Selling something you do not own is a loan, and every strange thing about shorting follows from that.
To short a stock you borrow shares from someone who owns them, sell those shares at the current price, and now owe the shares rather than the money. To close the position you buy the shares back and return them. If the price fell in between, you buy back cheaper than you sold and keep the difference.
Everything unusual about shorting follows from the fact that you owe shares, not cash. You have a liability denominated in something whose price can move.
When you buy a stock, the worst case is that it goes to zero and you lose what you put in. Your loss is bounded because the price is bounded below.
When you are short, your loss grows as the price rises, and the price is not bounded above. A stock that triples costs you twice what you originally received. This is not a technicality — it is the defining risk, and it is why a short position that is going wrong gets more dangerous rather than less as it goes.
It compounds: as the price rises, the position gets larger relative to your account, so the same percentage move costs more each time. A long position that falls shrinks its own influence. A short position that rises grows it.
Borrowed shares are not free. The lender charges a borrow rate, quoted annually and accrued while you hold. For an ordinary large company it is small. For a company that many people want to short and few people own, it can be enormous, because the supply of lendable shares is the constraint.
You also post margin — collateral against the possibility that the price rises — and if the position moves against you far enough, you are required to add more or the position is closed for you. That demand is what "margin call" means, and it is the mechanism that turns a paper loss into a forced trade at the worst moment.
Margin Call charges the engine's real borrow rate and real margin requirement on every short, and states both on the ticket before you confirm rather than after.
A short squeeze is the feedback loop the structure makes possible. If a heavily shorted stock rises, some shorts are forced to close. Closing a short means buying. That buying pushes the price higher, which forces more shorts to close, which is more buying.
Nothing about the company needs to have changed. The buying is mechanical, not informed, which is why squeezes overshoot violently and then give most of it back once the forced buying is exhausted.
The number that measures the risk is not short interest on its own but days-to-cover — short interest divided by average daily volume, or roughly how many days of normal trading it would take for every short to close. A heavily shorted stock that trades enormous volume can absorb the covering; one that trades little cannot.
Chiefly as a hedge. Being long a sector and short one weak name in it isolates the view you actually have. Outright shorts as a directional bet require you to be right about both the thesis and the timing, against an instrument that punishes being early more harshly than being wrong.
This is a description of a mechanism, not a recommendation. Nothing here is financial advice, and Margin Call is a simulation — the shares, the borrow and the losses are all imaginary, which is the point of practising with them.
Where this shows up: