Analysis of FOSS Gaming Failure: A history of PokerSource and Open Source Online Poker

Bradley M. Kuhn

LinuxConf Australia 2020

Monday 13 January 2020

ebb.org/bkuhn/talks/LCA-2020-Games/games-lca-2020.html or tinyurl.com/lca-2020-games

DISCLAIMER

This talk is only tangentially related to my day job (i.e., the talk is about software freedom!). While I don't think I'll say anything in this talk my employer would strongly object to, my employer probably has no opinion on the online poker industry, other than Conservancy's general opposition to any proprietary software. Nevertheless, the opinions in this talk are entirely my own.

I Don't Know Games

I'm better known for my work on software freedom policy and licensing…

…But, I did a side-gig working on a FOSS game for about 3 years.

This is the story of its failure…

(… and its transient success).

Poker Is Different

How would you define the game of poker?

“Correct” Definition of Poker

Poker is a gambling game of strategy played by people for money, using cards (priority of these things are in that order).

Brief Popularity History of Poker

Poker is very popular in USA, UK, and AU.

Played regularly in “home games”.

Saloon poker is mid-1800s USAmerican mythology (some of which is based in fact).

Casinos have offered it worldwide since the late 1800s.

Mainstream popular rise in 1999: UK's Late Night Poker TV series, featuring No-limit Texas Hold'em (NL HE).

How Poker Works

I won't present the rules of poker.

Sufficient for this talk to understand these points:

• Zero-sum, partial-information game where players are against each other, not “the house”.
• Game of skill with a strong element of chance.
• Without careful study, hard to know if you're playing well (poorly) or just (un)lucky
• House takes a rake (fixed fee for running the game for players).

Rise of Online Poker

In 1998/1999, first online poker sites, Planet Poker and Paradise Poker, launch.

Comparatively speaking, poker is straightforward to implement (2-D, minimal graphics needed).

Gameplay Implementation

While some tricky bits exist…

compared to modern MMORPs, etc., a poker game engine is almost trivial:

• Completely turned-based.
• minimal timing/network-speed issues.
• Client need not know secret information.
• minimal UI preferred by serious players(!)

The real challenges are playerbase scaling and operational support software.

Rise of Online Poker

Watershed moment: Chris Moneymaker (yes, it's his real name from birth. Journalists verified it.)

Moneymaker wins an US$86 entry fee online poker “satellite” to gain entry to the World Series of Poker (WSoP).

Moneymaker ultimately wins the 2003 WSoP Main Event for a prize of US$2.5 million.

Rise of Online Poker

During the Moneymaker era (2003—2011), there are three major online poker “networks” (all 100% proprietary): — Full Tilt Poker, Pokerstars, and Cereus Poker Network (UltimateBet and Absolute Poker)

Plus dozens of secondary and tertiary networks with “skins”.

Online poker market was estimated in the multi-billions USD by this time.

FOSS Was Both Early & Late

The history of Free and Open Source Software (FOSS) is riddled with playing “catch-up” to existing proprietary.

Ironically, IRC Poker had been the first “online” poker system (for play money only).

By the early 2000s, no one in FOSS was maintaining its infrastructure.

Ok, Maybe A Bit About Poker Rules…

Poker hands are “ranked” in an order.

You win a given hand by either getting everyone else to give up (fold), or by having a better/best hand if 2+ players stay in.

Most basic order of hands: no-pair, one-pair, two-pair, three-of-a-kind (trips), straight, flush, full house, four-of-a-kind (quads), straight-flush.

What Beats What?

Variant poker games have different rankings (such as lowball, split-pot games, etc.)

Most basic software you need to start:

{ <, >, == } operators to evaluate hands for every conceivable poker game.

The One True Hand Evaluator

The first comprehensive poker hand evaluator and odds calculator that ever existed, amazingly, was under the GPL.

Highly optimized, using minimal storage for hand representation encoding and bitmasks to quickly compare hand values.

Started in 1994 by Cliff Matthews (ctm), it supported every poker game played at the time.

The One True Hand Evaluator

I don't know for sure, but I suspect every online poker system uses (a fork of) it.

If only network services copyleft (i.e., Affero GPL) had existed in 1994. :)

FOSS Tries to Leapfrog

A French company called Mekensleep attempts to make the ultimate 3-D online poker site (spoiler alert: it failed).

But they hire my lifelong friend, Loïc Dachary as their CTO.

Loïc morally will never write proprietary software, so the entire system is under the Affero GPL and/or GPL.

No One Wants Poker to Look Like a Video Game

Poker, in essence, is gambling.

The goal is to maximize hands per hour (MORE GAMBLING).

No one plays just one table (usually at least 2-3).

The player-preferred UI, thus, has barely changed since 1998.

FOSS's Opportunity Lost

Online cheating and government raids became a plague for poker.

Cereus (UB & Absolute) founders actually wrote a “god mode” to cheat customers.

Full Tilt mismanaged funds (comingled player balances with operating expenses) and played a shell game.

Transparency of FOSS could have helped here.

The Harder Software

Focus on making only the game work meant that infrastructural software for site management wasn't written:

• Cashier services
• Player database management and marketing
• Collusion detection

The Harder Software

Even though Pokersource was a “complete system” — fast, efficient and scalable server, two clients, including the first Javascript client in online poker history — you couldn't launch a online poker business with it. Then …

Repurpose to Play Money

Loïc built a consulting business, called OutFlop, around the existing code.

Primary client was a French social media site who deployed play money version for prizes.

We were a network of contractors, and Loïc shielded each contractor from deployment — doling out small tasks (2—10 days work) and paying by project.

Was this Success?

For approximately three years, about 10 contractors had paying, part-time jobs improving a 100% FOSS game codebase.

Should we really define FOSS success as a project that lasts a generation (like Linux)?

Why FOSS Online Poker Is Likely Hopeless

October 13, 2006: USA's Unlawful Internet Gaming Act

USA DoJ warns all poker sites to cease USA operation immediately; PartyPoker is the only one that does.

April 15, 2011: All poker sites still operating in the USA are shut down.

Full Tilt's pyramid scheme is revealed within weeks.

Why FOSS Online Poker Is Likely Hopeless

Pokerstars quickly makes a deal with USA DoJ to pay a fine and acquire FullTilt's player debt.

Pokerstars now holds a strong majority in the market; other players are long-standing stalwarts.

More importantly, the investment in operational software would be prohibitive for a FOSS competitor.

Is There a Future in Play Money?

The FOSS code is there (even with full test coverage on poker-network, thanks to me :) to deploy play-money poker.

Why isn't anyone using it?

Zynga, as it turns out, dominated the play-money market because they were early on Facebook.

Would there be value in play-money Mastadon Poker as a social media add-on?

Is There a Future in Play Money?

Fact is, people who love poker enough to deploy a FOSS play-money site think play money poker is not even a game worth playing.