CM 002: Market Manipulation, NFTs, and PMs for Policy Makers

Low liquidity woes; Reality Cards CEO Interview; and Policy as Avenues for Prediction Market Growth.

💰 Welcome back to Crowd Money, a new monthly newsletter covering real-money prediction markets (PMs). Every month, Andrew and Clay take the Crowd on a tour of these markets giving you insights, trades, and new ideas, and introduce you to top-market traders, researchers, and innovators in the space. Topping it off, we’ll be making trades of our own on a quest for ever-increasing triple-digit returns.

Two weeks after our newsletter, we release the Crowd Money Cast, a podcast where we discuss our latest prediction market thoughts and interview key players in the space. In the first episode, we spoke with Kalshi co-founders Tarek Mauser and Luana Lara Lopes and talked about how prediction markets compared to prediction platforms (or polls) like Metaculus and Good Judgement Open.

We’re happy to announce the Crowd has already grown to over one hundred members🎉🎉 (101 to be exact)! Interested in the hottest markets of early October and the biggest PM news from September? #JoinTheCrowd to get next week’s supplement in your inbox and be a part of the next 100.

And don’t forget to ❤️, it’s like a tip—but free!

The Crowd Asks…

As we talked about in the first issue of Crowd Money, market makers are key to ensuring liquidity. But sometimes they fall asleep…or there just is not enough liquidity to go around. In the comment section of issue 001, Crowd member Ben asked: “How high is the risk for market manipulation in real-money markets which may have lower (real) liquidity?”

In short: Moderate to high, but with questionable impact.

While the initial risk is high, there are counteracting forces at play too. As you will shortly see, in low liquidity markets (such as when the automated market maker is down) it is easy to manipulate the market price. However, two things are going to happen. One group of people is just going to ignore the market price and submit offers closer to their forecast, while another is going to take you up in your market manipulation attempts by either buying under-valued contracts or selling over-valued ones.

Sure, a random spike from 10c -> 99c might give one pause...but that pause eventually turns into a scramble to transfer more funds into your account. Eventually the taxman (or resolution-criteria-man in this case) cometh, and the question either did or did not resolve positively.

This is largely why the impact is questionable. You derive your own forecast, and in a low liquidity market you should really not be relying on market prices as key signals for your forecasts. If you do, that’s on you.

So what does market manipulation achieve? It parts the fool from his money, while enriching those with their own, confident forecasts. It costs a lot of money to maintain, and as a result eventually brings liquidity to the market. In the end, there’s nothing those manipulation dollars can really do to manipulate the price when the resolution arrives (unless you like, finance a coup or something).

The real impacts come in the distortion of the order book and making it hard for individuals to enter the market (relatively minor), while also distorting the information produced by the markets themselves (definitively adult). For example, it is hard to deploy prediction markets as tools for policy makers or systems for filtering the news for individuals, if the price is being manipulated. And manipulation can definitely happen (although less likely now in this case).

Which brings us to Futuur, a project that has caught our eye particularly for this reason. The platform combines the prediction platform with the prediction market, so information comes from both truth-hunting forecasters and alpha-seeking traders. Neat!


Prediction Markets and Policy Makers

In the first issue of Crowd Money and episode of the Crowd Money (pod)Cast we talked about the various use cases for prediction market platforms and individual markets. Some markets seem primed for hedging opportunities while other markets may be helpful in understanding what news is meaningful and not simply sensationalist media. One of the most interesting applications of prediction markets to us, however, is the use of information from prediction markets to inform policy.

The ceiling for this use case is obviously quite high. Take this market on Polymarket regarding California Governor Gavin Newsom’s gubernatorial prospects, for example. 

Although the recall election has since passed, and this market has likely reached its terminal pricing, there was a time when this market had massive import. And as such, this market is still a great case study in just how impactful some of these markets can be.

