manifold is opensource. Why do you want 20k for play money mvp?
I aim to develop a prediction market platform - like Manifold, but using real money - so people can bet on which scientific studies will replicate.
For more than 60 years, people have been warning about issues reproducing science - and yet neither the public nor the private sector have responded adequately. My project aims to solve this by using human nature and the proven power of prediction markets.
Although the term “replication crisis” did not become common until the 2010s, different scientists and statisticians were raising the alarm about issues with replicability as early as the late 1950s and early 1960s. Their concerns were roughly concurrent with scientists and scholars of science and technology studies noting a shift towards “big science” and further professionalisation of science, its practices and its publication.
As researchers at the Center for Open Science have pointed out, current scientific publication models do not incentivize replication. Public funding bodies will - barring some exceptions - favour novelty over replication, as argued by Nosek et al. Higher-prestige journals publish papers that are less likely to replicate precisely because their results are so surprising. The private sector will not necessarily fund replication studies.
This is where prediction markets for science can have a positive impact. By betting on which studies will replicate, we can attach financial incentives and prestige to replicability. Over the past eight years, multiple studies have shown that both experts and laypeople can use prediction markets to accurately predict which study results will replicate. Moreover, Korbmacher et al. writing in Nature identify prediction markets as part of positive “procedural change” in the practice of science.
Although real-money prediction markets are still a stretch in the US, I grew up in the UK where we bet money on all sorts of outcomes - from which soccer team will win to who will be our next Prime Minister. Culturally, we’re so used to betting that getting people to wager money on science is not a big stretch. Moreover, offshore gambling is legal or in a grey area in plenty of countries outside the US. Since plenty of people in these countries also care about betting and science, they represent a vast untapped market. We have the potential to tap into appetites to make science better across the world.
I care fiercely about the replication crisis and want science to work far better than it does. Prediction markets for science show promise and play-money prediction markets have been around for several years. Track records demonstrate a) we know people like to predict stuff b) we know that prediction markets work c) we know how to make prediction markets more fun to use, more useful and more profitable.
I want to take the work people have already done and go further. I want to create a platform where anyone can create a prediction market for whether the results of a published study in any scientific field will replicate, where anyone can bet real money, and where we can publicise the platform and keep it running for years - there have been previous prediction markets for science but the projects are smaller and more limited in scope.
Those projects are so promising that they need and deserve to be scaled up - as I propose to do - and made more accurate by betting with real money - as I propose to do. With my funding ask I want to build a play-money minimum viable product, apply for a Curacao gambling license so that users can legally bet real money, and within the year I hope to hire a team and be bringing in revenue from per-transaction fees.
I grew up in a culture where betting is very normalized, so I understand how and why people will bet, and their attitudes to betting and prediction.
I have an integrated master’s degree in physics from the University of Manchester (graduated 2018 - an integrated master’s degree is a 3-year bachelor of science plus an extra year of research for the MPhys award), a Master’s degree in Science Communication from the University of Manchester (graduated 2019), and a PhD in the history of UK space science from University College London (I passed my defense and was awarded my degree in 2023). I am the world’s foremost expert in the social and cultural history of the UK space program.
During my research year in my physics degree, I spent a year developing a Raman spectroscopy experiment. During the course of that year, I spent hundreds of hours sitting in a dark room learning how to troubleshoot a rather temperamental laser setup. These experiences taught me about how people do science in practice.
During my Master’s degree in Science Communication, I:
- Commissioned articles for and copy-edited an astronomy magazine with a readership of over 2000
- Conducted original research with planetary scientists at NASA about their use of video games for public engagement
- Produced an advocacy policy brief for the Indonesian government, proposing the creation of a dark sky reserve in West Timor.
For my PhD, I:
- Lived and worked among the scientists, engineers and administrative staff of the Mullard Space Science Laboratory in the English countryside
- Conducted more than 100 hours of oral history interviews with scientists and engineers to collect their reflections on how they do science in practice
- Catalogued spacecraft parts and technical drawings for the Mullard Space Science Laboratory, laying the groundwork for preservation of these important artefacts
- Catalogued and collected records of spacecraft components for the Science Museum in the UK
- Developed and carried out the first demographic survey of amateur astronomy in the UK
- Arranged talks and activities for late-night events at the Science Museum
- Contributed research about the cultural history of Mars observation to an upcoming exhibition at the Science Museum
- Investigated how undergraduate physics students at University College London were affected by COVID-19 lockdowns in the UK.
I have subject matter expertise in seeing how people do science in practice, I have expertise in project and time management on a shoestring budget, and I have experience of managing an interdisciplinary years-long project on an extremely minimal budget. I also did this interdisciplinary years-long project during a pandemic and had to change my entire methodology, so I have proven myself extremely capable of adapting to unexpected adverse circumstances.
Given my track record with carrying out interdisciplinary research under adverse ever-changing circumstances, I am confident that I can build on what other people have already done to make a larger-scale for-profit prediction market platform.
