Can you steal game gold pieces? The Court of Appeal says yes

If Runescape gold is property, does that mean that golden capes are also property?

It’s been a while since we had a proper digital property and virtual gold story here at Llama Towers, I have to admit that it’s been mostly AI for the last few years. But a recent decision in the Court of Appeal (Criminal Division) in England and Wales has prompted me to write a few words about the always fascinating world of property rights in digital assets (thanks to Andrew Ducker for the heads-up).

The case is R v Andrew Lakeman [2026] EWCA Crim 4, where the Court of Appeal has just handed down what may be one of the most significant UK judgement on virtual property in over a decade (debatable of course, we have D’Aloia v Persons Unknown as well as others). In this case, Popplewell LJ ruled that “gold pieces” in the MMORPG Old School RuneScape constitute “property” under the Theft Act 1968. This isn’t just a quirky gaming story that brings us back to the legal debates of 2006, this is a decision that reminds us of the importance of digital property and digital assets, something that has been lost a bit since the NFT craze of a few years back.

The facts and lower court

The facts are the stuff of every MMO player’s nightmares. Andrew Lakeman worked as a content developer at Jagex Ltd, the company behind RuneScape. He had no legitimate access to player accounts, but allegedly hacked into the credentials of Jagex’s account recovery team. Using that access, he infiltrated 68 player accounts, stripped them of approximately 705 billion gold pieces, and sold them on the black market for Bitcoin and fiat currency. Jagex valued the haul at approximately £543,123 GBP.

For those unfamiliar with RuneScape’s economy: gold pieces can be legitimately acquired by grinding in-game tasks (woodcutting, mining, questing), through the “grand exchange” marketplace, or by purchasing “bonds” from Jagex for £6 each, which can then be traded for about 13 million gold pieces. Crucially, while Jagex prohibits real-world trading of gold pieces in its Terms and Conditions, a thriving black market exists where gold changes hands for roughly half the official bond rate.

Lakeman faced five counts: unauthorised computer access under the Computer Misuse Act 1990, theft under the Theft Act 1968, and three money laundering offences under POCA. All of this hinged on one question: are gold pieces “property” that can be stolen?

In the lower Crown Court decision, HHJ Grey ruled for the defence. His reasoning rested on two pillars. First, he applied the “rivalrousness” test drawn from the Law Commission’s reports on digital assets. I’ve talked extensively about rivalrousness before, but for completeness, something is rivalrous if one person’s use necessarily precludes another’s use, if you have a cake and I eat it, you can’t eat it (insert “have your cake and eat it” joke here). The judge found gold pieces failed this test because “one gold piece is the same as any other, and their supply is infinite. The fact that existing players have wealth does not preclude new players from joining the game and getting more wealth, without taking it from existing players.” Second, he invoked the “pure information” principle from Oxford v Moss (1979) 68 Cr App Rep 183, reasoning that gold pieces were more akin to pure information than to property.

The Court of Appeal

The Court of Appeal emphatically disagreed, and its reasoning deserves careful attention because it explains nicely how criminal law should approach digital assets. Popplewell LJ rejected the notion that property under the Theft Act must satisfy the same criteria as property under private civil law. He cited R v Hinks [2001] 2 AC 241 (HL), where Lord Steyn observed that “the purposes of the civil law and the criminal law are somewhat different” and that “it would be wrong to assume on a priori grounds that the criminal law rather than the civil law is defective.”

This matters enormously as Jagex’s Terms and Conditions explicitly state that gold pieces are “not your own private property” and merely “a measurement of the extent of your licence in a Jagex Product.” If civil law property rights were determinative, the defence might have prevailed. But the Court held that these contractual provisions are “not determinative as to whether they are property for the purposes of the definitions in the Theft Act.”

Here there is a useful analogy to illegal drugs, as English criminal law has confirmed that Class A drugs are property capable of being stolen, even though no civil court would enforce property rights in them. The criminal law cares about keeping the peace, not vindicating individual property rights.

