This Trade Could MINT (Patreon)

Is any one of you an incredibly handsome East Asian or Indian dude (preferably with some kind of acting background but not necessary) who wants to be in a short video? 

Can't promise it won't involve a black leather couch, but can promise you will be compensated (decently) well. 

Adding this edit here so no one who might be seriously interested is scared off: 

The video is "The Guy on the Other Side of Your Trade"; a parody of those "Most Interesting Man in The World" Dos Equis commercials. 

And yes I know Asia is a country in India I'm not that racist. 

In all seriousness, please DM or email me at benjaminfinanceyt@gmail.com if you're interested! I think it will be a fun project. 

H8ers will say this is photoshopped. 

I had a dry January experiment that turned into dry for the rest of the winter experiment that might turn into dry for the foreseeable future experiment.  

Positives 

• More time and money
• Less embarrassment and pain

Negatives 

• Collapse of social life
• Hot take: my brain actually hasn't really gotten much better. I'm not noticeably more focused, and there are still complex, usually trading related, problems where I still feel stupid. Maybe that's just how I am lol.  

However it would be a crime to not drink at my cousin's wedding, so that might be the end. I have a few more months. 

The Lottery Trade

One of the observations I've had while gambling on Polymarket is that shares tend to be "sticky" around very high prices. 

What I mean is that they don't seem to converge to 1 fast enough. You'll see shares sitting in the 95-99.9 range when the probability of an outcome is definitely greater than that.   

For example, I took this screenshot yesterday at around 11PM eastern. 

At this point, the probability of a No resolution was certain, but market makers were still offering these 99.9c no shares. You could buy them and get a risk free $6 on $6k. 

That may not seem impressive, but when you annualize this figure it turns out to over 30% returns. This also is far from the best opportunity I've seen. 

I think there are some behavioral biases that contribute to shares not converging to their yes or no outcome fast enough.

There are also structural reasons why market makers will sell you these. We'll get into those in a minute. 

Why do the Trade Exist? 

Reason 1: Premium for Holding Risky Stuff

I think in most markets, any lottery style, low-likelihood thing, (could be insurance, OTM options, etc) is usually overpriced. 

Who would sell a deep OTM option they thought was fairly priced? 

With some highly improbable thing, it's really difficult to say statistically how likely it is to happen. 

A lot more things can happen than have happened, and the seller is taking on a shitty, negatively convex payoff profile in exchange for a small fixed profit. 

Because of that, sellers want a risk premium. A premium above a guess of the fair price that compensates them for the fact that we don't know wtf is gonna happen. 

You can see this across a ton of different assets that are far more efficient and liquid than bets on Polymarket, which makes me think it's probably true on Polymarket.

We can see this in the historical data as well. 

The graph below is from a paper I read talking about Polymarket. It's a really good read if your tryna trade prediction markets in a big way. 

This graph basically buckets prices of no shares on Polymarket, and then compares them to the fraction resolved to a "no" outcome. 

Put simply, it's a way of visualizing whether certain ranges of shares (0-10c, 10-20c, etc) are over/underprice on average. 

The dashed, diagonal line represents where each bucket would sit it prices were perfectly efficient. 

We can see that on average, NO prices are too cheap. 

Events on Polymarket resolve to NO more often than prices imply (as we can see in the orange lines sitting above the diagonal dashed one). 

This just means if you got jacked to the tits in NO shares across a bunch of independent markets, that would have made money historically. 

Reason 2: Exchange Incentives

This bet also makes sense intuitively with how market makers are incentivized to trade on the platform. 

Polymarket pays 50-70k USDC/week to liquidity providers. 

"Liquidity provider" is just a fancy name for any mfer that places resting orders in the book. The biggest share of these rewards goes to the platform's top market makers.  

When the midpoint of a market is between 10-90c, you can earn rewards by placing one-sided resting orders. Just offering to buy or sell a given contract. 

When the midpoint is between 1-10 or 90-100, in order to capture rewards you have to place two-sided orders; a bid and ask.  

This is different than a platform like Kalshi (or even the US equity option market) where you'll notice very improbable events (or deep OTM options) won't have a bid. Market makers will sell them to you, but won't buy them from you, because the probability of them making money is incredibly low. 

Also, unlike other exchanges that might have volume rebates for liquidity providers, Polymarket pays rewards based on the amount of time your orders spend resting in the book. 

This means you don't need to actually trade with anyone to get these rewards, you just need to quote a competitive size near the midpoint. 

I'm guessing market makers have models that tell them when the amount they'll make in rewards is greater on a market is greater than the amount they'll lose when people buy these contracts. 

Given traders' preference for lottery-style payoffs, these super risky low-return contracts rarely actually trade. 

Just monitoring the trades on the platform, you'll notice once the midpoint in a market hits 95c+, trading volume slows down. 

All of this to say, we can see historically based on the paper above that this trade has made money, and I think there's a really strong economic rationale for why it would still exist. 

For what it's worth, I have also been trading it and making money. Admittedly, this doesn't really mean anything when you realize how many observations you need to prove it's "real" statistically, but hopefully it makes you feel better. 

How do I prevent implosion?  

This trade involves buying a bunch of shares that payout dollars when they win, and lose hundreds when they lose -> the definition of picking up penis in front of a steam roller. 

I figured the best way to prevent implosion would be to do some sims. 

We start with a $100 bankroll, and assume the bankroll is divided across 20 independent bets.

In each trade, we have a 96% chance of making 5 cents, and a 4% chance of losing 95 cents, giving our trade an expected value of 1 cent per share.

We trade every day, every 3 days, every 5 days, every week, or every two weeks.

• "We start with a $100 bankroll, and assume the bankroll is divided across 20 independent bets."

This 20 number is arbitrary, and far from the mathematically optimal betting fraction of your bankroll to maximize growth. 

If you input the following stuff into a Kelly calculator, you'll find the optimal bet fraction is likely going to be MUCH larger than you're comfortable with. 

I found this 20 number by simulating a bunch of stuff and going "what amount of diversification is realistic given the number of independent bets on Polymarket? And what makes the median outcome in our simulation look decent?"

We also don't know if our edge going forward will be 1c/share (or even positive). These sims are based on the empirical results of that paper, and don't really mean anything above what will happen in the future.

However, this is just a cruel reality of the market, and I think the economic rationale here is strong enough that I'd make this gamble. 

The conclusion: being that your edge is a small one, you want to do as many independent bets as possible to get it. 

In blue we can see the median path of each. 

I'd probably keep from ever putting more than 1/10th of my money into 1 thing, but that's me. 

In practice, being that you're constrained by the number of  liquid markets on Polymarket, you may want to yolo it and diversify less. It depends on if your money is being used elsewhere. 

Other Tips for Doing This Trade

• The best place to do this trade is in one-off, weird, markets that aren't standardized. Market makers in these markets don't have external data sources to get a fair value; they're likely just going off what the flow tells them. This means those markets are far more susceptible to the kind of biases we're trying to exploit here.  
• At the end of the month, a lot of trades are closing. This is a good time to look for expensive shares to buy. You can filter by "Closing Soon" on Polymarket. 

Video is coming I just added a ton of stuff so it's taking a minute but I think it's gonna be a good one. 

Portfolio Update

I have a small momentum trade in ETH right now. I'm also about to do all of my end-of-the-month trades. 

I'm wondering if there's some pair trade you could do between SMCI and NVDA? Anyone got any ideas? I have no idea on a value basis which is better.