When Science Cannot Predict Success, Should Luck Decide?
As research funding experiments with lotteries, science confronts an uncomfortable question: can chance sometimes choose promising ideas better than experts?
The Problem With Choosing the “Best” Science
Science is built on a simple expectation: good ideas should win. Researchers develop proposals, experts examine them, funding agencies compare their potential, and the strongest projects receive the money needed to continue.
At least, that is how the system is supposed to work. In reality, predicting which scientific idea will eventually produce an important discovery is extraordinarily difficult.
The problem is not necessarily that scientists or reviewers are incompetent. It is that the future value of genuinely original research is often impossible to measure in advance.
A proposal can appear methodologically perfect and still produce nothing significant. Another may look uncertain, unconventional or even impractical before eventually changing an entire field.
This has led some research funders to experiment with an unusual idea: once proposals have crossed a minimum quality threshold, why not use a lottery to decide which ones receive funding?
The idea may initially sound absurd. Science is supposed to be rational, evidence-based and merit-driven. Allowing chance into the funding process appears to challenge everything that system represents.
But the lottery is not designed to replace scientific judgement entirely. It asks whether scientific judgement should be used differently.
The Lottery Begins After Quality Control
The important distinction is that lottery-based research funding does not mean randomly giving money to anyone with an idea.
Under this approach, expert reviewers first examine proposals and remove those that fail to meet the required scientific or methodological standards. Only proposals considered sufficiently strong enter the lottery.
The reasoning behind this system is surprisingly straightforward. If ten proposals are all excellent, but only three can be funded, reviewers may be forced to make extremely fine distinctions between ideas whose future potential cannot genuinely be predicted.
At that point, the traditional process can create an illusion of precision.
A panel may rank Proposal A above Proposal B, but that does not necessarily mean Proposal A will produce better science. It may simply mean the reviewers found it more convincing at that particular moment.
The lottery therefore asks a provocative question: if experts cannot reliably distinguish between several equally strong ideas, is pretending that they can really more rational than using chance?
The Hidden Bias of “Merit”
Traditional peer review is often presented as the fairest possible system because decisions are made by experts. Yet human judgement is never completely free from bias.
Reviewers can be influenced, consciously or unconsciously, by institutional prestige, established research traditions, familiar methodologies or the reputation of the researcher.
This does not mean that peer review is fundamentally corrupt. It remains one of the most important mechanisms through which scientific standards are maintained.
But peer review is much better at identifying bad science than predicting revolutionary science.
That distinction is crucial. A highly original project can be difficult to evaluate precisely because it does not resemble what has already been done.
A proposal based on established methods and familiar assumptions may appear safer, more credible and easier to fund. The result can be a system that unintentionally rewards certainty over experimentation.
A lottery among proposals that have already passed a rigorous quality threshold could therefore reduce the influence of small, subjective differences.
Instead of allowing prestige or persuasive presentation to determine which of several strong projects receives funding, chance becomes a neutral final mechanism.
The greatest scientific discoveries are often celebrated for being unpredictable. Yet the funding systems behind science frequently demand that researchers predict their results before they are given the money to discover them.
Can Chance Encourage Better Science?
One of the strongest arguments for funding lotteries is that they may encourage researchers to take greater intellectual risks.
Scientists often spend enormous amounts of time writing applications designed to convince funding panels that a project will succeed. That creates an incentive to propose research that appears achievable, measurable and safe.
But scientific progress does not always emerge from safe ideas. Some of the most important discoveries begin with uncertainty.
Researchers do not always know where an experiment will lead. If they did, it might not be research in the most meaningful sense.
A lottery-based system could create more room for ideas that are difficult to rank but clearly worth exploring. Researchers would still have to demonstrate that their proposal meets serious standards.
But once inside a pool of credible projects, they would not necessarily need to outperform every other applicant through increasingly elaborate claims about predicted outcomes.
This could also reduce the enormous administrative burden associated with grant applications. Scientists often spend months preparing proposals that have only a small chance of being funded.
Reviewers then spend additional time ranking applications whose differences may ultimately be impossible to measure objectively.
When competition reaches the point where excellence becomes almost impossible to distinguish, the question is whether more judgement produces better decisions or simply creates the appearance of precision.
The Case Against Leaving Science to Luck
The lottery idea also has obvious weaknesses. Scientific research is expensive, and public funding agencies have a responsibility to demonstrate how taxpayer money is being spent.
Critics can reasonably ask why a government should use chance when it has access to leading scientists and technical experts.
Not every funding decision is suitable for a lottery. Major infrastructure projects, expensive laboratories and high-risk medical trials may require detailed strategic evaluation.
Research priorities also matter. Governments may legitimately decide to invest more heavily in areas such as climate science, public health or emerging technologies.
The strongest case for lotteries is therefore not that chance should replace expertise. It is that chance may sometimes be more honest than false certainty.
Expert judgement remains essential for determining whether a proposal is scientifically credible. The lottery enters only when reviewers are confronted with a group of projects that are all worthy but cannot all be funded.
The Future of Scientific Decision-Making
The debate over research lotteries ultimately reveals something deeper about science itself. Modern institutions are under enormous pressure to measure performance, predict outcomes and demonstrate efficiency.
Funding agencies want to know what a project will produce before investing in it. Researchers are expected to explain their expected impact years in advance.
But discovery does not always operate according to a business plan.
Some of the most valuable scientific outcomes cannot be scheduled, guaranteed or precisely forecast. They emerge from experiments that fail, questions that appear unimportant and observations whose significance becomes clear only much later.
The more aggressively science tries to predict discovery, the greater the risk that it funds only the kinds of discoveries that already look familiar.
That does not mean abandoning standards. It means recognising the limits of prediction.
A carefully designed lottery could become one tool for dealing with those limits, particularly when experts agree that several projects are all good enough to deserve support.
Perhaps the uncomfortable truth is that science does not become less rational when it admits uncertainty. It becomes more rational when it stops pretending uncertainty does not exist.