The Court then found the Crown Court’s rivalrousness analysis was fundamentally flawed. Popplewell LJ’s critique is withering:

“We differ from the Judge in his reasoning for reaching the contrary conclusion on rivalrousness. The two reasons which the Judge gave in [25] of his ruling do not, with respect, bear analysis. The first was that “one gold piece is like any other, and their supply is infinite”. This does not, however, distinguish them from many other forms of rivalrous property. One paper clip from a given manufacturer is like any other; and the manufacture and supply of them infinite, in the sense that is not capped at any finite number. Yet each paper clip constitutes property. The same is equally true of gold pieces.” At [77]. 

The fact that new gold pieces can be created doesn’t affect whether the gold pieces you currently possess are rivalrous. If player A uses their gold pieces, player B cannot use those same gold pieces. That’s what rivalrousness means. The possibility of acquiring different gold pieces is irrelevant.

The Court also distinguished gold pieces from “pure information” as understood in Oxford v Moss. That case held that knowledge of exam questions wasn’t property because (1) knowledge exists only in a person’s head, and (2) transferring knowledge from A to B doesn’t deprive A of that knowledge. Gold pieces are nothing like this. They exist outside any individual’s mind, they function as identifiable assets distinct from the code that creates them, and transferring them from A to B does deprive A of them. The Court drew heavily on Professor David Fox’s analysis of crypto-tokens as “ideational things”, more than mere data, grounded in but not confined to their digital design.

This brings us to what I consider the judgement’s most sophisticated insight. The Court carefully distinguished between gold pieces as functional assets and the underlying code that represents them. Lakeman’s defence argued that code is just information, freely replicable. But as the Court noted, the same is true of Bitcoin, the code is public, the blockchain is public, the only private element is the key. Yet Bitcoin is widely recognised as property.

When gold pieces are “transferred,” the code allocating them to Player A is deleted and new code allocates them to Player B. The Court held this is immaterial: “When gold pieces are transferred there is a visual and functional transfer. It is immaterial that this is effected by one piece of code being replaced by another.” The asset is the functional ideation, not the instantiation.

The Court looked abroad for guidance. The Dutch Supreme Court had already convicted someone of theft for physically attacking a 13-year-old and forcing him to transfer a virtual amulet and mask in RuneScape. The Australian Supreme Court of Appeal in Yeates v The King [2025] VSCA 288 (decided just weeks before this judgement) held that Bitcoin is property under theft legislation identical to the UK’s.

Popplewell LJ observed that if the defence were right, the gold pieces “could not be the subject matter of an English charge of theft or robbery in similar factual circumstances to the Dutch case, which we would regard as a surprising and unsatisfactory situation.”

The defence made much of Jagex’s Terms and Conditions, which state that gold pieces have no “inherent or real world monetary value,” cannot be exchanged for real money, and are not the player’s “private property.” But these provisions backfired in two ways: Firstly, as noted above, the criminal law is not bound by private contractual arrangements. If B dishonestly sells A’s car despite a contractual prohibition on sale, B is still guilty of theft. The prohibition doesn’t change what the car is. Secondly, the Terms themselves are somewhat internally contradictory on the property question: clause 14(6) says gold pieces aren’t “private property” but merely measure the “extent of your licence,” which the Court rightly observed makes little sense since the number of gold pieces doesn’t measure anything about the licence.

Concluding

It is refreshing to be looking at digital property again. While I find AI fascinating, the endless bickering over copyright infringement sometimes gets extremely repetitive, and variety is the spice of life. It is lovely to see that digital assets are still worthy of writing about, and if anything, this gives me a chance to update my old virtual worlds slides.

I found the ruling in R v Lakeman was thoughtfully and well-reasoned, and I think that it arrives at the right answer through careful analysis. Criminal law should be able to punish someone who hacks into accounts, steals items worth hundreds of thousands of pounds, and converts them to cash. To hold otherwise would be to privilege form over substance, elevating contractual boilerplate and philosophical abstractions about “information” over the economic reality that millions of players experience every day.

The Court’s willingness to distinguish criminal from civil property concepts, its proper understanding of rivalrousness, and its sophisticated treatment of the code/asset distinction provide a template for future cases involving virtual property, NFTs, and digital assets of all kinds.
The case now returns for trial. Given the strength of this judgement, Lakeman’s legal team will need to find different ground to contest the charges.

At least I got an excuse to re-read one of my favourite old posts. Ah, digital theft, those were the days!