Put Your Money Where the Truth Is — Andrea López de Vicuña | GoMarket

As you probably announced, I'm going to talk about prediction markets. And in this side we've been listening for a while about which one is going to be the next on-chain killer app. We've talked in the past about NFTs where that were a perfect tool to move ownership. We talked about payments that were the perfect tool to move money without borders. We talked about uh DeFi that was the perfect tool to uh move capital as well but without banks.

And now I want to change the question a little bit. And I want to ask how do we know what's true? And that's what prediction markets are trying to solve on this side. So, what they do is that they change the incentives um compared to the other information platforms that we were using in social media apps. What's what's paid is the engagement in normal media.

Um the one that's paid is the one that owns that media. And here in prediction markets the one that get paid are the ones that are that know the final outcome. The one that are true, like the ones that are correct, not louder or no competition at the end, just true. Excuse me. So, what I want to talk about is that part that they are more than just betting platforms on that side.

Um we've talked a little about this about this information infrastructure that we currently have. Um I don't know if you're used to this situation, but in some places where uh the media is controlled by three countries um and depending on the party that wins the elections the information changes. There you can see that this information is not decentralized, let's say, and it it can be really easy manipulated. So, in this type of of markets is where prediction markets can play a great role in in creating and substituting sub- substituting this information infrastructure that we have. But, they can also work as a hedging instrument.

Let's think about this use case about a football team that um is about to get disqualified for the Premier League. So, on this part is not only assuming that it's going to be heartbroken after the the match, it's going to be assuming that it's going to have a really great financial burden. So, for this they can use a prediction market just to bet yes on its own relegation. That will mean that it will act as an insurance in terms like if they were they were not correct, it's fine. They will assume a little bit of a loss, but that's better than being disqualified for the from the Premier League.

On the other hand, if they actually get relegated, what they will do is that they have some financial insurance to cover up this mess until the next season and until they can compete again. So, what I want to now highlight is this part that this can only happen if we have these prediction markets that are acting in a in a regional scale in a regional basis. What the tools that we've designed already um are only are only efficient in global markets and cover these two particular use cases. Great. The US presidential election was one of the biggest markets that happened in Polymarket, same as the major sports final.

What happened now with the World Cup? It moved a lot of volume, a lot of money. But, it's not covering properly and we'll see some really easy to spot mistakes when we are trying to target these regional markets because we're going to find markets that are completely empty or that no market maker wants to put their money on those. So, that's what I want to bring up this this mismatch that we are seeing. So, mainly the usual markets that work are for English speakers, they have to be global, and they have to be big.

Well, at the end, this is not covering the three biggest questions here. Like, where's the volume? Where's the liquidity? And where is the interest on that markets? So, sometimes we start thinking about the deep um work system or the mechanism behind the the prediction markets, which I'm going to explain it briefly, but what I want to highlight again is that this is not actually the engineering problem that we have to solve.

Like, um either having an AMM bonding curve or a club won't change these results of having um these markets that looked completely empty. So, the difference with between these two things are basically that for an automatic market maker, you always have a price. You always have someone that's betting against you because that's literally that bonding curve. That's what it's doing. And on the other side for the central limit order book, the difference is that e- that unless you have real trades, real bets, there's no price to go against to.

So, in our case, cuz I'm here representing a prediction market called Gnosis Market, we are transitioning from an automatic market maker to a club. And we are particularly doing this because of capital efficiency. At the end, where when you have an automatic market maker, what it means is that you have to um deposit big amounts of liquidity that are completely inefficient to protect against a bets in the other side. So, for that um the main situation is that the probability also uh moves really fast and really inefficiently, let's say. That's why those spreads are completely useless.

If a user bets an amount of $50 and the amount seated in that market is low, that's going to alter the market in a tremendous amount. What doesn't happen with a club mechanism. Anyway, the cause that we are facing is that market makers are not attracted enough about these regional markets because they don't have enough um incentives to participate on them. So, what I was saying, the diagnosis is clear. It's just dead markets.

And that's going to happen because of three different symptoms. Either you don't have liquidity, that means these market makers are not attracted, they don't have any type of interest in investing on you. There's no volume, even if you can bring these market makers to the platform. There's only better than one side, so it's going to look that the market is not changing correctly depending on what's the final outcome of of it, so it's not going to represent accurately the probabilities. And on the other side, we can simply have these um these not mechanical aspect that is that nobody has interest

[music]

on that particular market. So, how can we fix it? On that part, what we can do is making a little bit of a Frankenstein between these automated market maker and this club system with a bounded loss backstop. What does that mean? That you have for a certain moment, you have it all topped.

