Iryna Pasichnyk / Getty ImagesThis month, the Australian Research Council (ARC) announced that some funding in its new “Explore” grant scheme will be allocated to scientists by a conditional lottery. Reviewers will classify every application as highly competitive, competitive, or not competitive. Highly competitive applications will be funded and non-competitive ones will not, which is standard. However, the “competitive” applications are where it gets interesting: they will go into a lottery, with winners receiving funding. It might seem like a strange idea to distribute millions of taxpayer dollars by random chance – but evidence from around the world shows it might actually be a very good idea.Random rulesLotteries are an age-old way of making difficult choices. In ancient Athens, public officials and jurors were chosen by sortition, so that office could not be captured by the wealthy and well connected. In more modern history, lotteries have been used to admit babies to London’s Foundling Hospital, to conscript Australian men during the Vietnam War, and to allocate car licences in Beijing to reduce traffic pollution.Research funding is a more recent application of lotteries, but they are increasingly being used by funders overseas. New Zealand’s Health Research Council led the way in 2013, followed by Germany’s Volkswagen Foundation, the Austrian Science Fund and the British Academy. The Swiss National Science Foundation now lets reviewers draw lots as a tie-breaker in all its funding schemes. A 2022 Nature editorial made the case for lotteries as a tiebreaker of quality in research funding.Allocating funding by lottery may sound antithetical to science. Surely all applications can be ranked as more or less deserving? However, applications are lengthy documents with complex methods and unknown returns, so peer reviewers often disagree about their merits.Hence, winning funding is already somewhat random, because it depends on the luck of the draw of the peer reviewers. Getting a reviewer who knows your area and understands your methods increases your chances. Drawing one who is sceptical of your approach, or thinks there’s a better method, can sink you.The fairness argumentAnother argument for conditional lotteries is fairness. Every human is biased, and those biases can be amplified when making knife-edge decisions between equally qualified applications. For example, if two applications are impossible to separate, but one comes from a well-known university, that prestige can influence the decision, even subconsciously.There is evidence that conditional lotteries increase diversity in applications, too. When the British Academy introduced a lottery, the overall number of applications increased, but there was a greater increase among researchers with an Asian and British Asian background. These applicants likely felt they had a better chance when the final decision had no human influence.Efficiency and impactConditional lotteries are also efficient, reducing the time reviewers spend splitting hairs between equally good applications.A scheme in Germany has even used a lottery to win the right to apply. This saves scientists the time of writing unsuccessful proposals, which is the largest cost of any scientific funding scheme. Reducing the number of applications also reduces the costs of running the funding system. Running costs are likely to grow as researchers use AI to write more proposals, meaning more funders may turn to a lottery to make their task more manageable. The German scheme also saw a large increase in the number of female applicants, likely because women perceived a greater chance of success when human biases were reduced.The biggest gains might come in the returns to society. Researchers know that risky applications are less likely to win funding in traditional systems, so they often propose safer ideas. Explicitly including chance in funding decisions could free researchers to put forward truly novel ideas. There may be more failures, but also more breakthroughs.A conditional lottery also adds a way to measure these returns. Because funding is allocated by chance, winners and those who miss out are, on average, alike in every other respect. We used this approach in New Zealand to run the world’s first randomised trial of the impact of funding on researcher productivity. Funders can thereby accurately estimate the impact of their grants on a wide range of outcomes, from publications and citations to patents, spin-out companies, and changes to public policy.A first for AustraliaThe ARC Explore grant is Australia’s first major conditional funding lottery. Those who have never seen how funding decisions are actually made might decry funding lotteries as an abrogation of responsibility. Yet the citizens of Athens, the world’s first democracy, trusted the lottery to choose their officials. The evidence is that chance can also be trusted to choose between research proposals that experts already judge to be of comparable merit.Chance has always played a part in who gets funded. The ARC has simply decided to be honest about it, and, as an added bonus, create a perfect way to measure the benefits of research funding. If all goes well, more Australian funders may do the same.Adrian Barnett received funding from The British Academy to investigate lotteries and randomises applicants for their schemes. He is a member of the NHMRC Research Committee. Philip Clarke receives research funding from British Academy to investigate the use of lotteries in research funding and receives research grants from a variety of sources both in Australia and overseas including the UKRI, NIHR, JPAL and the MRFF.