In July, Newsom passed a $100 billion piece of legislation, the California Comeback Plan, which allocated funds towards Covid-19 relief programs and California’s carbon-neutral goals, among other things. $5.1 billion of this plan was earmarked for “drought support, water supply and natural habitat restoration projects around the state to build climate resilience in the face of more extreme cycles of wet and dry.” If you are a developer looking to secure a big government contract to construct this climate change mitigation infrastructure, Newsom’s political future is especially relevant.

  • Two other examples from Polymarket were featured earlier this month in The Pomp Letter, including the likelihood of the 2021 Tokyo Olympics being canceled as well as the third-wave of Covid brought to you by the Delta variant.

The floor for this use case is clear as well. In markets with low monetary liquidity, as we observed this week, information is harder to glean. Not only is information liquidity also low in these markets, but information in these markets can be drastically affected by low amounts of volume, making these markets less reliable for big decisions. So for a market to be useful for policy, it needs to cover an important question and reach a threshold of liquidity, and this combination has been difficult to achieve on many of the prediction market platforms we have used so far. 

There is also the question of whether policymakers should use information from real-money prediction markets like Polymarket and Kalshi, or prediction platforms like Metaculus and Good Judgement Open. Will this be an either-or decision, or a both-and? Will policy makers have a single market platform that they trust most, or reference a broader group of platform providers?

Leave a comment

There is research indicating using real-money prediction market information in the final days or weeks before resolution will give you the most accurate data to form policy, whereas prediction platforms are more useful further out. But given the gravity and import of policymaking, it may yet be too soon to make a permanent decision on this matter.

And then there is the question of prediction market adoption within government bodies. This trend has gained momentum recently, with the Virginia Department of Health partnering with Metaculus to run the Keep Virginia Safe Tournament to inform important policy decisions within the state. But as we know, some politicians might view accuracy as a liability, and would rather please their constituency in order to get reelected rather than make unpopular decisions based on compelling data. So while some politicians might embrace prediction markets as another tool, another input to inform their policy, others may view it as a threat. And if they do, they will surely be the loudest voices in the conversation.

Word on the (R) Street

We spoke with Mary Brooks, a senior research associate for Cybersecurity and Emerging Threats at R Street Institute, about this issue for further insights. Mary wrote an article on prediction markets and policy for Lawfare in July, and was generous enough to share her observations with us as she continues to conduct research and write content on the growing space:

We've seen relatively few examples of governments soliciting crowd-forecasting to help guide policy making. Three things need to be defined before this can happen at scale.

First, the information taken from a market should be fairly reliable. Otherwise it’s just more noise.

Second, policymakers need to decide how they can responsibly use crowd-source-derived information. No one is ever going to say, 'Well the market thinks this policy will be created so let’s do it.' Instead, a market might be used to get more information about an issue, or to corroborate inputs and compare against other information sources.

Finally, the market must be structured so that someone with an agenda can't skew results toward the outcome they want--even though the causal relationship between the market and the policy will likely never be so direct. The market needs to reassure users and the public that answers can't be manipulated to influence policy outcomes.

As Mary illustrates above, the state of forecasting and prediction market platforms today are not necessarily conducive to policy making. This is not to say that information from prediction markets cannot be used to make decisions, but rather that the platforms and dialogue around forecasting do not lend themselves to action.

The goal of Metaculus and Good Judgement Open is ostensibly to achieve a low brier score, gain respect with the forecasting community, win tournaments with accurate forecasts, and become a Superforecaster. On Kalshi and Polymarket the goal is to generate alpha.

The goal of policy, however, is to take forecasts one step further, and use that information to make a decision. Most forecasters and prediction market bettors are not thinking about anything past their probability of an event taking place. Once focus shifts to using prediction market information for decision-making (this will likely happen after the market making and liquidity issues we discussed in our first issues are solved) policymakers may begin to adopt the technology at a larger scale.

And once they do, the ceiling for prediction markets will rise. The same way that entities like Tesla, Square, and El Salvador adopting Bitcoin gave it legitimacy and staying power which will allow it to reach its potential, the same is true for prediction markets. Policymakers using prediction market information will validate the quality of the data coming from these platforms, as well as the significance of the questions and markets being created. Once this happens new user inflows will rise and prediction markets will be able to scale in a sustainable way. Similarly, policy may also become more purposeful and objective, with prediction market data validating the decisions that politicians make through more transparent and quantitative means. 

