Scaling Prediction Markets | Henry Lau, Jack Simison, and Lionel Williams

Hello everybody. Welcome to the prediction market topic here. So right now we're going to have a exciting panel. We're going to have Oh, my name is Lucky by the way. If you need anything, I'll be around for the next 3 hours with you guys.

So we have Lionel Williams from Light Node Ventures. That will be He will be our moderator. And then we have a Henry Liu from Sky Walk. And then we have a Jack Simpson from Nascent. Welcome guys.

You have your microphone. Perfect.

we're good. You guys can hear me? We're good? Okay, we'll take it. How you guys doing?

Welcome to ETHDenver. Welcome to the prediction markets panel. We're going to get a little spicy here today. I'll be your moderator Lionel Williams, the vice president over at Light Node Ventures. We're strategic growth partners for VCs as well as projects helping pre-seed to series A projects as well as VCs with their portfolio companies introducing them to LPs and also family funds.

I'll pass it on to Henry. I'm Henry Liu. I'm the founder of Sky Walk. We're a liquidity layer for prediction market assets. Um just this morning we announced that we for the first time brought Polymarket assets over to Ethereum mainnet.

Um and also to base for trading. Everyone, I'm Jack and I'm with Nascent. I'm an investor there. Uh we focus on both venture and also liquid strategies across the board within crypto. Got you.

So let's jump into it, guys. So prediction markets are now moving from like a niche speculation more into real financial infrastructure. All right. So the themes will be just like trusted resolutions, durability, liquidity, and also real-world utility beyond speculation. So let's like unpack these themes.

I'll start off with you, Henry. What has to be true for prediction markets to become systematically important not just interesting? Because I think we're in a time where people just think, "Oh, this is just interesting." But like what's the systematically aspect systematic aspect of it? Yeah, I think the thing that makes them interesting is when prediction market outcomes affect more than the people who are just trading them.

Got you. And so we we see that as an information source. I you know, I have friends who monitor some of the geopolitical prediction markets. They actually set up alerts when prices breach certain levels cuz that might just be an early warning sign for friends and family back home wherever they might be. I think that's a very useful tool.

And then on an economic sense, having these assets of course also help people end users who might not even be using prediction markets themselves through hedging or other techniques. That would be great, too. Absolutely. You want to hop in, Jack? Yeah, happy to.

I think for prediction markets to become more systemic to the overall macro structure of investment markets globally, one, we'll need a lot more liquidity. And obviously that's determined by basically better resolution, better oracles, and then also institutions just overall becoming more comfortable with prediction markets in general. We're seeing a little bit of that with the high-frequency trading firms like Jump Crypto, DRW, etc. etc. especially with Jump taking equity stakes in both Kelp She and Polymarket as of late.

Nice. Nice. That's a great transition into the next topic of liquidity. What liquidity for Henry? I'll start off with you, Henry.

So why is liquidity the single most important variable of prediction markets? And then also is there a bare minimum liquidity threshold where a market becomes like more so informative? Yeah. Um for Sky Walk, we only focus on markets that are trending towards a hundred million dollars in volume or more over the lifetime of that market. So if you look at over the 30,000 events that are traded, I would say 29,000 plus of them do not qualify.

And so our key launch market right now is around the US midterms. We know based off of past political cycles, you know, this will be a very well-traded market by the time it resolves. Um Liquidity is really important because especially for prediction markets, you you you need a counterparty on the other side of the trade. Um and more importantly in the context of DeFi, you need atomic liquidity. So you need some sort of presence there to make it work for the rest of the ecosystem.

A lot of DeFi doesn't work without atomic liquidity. Got you. Got you. And I'm going to shoot it over to Jack here. How do platform designs and microstructure decisions influence liquidity dynamics?

Like especially coming from your your neck of the woods as far as being a VC and being able to understand the ins and outs of the of the actual like details of it. So love to hear your thoughts. Yeah, absolutely. When it comes to liquidity especially with high-frequency trading or anything of the like, depth isn't as important in comparison to big institutional firms coming in to hedge their exposures off platform either in off-chain completely or on-chain and just other investments that they're exposed to. Um but again, going back to high-frequency trading, liquidity doesn't matter as much still matters tremendously.

If you're exiting and entering positions with size, obviously you can't do that with Polymarket or Kelp She today unless you're betting on Super Bowl markets within Kelp She. Um but specifically for the microstructures involved and for hedge funds to basically involve themselves more on-chain with Polymarket or off-chain with with Kelp She, um liquidity's going to have to just continuously increase. Got you. Got you. And so like once liquidity is flowing, technology, AI bots, execution strategies starts to kind of like shape who wins and who loses, right?

So I'm going to stick with Jack. AI tools have dramatically lowered the barrier to building more so infrastructure and trading bots. How has it changed the market microstructure? Yeah. So as you guys probably all know here, Claude 4.