That you have an automatic market maker working until some users are really trading in your own. So, that way you limit the the need of having external market makers going into your platform, and you can control your your liquidity stagnation, let's say, also for until the market gets traction. Or another really nice um way to do it is like it happens with some illiquid illiquid assets in in financial markets, that is that instead of like having the markets available to trade during 24 hours, you say like, "Okay, let's put a time frame where you can come in and bid a little bit for a specific amount of time." That way you fix the price, and that makes more sense and makes the price more robust. On the other part, that was the third pillar in the in the two slide two slides before, that was this part of the interest.

That's not a mechanical problem in term of in terms of like how do you design the mechanics of your prediction market, it's more about how do you market it, like literally on that part. So, how do you market these regional these regional markets? Well, depending, you can go to local curators, people on site that they read their news in a regional basis. You can go for time box listing, that means that instead of like having a market going for a long time, you just put a push it a a close to the finish line of a specific event. So, instead of like having it there for 3 months, you have it for 15 days or 24 hours just because before that events that event starts.

And the other part, this last one, is to decentralize the information access, is instead of like going directly to one source of truth, like Reuters, for example, you go with a multilingual partners that known that are sovereign and they know their source source of truth in that specific region. So, on our part, we've solved the primary design of these markets, how do they work, how did do they operate, and this final question is like how do they resolve? What happens when the market ends? That's kind of like the oracle question. And again, this has some problems with the design that I was mentioning because they are mostly based on global markets.

Original markets usually don't have a centralized source reported, or maybe you do, but in that way, you're again centralizing the information in just one source of truth. And on the other one, for the dispute window, imagine that uh with I don't know if you remember this part with the market of Cristiano crying after Portugal was eliminated, if someone doesn't know if he was crying or not, and you open a judge window to literally have people voting if that happened or not happened, you can't have people In this case, yeah, but in a regional market, you don't have people that know literally the truth. You can have people in that governance um governance board in San Francisco talking about what's happening in Belgrade and voting about that. So, what do What are we doing in Gnosis Market to fix this situation? We continue with this optimistic oracle.

We need to be fast in this industry. We don't want people to be waiting to get back their earnings for 24 hours, 2 days. We want it as fast as fast as we know the truth. So, this means that uh the truth, if it comes from an elaborated source of truth, let's say for these crypto markets, we already know that uh CoinGecko piece is not going to change It's not going to lie to me. I'm just going to close the close the market, say yes, and unless someone puts a bond on it and disputes it, I'm not going to think bad about it.

I'm going to believe that that's okay. And the other way to do it is again, um, as I was saying, uh, making markets really close to the final end. But creating these curated governance regional forums. So, in that way, if a market that is regional needs to be disputed, it's only going to be these board of governance very established in those in those sources, as well. Instead of like having only one source of truth, let's say, just calling one API for the for the resolution, we need to have two of those.

So, um, on these parts, the other issue with market makers, that that's the problem we're mainly facing these days, is not that if we convince him if if we convince them to trade on on on our regional markets that they are very thin and maybe they are not attractive to them, is that if they are not sure that that resolution that we are choosing is good enough, they can pull the money back before they just before the market's going to end. And that's the moment where real bettors need that liquidity to get back. So, to avoid that situation, we need to inform them of what's the resolution mechanism of each market that we are deciding in that particular case. Anyway, everything comes with with risks, as it happens in whatever the industry. So, in this case, the forest that that we are mainly facing for prediction markets is like manipulation.

We've seen this in the news a lot. Insider information, like in finance in financial environments. Morbid markets, we cannot have markets talking about when's going to be the assassination of a particular public person, for example. Or, uh, capture arbitrage. So, for these again.

What are we trying to do then? And with this um the presentation is done. Like we need to fix the infrastructure, the microstructure of these regional markets. We have to fix the article to be a really source of truth that um it's real for these a specific um regional markets and we have to fix this distribution with people. We need to find curated list of people that are interested in very small um very small topics.

So, yeah. Uh to conclude on this, like as I was bringing at the beginning, I think that we now should move the question of like what type of um value are we moving to what type of truth are we are we moving actually and what can we believe and in what markets can be based on. So, thank you very much for listening. If you have any other questions, I'm glad to respond them.