Thinking about kalshi, polymarket, spreads, predictions, etc.
https://wifpr.wharton.upenn.edu/blog/a-primer-on-prediction-markets/
Ai summary
Here is a summary of the Wharton blog post **"A Primer on Prediction Markets"** by Victoria Liu and Djavaneh Bierwirth:
Prediction markets—platforms where users bet on the outcomes of specific events—have gained massive traction recently, notably through sites like Polymarket and Kalshi during the 2024 Presidential Election. The article breaks down why these markets are so effective and what they mean for the future of finance.
### How Prediction Markets Work
The core strength of prediction markets is that they financially incentivize people to tell the truth, do their research, and crowdsource information. To function effectively, they rely on a few specific mechanisms:
* **Contract Types:**
* *Winner-take-all:* Predicts the probability of an event happening (the most common type).
* *Index contracts:* Predicts the mean of a continuum of possible outcomes.
* *Spread betting:* Predicts the median outcome based on a chosen cutoff.
* **Market Design:** Modern markets often use **Automated Market Makers (AMMs)** rather than traditional manual ledgers, acting as the universal counterparty to trades. They also frequently use a *pari-mutuel system*, where all bets are pooled and split among the winners, which helps boost liquidity and participation.
* **Key Requirements:** For a market to be accurate, it needs clear definitions, agreed-upon data sources to verify the outcome, good incentives, and safeguards against spam and manipulation.
### Implications for the Finance Industry
Historically, prediction markets were niche—like the Iowa Electronic Market launched in 1988 for political betting. Today, platforms like Polymarket handle billions of dollars in volume because they reduce friction with real-time analytics and easy-to-use interfaces.
For the finance industry, these markets act as a new type of derivative. Instead of hedging against price changes (like traditional derivatives), investors can hedge against **specific event-based risks** (e.g., policy changes, cabinet appointments, or specific dates).
### Policy and Future Outlook
As these platforms become more liquid and popular, they face some challenges:
* **"Noise Bettors":** A flood of uninformed retail bettors can decrease market efficiency, meaning platforms need to ensure they are adequately incentivizing *informed* traders.
* **Regulatory Concerns:** Policymakers have to balance innovation with consumer protection. Because prediction markets closely parallel gambling, regulators are asking hard questions about consumer manipulation and how these markets might influence real-world events.
Ultimately, the authors conclude that as prediction markets continue to evolve and become more liquid, they will offer powerful new utilities for both synthesizing public opinion and advanced risk hedging.
In a traditional financial market, if you want to buy a share, you have to find someone willing to sell it to you at that exact price. This is called an **order book** system.
In prediction markets, relying on an order book is a massive problem. Because people are betting on thousands of very niche, specific events (like "Will the temperature in London exceed 80°F on Tuesday?"), there often aren't enough buyers and sellers online at the exact same time to match up. If you try to bet, you might be standing in an "empty room."
An **Automated Market Maker (AMM)** solves the empty room problem by replacing the human counterparty with a robot—specifically, a mathematical algorithm running on a smart contract.
Here is exactly how that algorithm keeps the market moving:
### 1. The Liquidity Pool
Instead of matching you with another trader, an AMM trades against a "pool" of money. People (called Liquidity Providers) deposit funds into this pool in exchange for a small cut of the trading fees.
* Think of the AMM as a vending machine loaded with equal amounts of "Yes" and "No" shares for a specific event.
### 2. The Algorithmic Pricing (The Math)
The AMM uses a fixed mathematical formula to determine the price of a share at any given moment based entirely on supply and demand.
While the formulas can be complex—a common one in prediction markets is the **Logarithmic Market Scoring Rule (LMSR)**—the basic concept is always the same: **price increases as scarcity increases.**
### 3. The Teeter-Totter Effect
Imagine a market asking, "Will the US land humans on Mars by 2030?"
* At launch, the AMM sets both "Yes" and "No" shares at $0.50 (implying a 50% probability).
* A surge of optimism hits the news, and traders rush to the AMM to buy "Yes" shares.
* The AMM obediently sells them the "Yes" shares from its pool, but because its algorithm dictates that it must balance its risk, it automatically raises the price of the next "Yes" share to $0.55, then $0.60, then $0.75.
* In a mirror reaction, it drops the price of the "No" shares to $0.45, then $0.40, then $0.25 to entice traders to take the other side of the bet.
### Why this is a superpower for Prediction Markets
* **Always Open:** You can buy or sell instantly, 24/7, regardless of whether anyone else cares about the event. The AMM will always give you a price.
* **Instant Probabilities:** Because the AMM dynamically adjusts prices based on the volume of bets, the resulting price perfectly mirrors the "crowd's" perceived probability of the event occurring. If the AMM is currently pricing "Yes" at $0.80, the market fundamentally believes there is an 80% chance it happens.
> **The Catch — "Slippage":** If you try to place a massive bet in an AMM pool that doesn't have much money in it, your own trade will drastically warp the supply, causing the algorithm to aggressively hike the price while you are buying.
—-
How do Liquidity Providers make money in an AMM, and what is impermanent loss?
—-
This is one of the most notoriously confusing concepts in decentralized finance, but it makes a lot of sense once you break down how the math of the pool actually works.
