this, the list of commonly-played variants in a casino setting is considerably shorter; most mixed games use three, five or eight variants. The WSOP Dealer’s Choice events this year will set the record for most variations used in a single tournament at 19. There are a few cash games being spread in Las Vegas which incorporate more variations than that, but not many more.

The reason is that, to attain even limited adoption in a serious context, a poker variant needs to meet several other “soft” requirements, some of which are considerably more restrictive to our options than that hard ones. On top of being recognizable as poker, a successful variant needs to be accessible, exciting and feature the right balance of luck and skill to appeal to both recreational and professional players.

Table size requirements

Poker is a social game, and although heads-up play is popular among high-level players, recreational players almost universally prefer to play in a larger group. Since recreational players form the foundation of the poker ecosystem, a viable poker variant needs to be able to handle at least five or six players, and preferably eight to ten, since this is what people have become accustomed with the rise of Texas Hold’em.

This imposes two main constraints on variant design: First, we must ensure that the deck does not run out before the end of the hand – at least, not often. Stud and Draw games both contain provisions for dealing with an exhausted deck, but they almost never arise in actual play and are thus more a point of trivia than a design flaw. This should be the case for any new variant we come up with as well: The 52-card deck should be sufficient to deal the entire hand without reshuffling or other special-case rules, except in circumstances so rare that the average player will never encounter them.

Secondly, and more importantly, the game must be resistant to collusion. It’s unfortunately the nature of games that any game with more than two players is vulnerable to collusion to some extent (unless there is no strategy or player interaction at all). So we can’t require that a poker variant be completely collusion-proof, but the benefit must be small enough and the probability of detection high enough that it remains a rare and manageable problem, in line with existing variants.

Measuring the potential for collusion in a given set of rules is a difficult problem, but at the very least, the necessity to avoid it rules out certain otherwise tempting avenues of experimentation. These include giving players decisions which impact others’ hands or the community cards (the passing of cards between players rather than discarding and drawing from the deck for instance), or allowing players private information about anything other than their own cards.

Although handling many players is the harder requirement to meet, it should also be noted that the game must be playable by as few as two players, since players come and go in cash games, and tournaments eventually play down to a single winner. Any mechanics which require a specific or minimum number of players to function are likewise ruled out for this reason.

Information and hand strength

The flow of information in the game is likewise important. At a minimum, a successful variation requires that a player’s private information at the beginning of the hand is sufficient to affect their decision-making (in other words, that it is not outweighed by position and/or other public information) and that additional information gained during the hand is useful, but in proportion to the information inferred from others’ actions. In other words, it’s important that the strategy leaves room both for playing one’s own cards and for reading one’s opponents.

One important implication of these requirements regards the strength of starting hands. The better starting hands need to have a meaningful initial advantage over the worse ones in order to make the early rounds of betting relevant, but not so big an advantage that events later in the hand are unlikely to change anything. A massive initial advantage can sometimes be acceptable, but only for very rare hands, such as pat straights or better in 5-Card Draw, or rolled-up trips in Stud. For matchups between more commonly-seen starting hands, the favorite should usually be between about 55% and 80% to win in order for initial information to be balanced against that acquired later.

Later on, hand strength should likewise follow a sensible distribution, such that holding an unbeatable hand prior to the final betting round should be rare. At the same time, a wide spectrum of medium-strength hands should be possible, since the most interesting scenarios in poker are the ones in which neither player can be sure of whether he’s ahead or behind.

Getting the right distribution of strong and weak hands tends to constrain the number of cards you can allow the players access to for a showdown. For standard hard rankings, seven seems to be the sweet spot: most Stud games have seven cards per player, as does Hold’em. Even in Five-Card Draw, most hands tend to draw between one and three cards, meaning that a player ends up having seen an average of around seven. Omaha is of course the exception which proves the rule and is worth studying for this reason – in particular, its special “two plus three” rule for forming hands at showdown and why that’s essential to the game’s functioning.

