Fast Fold in Poker

Probability

Overview

I released my poker game app “Exotic Poker” this week on February 3rd. However, around the same time, another poker app garnered significant attention in Japan: “Edge Poker,” developed through a collaboration between DeNA and the globally renowned Yokosawa, was released on January 27th (I only learned of this on January 30th…). Of course, my app is completely unknown and probably not even worth comparing, but in terms of being an app that placed some expectation on using AI for poker, it might have been developed with similar hopes.

 The reception for “Edge Poker” seems mixed—players closer to beginners or casual users seem to praise it, while advanced players or those who want a more stoic experience seem negative. After actually trying it out, I got the impression that it’s probably aimed at casual players, but I’m not really sure how it actually is. I’m mostly a casual user myself, and I didn’t get as bad an impression as the critics say (or so I thought, but as I played, I started to get frustrated with NPCs who only ever went all-in or only ever folded. I thought the stamina system would be fine for casual players, but it might be tough after all…).

 With the game release complete for now, I was thinking about ideas for the next feature to add, when I came across the Fast Fold feature in Edge Poker. This was a feature that m Hold’em and Poker Chase didn’t have, and I wasn’t familiar with it. So, I was curious and decided to look into the Fast Fold feature, which is the subject of this article.

What is Fast Fold?

Fast Fold is a game format unique to online poker. It seems to be a system where you can leave the table the moment you fold and immediately move to a new table. When playing ring games with hundreds of players, it seems that as soon as you fold, you are matched to the next table, greatly reducing the time spent waiting for your opponents to act and making it a very efficient way to play poker. Representative examples include

  • Poker Stars’ “Zoom”
  • GG Poker’s “Rush & Cash”

(Both are cash games, and players from Japan who played them were undoubtedly engaging in illegal gambling (though many may be past the statute of limitations), which may be why it is not discussed much in Japan). Players’ stacks are carried over as is, and unless explicitly reset or rebought, the rules are basically the same as ring games, except for the features mentioned above.

Strategic Characteristics

Fast Fold games require different strategies than standard ring games (optimal strategies seem to differ from live games or regular ring games). According to ChatGPT, they have the following properties:

  • The game progresses quickly, and opponents change every hand. This makes strategies based on reading and exploiting opponents’ tendencies virtually nonexistent, making GTO-based play recommended.
  • With minimal waiting time, many players adopt narrow (tight) ranges. Optimal VPIP tends to be lower than in standard games.
  • Due to the rapid turnover, fewer players stick with weak hands, leading to fewer bluffs and bluffs being harder to pull off. Value-focused strategies or bluffs require larger bets.

While the ability to play a large volume of hands quickly might seem ideal for poker practice, it’s important to note that the slight rule differences can have an impact.

System Implementation Approach

Having released my own poker app means I now possess a personal poker application with freely customizable features—a privilege few in the world enjoy. This prompts me to consider how best to implement this functionality. Developing a single-player poker app isn’t particularly difficult, but multiplayer poker app development is comparatively high difficulty and requires significant server-side development skills. This likely means poker apps won’t be mass-produced in the future.

 When implementing Fast Fold ring games, instead of managing each table individually, you need to manage all participants in a single pool, similar to a multi-table tournament. This involves repeatedly returning players to the pool after Folds or Showdowns and instantly generating new tables once the required number of players is reached. Handling dozens to hundreds of tables simultaneously while accepting real-time processing requests necessitates a server function acting as a command center, similar to a multi-table tournament. This requires multithreading to handle numerous communication processes, which is likely the most challenging part of implementation.

 Another challenge is that in Fast Fold, the dealer button and positions do not rotate sequentially. Randomly assigning positions often leads to skewed distribution and unfairness. Ideally, we should count each player’s position assignments and solve an optimization problem using Hungarian algorithm to achieve fair position allocation.

Can Fast Fold be used in MTTs?

Finally, let’s consider whether applying Fast Fold in MTTs could create more efficient games. In “Edge Poker,” the target for Fast Fold implementation appears to be tournaments rather than ring games.

 Naturally, ideas like this are commonplace, and PokerStars apparently had (or has?) a game called Zoom Tournament that resembles Fast Fold MTTs. However, it seems there were issues, and the game design ultimately wasn’t very interesting. The following challenges were reportedly present:

  • Extremely high early eliminations made blind level increases largely meaningless
  • The number of hands played exploded, quickly dividing players into those with large stacks and those with short stacks
  • Elements effective in tournaments, like reading opponents’ tendencies, disappeared (it simply became less fun)
  • The unfairness of the blinds had a disproportionate impact on players’ wins and losses (with a strong element of luck in assigned positions).

To successfully blend Fast Fold with tournaments, a game design that resolves these issues (or introduces alternative fun elements) will likely be necessary. Personally, I’d like to challenge myself to develop such a game in the future, but I’ll revisit the specific game design and write about it again when actually implementing it.

Comments