Selling something you don't own is really a loan. Every odd thing about shorting comes from that.
To short a share you borrow it from somebody who owns it, then sell it straight away at today's price. Now you're holding cash, and you owe somebody a share.
To finish, you buy the share back and hand it over. If the price dropped in between, you buy it back cheaper than you sold it and keep the difference.
Here's the detail everything else comes from. You owe a share. You don't owe cash. Your debt is in something whose price moves.
Buy a share and the worst case is the company goes bust and you lose what you put in. Your loss has a floor, because a price can't go below zero.
Short a share and your loss grows as the price climbs. Nothing stops a price climbing. A share that triples costs you twice what you collected. That's the big risk of shorting, and it's why a short that's going wrong gets more dangerous the longer it goes wrong.
It snowballs, too. As the price rises the bet gets bigger next to your account, so the same percentage move hurts more each time. A normal holding that falls quietly shrinks its own importance. A short that rises grows it.
Borrowed shares aren't free. Whoever lent them charges you a borrow rate, quoted as a yearly percentage and clocked up every day you hold. For a big ordinary company it's small. For a company lots of people want to short and few people own, it can be brutal. There are only so many shares to lend.
You also put down a deposit, called margin, in case the price rises. If the bet moves far enough against you, somebody asks for more or closes it for you. That demand is what "margin call" means, and it's what turns a loss on paper into a forced sale at the worst possible moment.
Margin Call charges the real borrow rate and the real margin on every short. It shows you both on the ticket up front, before you confirm.
A heavily shorted share starts rising. Some shorts are forced to close. Closing a short means buying. That buying pushes the price up again, which forces more shorts to close, which is more buying.
Nothing about the company needs to have changed. The buying is mechanical, and nobody involved learned a thing about the business. So the price overshoots hard, then sags back once the last forced buyer is done.
The number that measures the danger is days-to-cover. Take the shares sold short, divide by how many trade on a normal day. Roughly, that's how many days every short would need to get out. A heavily shorted share that trades huge volume can cope. One that barely trades can't.
Mostly as a hedge. If you like a whole sector but think one company in it is weak, you can hold the sector and short that one company. You're left with the view you actually have.
Shorting on its own, as a straight bet that something will drop, asks you to be right about the company and right about the timing. And the tool punishes being early even harder than being wrong.
All of that describes how the mechanism works. None of it is a recommendation, and nothing here is financial advice. In Margin Call the shares, the borrowing and the losses are all imaginary, which is the point of practising with them.