6 and 4.5 that it was launched in January has given a lot of new people a lot of confidence that they can build arbitrage bots to compete with that of the likes of Jump Crypto or the others. And so what we're seeing is we're seeing a lot of potentially uninformed flow or less educated flow that's coming onto these markets and actively deploying their capital. Um and so within that, if you actually zoom into that data, you're seeing 90% of all of these quote-unquote arb bots that aren't really arb just taking uninformed risk losing, right? And that will get better over time.

Obviously as the foundational models improve, the agents that come out of those will also improve and will be able to see a lot more liquidity coming out from a bunch of different strategies that is going to create some unique dynamics within the markets themselves. However, I do see that continuing. I do see that impacting basically everything on the micro front. Uh but yeah. Got you.

And I'm going to go over to you, Henry. All right. So Polymarket charges fees on 15-minute crypto markets but not event markets. So what does that signal about where like the real revenue lies? You know, it's an interesting dynamic, right?

Because Polymarket has a regulated presence now and a non-regulated one. And there's an interesting game theory outcome as to whether they end up matching Kelp She's fees across the board or whether it's a fee war between the two. So I think turning on the fee switch on-chain has been interesting. I think they're on track to do like a good chunk of revenue off of these new markets already. Um but also I should note that a lot of the fees are similar to Kelp She or being used to incentivize liquidity providers and it's adding more depth to the book through rebates that market makers are getting back.

Got you. So we're talking about two different models here. So for Jack, different platforms use like fundamentally different resolution models as we're seeing. Decentralized optimistic oracles versus centralized resolution. What are the trade-offs in both of those?

Yeah, happy to speak to the resolution the differences between centralized and decentralized approaches. But going back to the 15-minute markets, I do view that as a highly specific product that Polymarket really discovered recently and has been deploying across chains. And they're also being able to charge on it because most people are just accessing it through APIs and their bot approaches. And they see that as All right, we kind of want to disincentivize this behavior but we also want to profit off of this behavior. And so they've added those fees.

Which I think is again a great approach. It's like a bot tax essentially.

Yeah, it's essentially a bot tax. Exactly.

There we go. Um but yeah, so going back to your question, when it comes to decentralized oracles like that of UMA for Polymarket, we have seen in the past when Polymarket wasn't as common, UMA basically be taken over by bad actors that have enough voting power to change the actual outcome or the actual resolution process for a given market. And so I think that's a key risk that a lot of people don't really understand when they're trading with Polymarket specifically. They don't understand UMA. They don't know how the voting mechanisms work and they don't know how the repeal process works.

So like just as a general recommendation to everyone that's getting into Polymarket trading, please take some time, understand UMA, and understand how it works, and understand the risks that you are exposing yourself to. And then when it comes to resolution, as we've all probably seen, there's been a lot of drama across a bunch of different markets within Polymarket and Kelp She when it comes to their specific wording on the resolution of the actual market. And so we've seen specific markets basically yield 100, yield zero, etc. when it comes to outcome based on some confusion between the actual investors investing in the market and the market determined outcome by Kalshi or by Polymarket. Like a a famous one recently was when GTA 6 delayed their outcome to next year, there was a market that would basically compared will Jesus return before before GTA 6 is launched, right?

That's a crazy market. Obviously, you're only seeing that in Polymarket. But it was a key example of how resolution would resolve to 50/50 instead of zero or 100 just based on the event not actually happening in time. And so another general piece of advice to everyone getting involved in prediction markets specifically is understand every single line item in that resolution before taking part in any trades.

Amazing. And I kind of want to speed it up just here a little bit. I know you talked about the Super Bowl just a tad bit. So let's get into Cardi B Super Bowl market. I know you guys have all probably watched a little bit of the Super Bowl in some aspect, but how did like the two platforms reach opposite outcomes per se?

So Yeah, I think that's just another great example comparing Kalshi and Polymarket. Kalshi ended up deciding to just basically resolve the market at the price the market was at when the game ended, aka around 74 cents or 76 cents. They decided to hold that resolution because they couldn't determine if Cardi B actually performed during the Super Bowl or she didn't. Whereas Polymarket decided that it wasn't technically a performance. And I may have that backwards, but I believe it wasn't technically a performance.

And so that's like a big frustrating factor that you can see. We're still seeing and we're probably likely going to see in the future until you know, more structure comes the actual resolution process. Got you. So institutional adoption depends on trust, not just the volume. And to kind of transfer that into like beyond like the gambling and the insurance realm, I'm going to you Henry here.

Can prediction markets evolve from speculation into hedging infrastructure? Yeah, they absolutely can. But I think one of the things that needs to happen first is that these assets need to be standardized and they need to just fit into the existing workflows that these institutions intend to work with. So expecting institutions to customize entirely new workflows just to service, you know, a few markets I think is is a very tough thing to pull off versus something that fits hand in glove. And it's it's kind of interesting cuz you can start imagining insurance-like entities or hedging-like entities wrapping around these assets, especially if they're liquid enough.