For references, you are welcome to contact:
Jon Agar (jonathan.agar@ucl.ac.uk)
Doug Millard (doug.millard@sciencemuseum.ac.uk)
Lucie Green (lucie.green@ucl.ac.uk)
Anasuya Aruliah (a.aruliah@ucl.ac.uk)
Mark Fuller (mark.fuller@ucl.ac.uk)
Twitter: @astrosnat
Bluesky: astrosnat.bsky.social
Website: osnatkatz.com
My entire PhD thesis: https://discovery.ucl.ac.uk/id/eprint/10178481/
I *am* a real person, but have been keeping the whole science betting thing under wraps. Partly because I only finished grad school quite recently.
$20k buys a play money MVP. $60k buys a play money MVP and a Curacao gambling license. $150k buys a play money MVP, a Curacao gambling license, and enough money to hire smart people who know how to make a legal offshore gambling platform.
Business model so far: https://www.canva.com/design/DAF2pWJ3KkU/Cr4t_QjuxDxq3nCu1Y1Oxg/edit?utm_content=DAF2pWJ3KkU&utm_campaign=designshare&utm_medium=link2&utm_source=sharebutton
Online pitch deck: https://www.canva.com/design/DAF2ppN_dzk/HKcQw3YOSARHD-8NljxL6Q/edit?utm_content=DAF2ppN_dzk&utm_campaign=designshare&utm_medium=link2&utm_source=sharebutton
"The replication crisis has led to positive structural, procedural, and community changes" - Korbmacher et al. Citations 94-100 also lead you to additional studies about the efficacy of prediction markets for replicating science. I leaned on those studies too: https://www.nature.com/articles/s44271-023-00003-2
50%
Josh Sacks
7 months ago
Academia seems like one place where substituting reputation for $$ makes sense. I'd focus on building a reputation system, rather than solving the hard problem of real $$
Alyssa Riceman
7 months ago
What are the advantages of creating a whole new real-money prediction-market platform specifically for predicting study-replication (which seems difficult and failure-prone considering the general troubles faced by prediction-market platforms), as compared with the obvious alternative of creating markets on study-replication on some preexisting real-money prediction-market platform? Are the preexisting platforms unwilling to support such markets in some way? Are there particular features they're missing which your platform will have? Other factors I'm currently failing to think of?
Osnat Katz Moon
7 months ago
@Alyssa the problem is that preexisting real-money prediction-market platforms have their own troubles. To the best of my knowledge, there are 2: Kalshi and Polymarket.
Kalshi is great but has run into regulatory issues with the CTFC, IIRC. Polymarket uses crypto. The intersection of "people who care enough about science to forecast about it" and "people who use crypto" is small. I'm trying to target a bigger market of "people who care enough about science to forecast about it". In an ideal world, I would persuade Betfair - the world's largest online betting exchange - to start taking bets on paper replication, because that seems much easier than creating a new prediction-market platform.
My hunch is that some of the troubles faced by prediction-market platforms arise from them focusing on US customers, and that expanding to Canadian, UK, EU, Chinese markets would help.
Jason
7 months ago
I don't understand the business model here. For play-money markets, the play money needs to be valuable enough that people will do deep enough dives to generate useful predictions. For real-money markets, the real money seemingly has to come from either from less-skilled-than-average participants, or from donors who want information. Unlike sports betting or even political betting, I don't foresee there being a ton of whales to supply less-informed money that skilled predictors can win. Unless there are donors propping up the markets, the bottom half of the predictors are still likely to be rather rational people; that means they will likely exit as soon as they realize they are more likely the source of revenue rather than likely to win money.
Osnat Katz Moon
7 months ago
@Jason Thanks. This has given me a lot to think about. Hoping to start getting users for a play money version ~May-June, and to talk to some UK research funders about the possibility of using prediction markets to help inform funding decisions, so we will see what happens then.
Osnat Katz Moon
7 months ago
@Elizabeth unfortunately, yes. There are a couple of ways around this:
1) decide "we're going to be an offshore gambling platform", grit my teeth, and go with the fairly tough but doable process of getting a gambling license or applying to be on someone else's gambling license; doesn't work for the US, means a greater focus on targeting non-US markets (e.g. Canada)
2) if you have an organization stump up a cash pool first and then participants in the org can bet, this has fewer legal difficulties (specifically, it doesn't count as gambling any more), but you also have a product aimed at organizations rather than individuals (internal prediction markets do seem to work where they've been tried, though!)
Aditya Arpitha Prasad
7 months ago
I really like this idea. I want this to happen in some form, so I hope this moves the bar on this.
Patrick Delaney
7 months ago
Greetings! I am creating an open source prediction market platform already which is fully functional so possibly could help shave around $20k off of your costs. https://github.com/openpredictionmarkets/socialpredict
Osnat Katz Moon
7 months ago
@PatrickDelaney Awesome stuff, just checked it out now! Yes, let's talk more. My email is osnatkatz at protonmail dot com