Inspiration

What it does

Below is a clear, readable Markdown story you can use for your submission. It follows your instructions, uses LaTeX where helpful, and stays simple for judges.

Project Story

What problem are we solving?

Some items are easy to buy and sell. Most aren’t. If you own a Pokémon card, an in-game item, a concert ticket, or even something like carbon credits, you often don’t have an easy way to turn that item into money. You need to find a buyer, negotiate a price, and wait. This creates huge friction and low liquidity.

Our goal was simple: let anyone create a market where people can instantly buy and sell any item at a fair price.

What inspired us?

A friend mentioned how hard it used to be to sell a Pokémon card. You could have something valuable, but unless you happened to find the right buyer at the right time, your card just sat there. That conversation stuck with me. Lots of items in games, online communities, or real life have the same problem. They have value, but no liquidity.

We wanted to fix that.

How are we solving it?

We built a system where someone can spin up a “merchant” for any item. That merchant:

Buys items from users immediately
Sells items back to users immediately
Sets prices using an algorithm that reacts to supply and demand
Charges a small spread, just like a real market maker

When many traders come and go, the algorithm adjusts prices in real time. If more people buy, price goes up. If more people sell, price goes down. At its core, the system tries to balance inventory and keep spreads tight.

We use a simple model:

[ p_{\text{mid,new}} = p_{\text{ref}} - \lambda\left(\frac{\text{stock} - \text{target}}{\text{target}}\right)p_{\text{ref}}, ]

so low stock raises prices and high stock lowers them.
It then computes a spread based on how far inventory is from the target, and sets

[ p_{\text{buy}} = p_{\text{mid}} + \frac{s}{2}, \qquad p_{\text{sell}} = p_{\text{mid}} - \frac{s}{2}. ]

where (\alpha) controls how strongly prices react to trade flow.

This gives users a fair price while allowing the merchant to manage risk.

How we built it

Before writing any code, we spent around six hours designing the full interface: every class, every function, every plugin boundary, and every data path. This planning made the system much easier to implement.

The tech stack:

Python + FastAPI for the backend
React for the frontend
PostgreSQL for storage
Mineflayer for our Minecraft plugin
A plugin architecture for connecting to games and real-world items
A pricing engine that adjusts quotes based on trade flow

We also did a lot of data-science work on the algorithm. We tested different update rules, looked at stability, and checked how sensitive the price is to bursts of trades.

How the system works

A market creator chooses an item.
The system sets up a merchant for that item.
Users can now instantly buy or sell from that merchant.
A pricing algorithm updates the buy and sell prices after each trade.
Plugins move items in and out of external platforms (games, storage systems, etc.).

For example, in Minecraft, our bot network holds real in-game items. When someone buys, the bot delivers the item. When someone sells, the bot retrieves it and adds it to storage.