I'll give you a prime example. I knew quite a lot of people who actually use Kamala Harris to hedge their portfolio and leading up to the election. I'm not going to pass judgment on what the implications of what that means, but I thought that was very interesting and they were able to do it because that market was so liquid. Got you. Got you.

Okay. So kind of to get just kind of like a in practice type of question. What would a stablecoin depeg insurance market look like? You know, that that's a tough one to to kind of price because um in general stablecoin depeg risk should be pretty low. And it's one of those things that where you look at something like how I don't know if next I think Nexus Mutual does a version of this insurance, you know.

It's I I oftentimes think that on-chain insurance is underpriced relative to the risk that's there. But the problem then becomes if the premium just feels too high and it's priced at the actual risk cuz no one wants to buy it. You know, so I think for I think that that market could work in short time frames. Harder to pull off over like the period of year cuz then you have to think about the cost of that money. And I I guess over time, like you were saying, it'll kind of get better in terms of everything.

Jack, do you have anything to add to that or you want me to can I keep on moving on, man? Yeah, just quick comment. Like obviously for hedging markets, especially when it comes to institutional holders, there needs to be you know, magnitude increase in the underlying liquidity available on these specific markets. And for things like Turbine, they are providing exactly what Henry discussed, which is basically hedging for stablecoin depegging. And right now their liquidity is just not where it needs to be for institutions to get involved.

Got you. So I'll leave this for both of you guys. Who controls the rails long term? Is it crypto-native builders, regulated exchanges, or is it more so asset managers that are coming in? I'll I'll I'll go with Jack just to kind of start off and then Henry, I want you to chime in for sure.

Yeah, I would say this has a lot to do with the legal structures that we're going to see come out and probably go after Polymarket and other prediction markets over the next few years. Right now it's kind of a regulatory loophole and I see that closing probably within the next 5 years. And so depending on what that outcome looks like, it'll very much determine who plays the game and who has control. Right? If we're going to see CFTC regulated markets in an institutional account or if we're going to see open access to everyone because it's deemed not gambling by the US.

Got you. Henry, what you got for me? Yeah, I'm looking forward to the DeFi exemption being part of the Clarity Act and seeing how that extends. So yeah, hopefully that's bullish for decentralized builders. Okay.

And last little spicy question, I guess you could say. Vitalik warned us about the unhealthy product market fit. Do you guys agree? I absolutely agree. I I just wrote a blog post about that today.

It's Sky block is answering one of the questions or one of the the tasks that Vitalik is is asking prediction market builders to do, which is to turn these prediction market assets into something that's more than just chasing after dumb money and and taking advantage of that. And to something that can kind of help people hedge their financial risks a lot more. All right. Got you. So we have about 2 minutes here.

We're going to do a rapid fire question theme here. So rules are you have one or two sentences to answer. No hedging and keep it fast and quotable. All right?

[laughter]

So we got the first first question. Centralized versus decentralized, all right? So prediction markets could see billions in capital flow in. Does that favor centralized resolutions or over decentralized systems or decentralized models scale? I'll leave that up to you, Henry.

That's a tough one. Uh I think a centralized regulated resolution can actually still be enforced on chain. Got you. Jack. I think the ability to go after someone in the event of poor resolution is a must for institutions.

So I think centralized resolution approaches are probably going to have more institutional flow. Got you. Are Okay, second question. Are truth in discovery versus narrative, okay? So are prediction markets generally improving truth discovery or are they financializing social and narrative trends?

I'm going with you, Jack, first. Both. Okay. Yeah, both both for me as well. Okay, we'll do next liquidity and herding, all right?

So does more liquidity always lead to better information or does does this amplify herd behavior and reflexivity? Yeah, I can answer that one. Someone just put $70,000 in the will Jesus come back before 2027 market. Polymarket just announced this function where anyone can actually add liquidity to a market. I I I'm not sure the signal is any better.

The odds have jumped to 6 or 7%. I I think it's a bit high, but um You got anything for me? We're good. We're good. We're going to the next one.

Bloomberg versus DraftKings. Are these platforms trending more toward Bloomberg style or is it more so going towards DraftKings style as far as like consumer gambling products? That's a hot topic. I'd love to argue that prediction markets are not gambling in any form. However, that's a hard argument to back.

So yeah. Got you. Got something here? Jack, I I imagine you've seen like 50 well, I guess prediction market trading terminal startups at this point, all right? So clearly like a lot of builders are are focusing on on this problem and it'll be interesting to see how that plays out.

Got you. And kind of to wrap up here, long-term outcomes. Where do you see prediction markets in 5 years? Is it more so a trillion-dollar financial infrastructure or still like a niche crypto product? Yeah, I think prediction markets are going to be huge.

I think there's going to be more and continuous growth in overall liquidity and traders, institutions, etc. etc. And once it get to a point where people can actually hedge their exposure outside of those markets, it'll become a very dependent and systematic thing. Same here. Got you.

Got you. Well, thank you guys. That wraps up the prediction markets panel. Thank you guys and have a great one.