PM Freestyle Interview 

Disclaimer: We do not own any assets created by Reality Cards.

There’s an exciting new project in the prediction market space, and it’s not the “traditional” prediction market platform that you’re probably used to. Enter stage left, Reality Cards: a new Decentralized Finance (DeFi) prediction market built on top of Non-Fungible Token (NFT) technology.

On Reality Cards, each event contract has a unique NFT for each outcome. Unlike other exchanges where you purchase “Yes” or “No” shares for a particular prediction market, you rent an NFT. Each potential outcome only has one NFT associated with it. Here’s where things get interesting.

Each NFT also has a rent price. As the first owner of an outcome NFT, you are able to set the rental price. A certain amount of USD Coin (USDC) is deposited and sent to the contract, and the rental price is deducted from that deposit each hour.

If you were thinking “Doesn’t this mean there can be outcome NFTs for events with no rent?” The answer is yes! If you want to rent an outcome that’s already owned, you can always just take it from the current owner for a higher rental price. But interestingly, the way that Reality Cards works, if an owner runs out of deposit on an outcome, the NFT returns the previous owner at the previous rental price, so even once you’ve lost an NFT it is wise to keep your deposit in the outcome.

Upon resolution, the aggregate rent of the owners of the winning outcome NFT are totaled. Then winnings are allocated based on the length of time you owned the outcome NFT for. The measure of time used when distributing the aggregate winnings is seconds owned. Additionally, the individual who owned the outcome NFT longest, gets to keep the NFT, which will subsequently hold value as a collectible and piece of art. It’s important to note here that both the winning and losing outcome NFTs are kept by their longest holder, so there is value to be gained irrespective of accuracy. Currently, Reality Cards’ open events span a range of topics, from sports to music to crypto, with new events coming on the horizon.

Interview with CEO Andrew Stanger

Q: The NFT space has proven to be very interdisciplinary with regards to the different artists and operators behind the most exciting projects. What was your path to Reality Cards?

A: The NFT scene is overflowing with cookie cutter projects- Reality Cards was born out of a desire to create something completely different. In particular, I was looking to create a project where it was possible to collect NFTs and win money at the same time. After extensive brainstorming, I arrived at the solution of adding a prediction market on top of an NFT marketplace, and Reality Cards was born.

Q: What do you think that NFT technology provides to prediction markets that the traditional PM platforms don’t have?

A: It makes betting synonymous with ownership- and thus, winnings are determined not by how much you bet, but instead by how long you owned an outcome for. No other prediction market on Earth can make this claim. 

This has a profound implication- there can be a strong incentive to bet in an NFT-prediction market before there is even any money to be won, an incentive that does not exist on any normal prediction market. This is because the start of an event- where there is zero pot- is when seconds of ownership of the NFTs can be acquired at the lowest price. 

Thus, there is an incentive for liquidity to be added immediately, without providing fees to liquidity providers. This is not an incentive that exists in any other prediction market.

Q: How can someone from the prediction market space map common concepts to this new project? Are daily rental prices reflective of the “odds” of an outcome? How does the lack of “shares” and “bids” affect market dynamics for Reality Cards?

A: Great question. Yes, daily rental prices are indeed reflective of the odds at that time. For example, if the prices of outcome A are $25/hour, and the prices of outcome B are $75 an hour, this reflects 25% odds of outcome A occurring and 75% of outcome B occurring (assuming only two outcomes). This information is presented in the UI as 'implied odds'.

However, it is granted that it is not easy to map onto common concepts easily, because it is not as simple as renting a Card where the implied odds are below what you believe to be accurate. Instead, the user must take into account the average rental prices across all outcomes to date. To be specific- a user should rent an outcome if and only if the current rental price is below the user's expected odds of that event occurring, multiplied by the average sum of rentals of all Cards. This sounds confusing, and it is- we are actively working on our UI to make this easier to digest, so that users are able to see at a glance whether an outcome is worth renting or not.

