The virtual currency Bitcoin may have a fundamental flaw according to two Cornell University professors. And they’re not talking about its lack of official recognition or oversight.

Emin Gun Sirer, who specialises in operating systems and networking, discusses the piece in a post on the “Hacking, Distributed” blog in which he details a paper he has prepared with colleague Ittay Eyal. They argue that the system relies on user acting in the best interest of the Bitcoin community, but that it creates too strong an incentive working against this.

The issue centers on the way new units of the Bitcoin currency are “created” when computers solve a cryptographic puzzle, a process designed to control the number of units in circulation and avoid inflation risks. That’s something supporters of Bitcoin consider a major advantage compared with government issued currencies.

As New Scientist describes, each time a computer solves a puzzle (known as successful mining), it adds a new entry to the complete list of Bitcoin transactions, known as the blockchain, which effectively creates a new puzzle to be solved.

In some cases two computers or groups of computers may solve the same puzzle and create different branches of the blockchain. The usual procedure is that when this happens and is discovered, the shorter branch is abandoned.

The problem Sirer and Eyal raise is that a fundamental flaw in the way the Bitcoin system is designed means it’s possible to engage in “selfish mining.” This means working to create a lengthy branch but keep it secret for some time. Eventually the branch would be released and immediately become the dominant branch and force all others to be abandoned. That could mean a big Bitcoin payout for the people behind the dominant branch and little or no reward for people who’d been working on the other branches.

While it’s highly unlikely an individual could get much benefit of selfish mining, Bitcoin allows users to pool their mining resources and then split the rewards proportionally. Sirer and Eyal believe there’s a serious risk that this could create an incentive for new users to join a selfish mining group and that the more people join, the greater the incentive becomes for even more people to join, creating a snowball effect.

To date, the assumption has been that selfish mining could only work if the group using the tactic made up a majority of all mining capability, something that seems unlikely. According to Sirer and Eyal that’s a misconception and as things stand, a small percentage acting as selfish miners could kick off the snowball effect.

They propose a technical change to Bitcoin but note this would only increase the threshold to 25 percent of miners needing to be selfish — and that there’s already a group that exceeds this threshold. They also believe that even a perfect technical solution couldn’t raise the limit to more than 33 percent.

Not everyone backs Sirer and Eyal’s conclusions though. Several commenters on the blog post argue that the system does in fact have more of an incentive against selfish mining, which is that the narrower the control and influence over Bitcoin become, the less credibility and trust the currency will have, reducing its value in comparison to “real world” currencies. They argue that Sirer and Eyal mistakenly assume selfish miners are motivated by the raw number of Bitcoins they can gain, rather than also taking into account the “value” of each Bitcoin.