Here is how Liquidity Providers (LPs) get paid, and the invisible risk they take on to do it.
### How LPs Make Money
When you lock your money into an AMM liquidity pool (like putting equal parts ETH and USDC into a smart contract), you are providing the inventory that other traders buy and sell against. In exchange, you earn yield in two ways:
1. **Trading Fees:** Every time a trader swaps tokens in your pool, the AMM charges a small fee (often around 0.3%). This fee is deposited directly back into the pool. Since you own a percentage of that pool, your underlying share of the assets grows slightly with every single trade.
2. **Liquidity Mining (Rewards):** To attract capital, many platforms will give LPs "bonus" tokens just for keeping their money in the pool.
If the pool gets a massive amount of trading volume, those 0.3% fees add up incredibly fast, resulting in high returns for the LPs. But there is a catch.
### What is Impermanent Loss?
**Impermanent loss** is the difference in value between putting your tokens into an AMM pool versus simply holding those same tokens in a standard crypto wallet.
When token prices move up or down, the AMM algorithm forces the pool to rebalance. Because of how this rebalancing works, you almost always end up with less total value than if you had just done nothing.
#### The "Arb" Drain (Why it happens)
To understand why you lose money, you have to look at the formula powering the AMM. The most common formula is **X * Y = K** (the quantity of Token A multiplied by Token B must always equal a constant number).
The AMM is completely blind to the outside world. It doesn't check Coinbase to see what the price of ETH is; it only knows the ratio of tokens inside its own pool.
Here is what happens when prices change:
* **The Setup:** You deposit 1 ETH and 1,000 USDC into a pool. (ETH is currently worth $1,000). Your total value is **$2,000**.
* **The Market Moves:** Suddenly, a massive bull run happens on external exchanges, and ETH rockets to **$4,000**.
* **The Arbitrage:** The AMM doesn't know ETH is now $4,000. It is still pricing it at $1,000. Wall Street bots and arbitrage traders immediately notice this. They rush to your pool and buy up all the "cheap" ETH, draining ETH from the pool and dumping USDC into it until the pool's ratio matches the external $4,000 price.
* **The Aftermath:** Because of the X * Y = K formula, your share of the pool has shifted. You now have **0.5 ETH** and **2,000 USDC**.
Let's do the final math:
* Value of your tokens in the pool: (0.5 ETH @ $4,000) + 2,000 USDC = **$4,000**.
* Value if you had just held them in your wallet: (1 ETH @ $4,000) + 1,000 USDC = **$5,000**.
You missed out on $1,000 of profit. That $1,000 difference is your **impermanent loss**.
### Why is it called "Impermanent"?
It's impermanent because the loss is only realized on paper. If the price of ETH crashes back down to exactly $1,000, the arbitrageurs will trade the pool back to its original ratio, and your loss drops to zero.
It only becomes a **permanent** loss if you decide to withdraw your money from the pool while the prices are different from the day you deposited.
Ultimately, Liquidity Providers are making a calculated bet: they are hoping that the trading fees they collect over time will be higher than the impermanent loss they suffer from price swings.
This is one of the trickiest ideas in crypto, so let's strip away all the complicated math and imagine a magical vending machine.
### 1. How you make money: The Toll Booth
Imagine a vending machine that trades Ethereum (ETH) for US Dollars (USDC). It’s not owned by a company; it’s run by a robot (computer code).
For the machine to work, it needs inventory. So, you (the Liquidity Provider) step up and fill the machine with your own money—let’s say 1 ETH and $1,000.
In exchange for letting the machine use your money, you get to act like a toll booth. Every time someone uses the machine to swap tokens, the robot charges them a tiny fee. That fee goes straight to you. If the machine gets used hundreds of times a day, those tiny fees add up to a nice, passive income.
### 2. The Risk: Impermanent Loss (The Blind Robot)
Here is the catch: **The robot running the machine is completely blind to the outside world.** It doesn’t read the news or check external prices.
Let’s say you loaded the machine when 1 ETH was worth $1,000.
* Next week, ETH becomes wildly popular. In the real world, 1 ETH rockets up to **$4,000**.
* Because the robot is blind, it doesn't know the price went up. It is still trying to sell your ETH for around $1,000.
* Smart, opportunistic traders immediately notice this. They rush to your machine, buy up your "cheap" ETH, and dump their Dollars into the machine to pay for it.
When you finally come back to collect your money, the balance has changed. Because the traders bought up your ETH, you now have mostly Dollars and very little ETH left.
Yes, you have more Dollars than you started with, **but you missed out on big profits.** If you had just kept your 1 ETH safely in your wallet instead of putting it in the machine, it would be worth $4,000. Because the blind robot sold your ETH too cheaply, you ended up with less total value than if you had simply done nothing.
That difference—the money you *would* have had if you just kept it in your pocket—is your **impermanent loss**.
### Why is it called "Impermanent"?
It’s "impermanent" (temporary) because the loss only exists on paper. If the real-world price of ETH crashes right back down to $1,000, traders will use the machine the other way, the inventory will balance back out, and your "loss" magically disappears.
It only becomes a **permanent** loss if you decide to take your money *out* of the machine while the prices are still out of whack. People who provide inventory to these machines are just hoping that the "toll booth" fees they collect will be bigger than any money they miss out on from the price changing.