However, there is a flipside here, which is that if the current sum of rental prices is ever below the average sum of rental prices, a user will always profit from renting all outcomes. This is never the case in other prediction markets. Thus, in certain instances, Reality Cards offers the user the chance to make risk-free winnings. 

Of course, this describes the mechanics of trying to win money- if your goal is to collect the NFT, it is much simpler- the user must simply own the Card longer than anyone else.

Q: Many prediction markets and platforms grapple with the question of purpose. The information gained from prediction markets can have a bevy of uses, from supporting social goods to generating investment alpha. Does Reality Cards view itself and its goals in any particular way? And is that reflected in the events which the company covers?

I view Reality Cards not primarily as a prediction market, but more as an NFT marketplace. Thus, I do not consider the extent to which it is making predictions to be paramount. Instead, it is a platform where artists can drop art that relates to a real world event. This is what defines our events- the events that NFT artists wish to commemorate through their artwork.

Check out Reality Cards’ development roadmap on their website, and let them know Crowd Money sent you if you check them out!

Prediction Plays of the Month

Summary of GG Portfolio / Plays

The Road to Glory continues this week, although with some unfortunate news: the first loss of our journey. Last month we bet on the Kalshi market on whether new home sales would exceed 700k  in the month of August. We made a position on No with a few hundred contracts at a cost basis of ~30 cents. The results came in on September 24th and there were 740k new home sales in the month.

Life has been busy and the results are recent, so we need to dissect what went wrong with this market. We’ll make sure to update you in next week’s supplement. But here are our initial thoughts.

First, the month lag between current news and these Kalshi markets can obscure observations. What we were seeing and downward pressure on the housing market might have kicked into effect in the month of September, but those dynamics won’t flow through until October. So we may have had accurate, albeit early, insights on the market. 

Second, it is also possible that our observations from this past month were simply wrong. For example, perhaps our theory that fewer homes would be bought in August, irrespective of macro trends, due to school starts and school enrollment requirements was incorrect. This might be a dynamic that is only significant from September onward. It is also possible that the eviction moratorium ending had more impact on Augusts’ home sales numbers than we anticipated. Other factors that influenced the outcome include July’s sales numbers being revised up to 729K from 708K units.

This four month new home sales high indicates a stabilizing of housing market dynamics, with Covid-19 related disturbances potentially alleviating. Next month’s housing market will be interesting to watch, but it likely won’t be the play we discuss!

Starting Balance 214.62

Traded: 167.67

New Balance (😉): $42.42

We may inject some more capital into this experiment next month, and make more plays each month, in order to speed things up. Plays from the Crowd

Plays from the Crowd

This section is one that we are very excited about. If you are a subscriber (Crowd member), you will gain access to several perks which will develop as this newsletter grows. We are hyped that one of those perks will come into effect next week. 

If you want to start making plays, Polymarket is a great place to start! Use referral code ‘CrowdMoney’ here to get up to $100 reimbursed on your first day of trading!


If you want to submit plays and you want to have more of a conversation about it, come find us in the Prediction Markets Discord server! We like to discuss plays there and get feedback on our content.

  • You can also submit plays in the comment sections or on Twitter.

We are also in partnership with the Young Professionals in Foreign Policy organization, a global group focused on developing the next crop of foreign policy leaders. Check them out and sign up if you want to network with like-minded individuals around the world!

Most importantly, we are introducing a new Crowd Money referral program! Share the newsletter with your friends to help us grow the crowd and unlock cool perks when you get 5, 10, or 15 (!!!) friends to sign up.


We really appreciate your support in growing this newsletter and every referral counts.

Finally, don’t forget to come back to next week to catch the hottest new markets for early October and our recap of the biggest prediction market news stories of September. More monthly content for the same low price of free.

And a reminder,  if there are topics you’re interested in, you have feedback, or just want to say hi, you can follow us at the below:

Share 💰 Crowd Money: Exploring Prediction Markets