Unlocking the power of cryptocurrencies to step inside the decentralised casino

Researchers from the Department of Computer Science at the University of York have presented the first ever analysis of gambling transactions taking place in cryptocurrency casinos on the Ethereum blockchain.

In uncovering extreme behaviours, the study could assist in identifying the potential for financial harm via unsustainable spending among players, ultimately leading to better consumer protection tools.

The study unlocks blockchain transaction data to provide a detailed summary of spending behaviours, revealing how people are affected by this new form of online gambling.

Decentralised gambling applications differ from traditional online casinos in that players use cryptocurrency as a stake. The codes running the casino games are also stored on public ledgers, known as blockchains and cryptocurrency platforms such as Bitcoin and Ethereum use the technology to process payments and calculate game outcomes.

When players place wagers, their transactions are recorded on a blockchain in an anonymised form. Whilst this data has always been publically available, the technologically advanced nature of the applications has presented barriers to research and regulation.

As part of the study, researchers developed an algorithm to extract and decode transaction data from the Ethereum blockchain. The work examined more than 2.2m transactions from 24,000 unique addresses on three applications operating on the Ethereum cryptocurrency network (dice2.win, etheroll.com and fck.com).
Focussing on simple casino type games of chance, like dice rolls and coin flips, researchers found that the average decentralised gambling application player spends less than in other online casinos overall, but that the most heavily involved players spend substantially more.

As with regular online casino players, researchers found that those who generally place larger bets are more likely to wager larger total amounts over the duration of their betting careers, however this appears to be amplified in decentralised gambling application use.
The data revealed that the most heavily involved bettors wagered an average of 1,000 ETH (equivalent to approximately $100,000), which dwarfs the average 1.1 ETH (~$110 at time of study) presented by the majority of bettors.

In their assessments of typical ‘player’ behaviour, researchers were also able to identify the presence of non-human players, known as bots, in the data set. Bots may exist for a number of reasons, for example to artificially inflate the perceived popularity of the applications they are transacting with, or to attempt to win the jackpot from an application once it becomes statistically worthwhile to pursue.
The researchers cite that whilst they cannot infer the reason behind the bots’ existence, the identification technique they have developed presents an interesting area for future investigation.

Oliver J. Scholten, a PhD Researcher with the EPSRC Doctoral Centre for Intelligent Games and Game Intelligence, and member of the research team commented: “The study illustrates the power and scale of transaction data that decentralised gambling applications can provide to researchers”.
Scholten added: “These applications could really change the way people gamble online – we in the research community need to know how they operate, how to analyse them, and ultimately how to identify those in harm’s way.”

Dr James Walker from Digital Creativity Labs commented: “The work draws attention to cryptocurrency transactions as a tool for large scale in-vivo gambling research, and presents a robust foundation upon which multiple avenues of further analyses can be performed.”

The article is published by The Public Library of Science ONE at https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0240693