Betting round requirements

There are also practical constraints on the number of betting rounds. Two is the minimum if we want players to obtain additional information mid-hand, but the relative lack of popularity of Five-Card Draw and 2-7 Single Draw shows that it is probably not enough to make a really good variant. It’s generally the case that a deeper game tree will lead to more options and thus more strategic complexity. Indeed, most of the more advanced strategies in a game like Hold’em – slowplaying, pot control, multiple-barrel bluffs and floating – all rely on the existence of multiple streets.

On the other hand, too many streets of betting leads to problems in the form of pot bloat. This can be an issue in one of two ways, depending on the game’s betting structure. In a Limit game, a large pot tends to trivialize later decisions. Faced with a bet which is small in relation to the pot, a player should almost always call; even in existing games like Stud and Limit Hold’em, folding to a single bet in the final betting round is usually a mistake, and adding additional betting rounds would only make the problem worse.

If Limit games tend to become dry and anti-climactic in the later betting rounds, Pot-Limit and No-Limit games have the opposite problem. In these games, the size of the pot tends to grow exponentially: a half-pot bet will double the size of the pot if called, and a full pot bet will triple it. In most of these games, it’s normal for pots to grow to be around 100 times the value of the blinds and antes by showdown. Adding additional streets would balloon these pots even further. Although the drama this entails would seem to be a desirable feature, it makes small pots won on early streets seem trivial in comparison; in the worst case, at a tight table, it can mean that the vast majority of hands are essentially meaningless, and the overall outcome of the game is decided by a very small number of very large pots.

Here too, looking at the relative popularity of games, we find what appears to be a sweet spot at four (flop games, triple draw games) or five betting rounds (stud games). Games with two betting rounds exist in the form of Five-Card Draw and 2-7 Single Draw Lowball, but are less popular. Oddly, no commonly-played variants feature three betting rounds.

Hold’em and the ever-narrowing design space

Taking all this into account, we start to see why Hold’em is a tough game to beat. It combines the empirically most-popular number of cards – seven – with the likewise tried-and-tested four rounds of betting. Preflop hand odds tend to run anywhere between even money (for the classic pair-vs.-overcards coin flip) up to around 80% for most common match-ups, and cap out at just under 95%. Postflop, there are also a lot of common situations which give anywhere between 10% and 40% odds for the underdog, meaning there is usually one player who is significantly but not overwhelmingly in the lead. All hand types are seen at least occasionally, and it’s not uncommon to see anything from high card to a full house win a showdown. Finally, by the river, players have knowledge of 5/7ths of their opponents’ cards, while the hidden information is still the same as it was at the beginning of the hand, maximizing the amount of information players can infer from their opponents’ bets.

In other words, Hold’em does in fact seem to be pretty close to the ideal game which can be made through mixing and matching of existing poker mechanics. Although mashups like the recent Badugi hybrids provide low-hanging fruit for short-term novelty, anything with the potential to supplant Hold’em will likely need to move further outside the box. Next week, we’ll take a look at which corners of the design space have not yet seen much exploration.

Next: Successful Experiments

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

Get your daily dose of poker news with the PTP Hit and Run.

Daily Fantasy Sports Promo and Referral Codes >>>
comments powered by Disqus
IpVanish VPN


Don't want to deposit? Try free bankrolls.
Room Bankroll
Sign up at Titan Poker Titan Poker
Major room, easy qualify
Sign up at Sky Poker Sky Poker
Good for MTT / SNGs Only

Follow PartTimePoker ®

RSS Feed
Donate Bitcoins


About Part Time Poker ®

Online since 2004, brings together a unique combination of top-paying poker rakeback and prop offers and a variety of poker-related content including poker news, strategy articles, free poker training reviews and the latest Poker News and Gossip. - your best source for online poker.

Inside PTP

Rakeback and Free Bankrolls

PTP offers several rakeback and free bankroll offers for our viewers. If you're not familiar with rakeback, read our guide to online poker rakeback. To learn how much you could be earning with rakeback, check out our rakeback calculator.

Free and no deposit bankrolls (also called free poker money) are essentially promotional deals we've arranged with rooms where they give you a small amount of money (usually $10-$150) started.