Say the stock has been increasing quickly, and is currently at $1000. Would selling it if it reaches $900, and rebuying if it jumps back up to $1000 quickly be a decent strategy? If I'm wrong the most I can lose is $100 + transaction fees, if I'm right I can potentially gain a much larger amount if it crashes hard enough.
The normal sequence is to set a stop-limit below the current value, and move it up in line with any increases. You lock-in the gains. Forget about 'buying low, selling high'. Everyone wants to do that. By definition for every winner there is a loser. You cannot time the market. Churning has costs. Buffet is a fan of buy and hold. So am I.
No. In order to make money, you have to buy low and sell high. You are suggesting to sell low and buy high.
Are you talking about a stock you already own or are you asking about short selling?
If you already own stock that's currently worth $1,000, then setting a stop-limit that will sell if the price falls below $900 will put a limit on your losses if it drops and carries on doing so. If it was a temporary "blip" and it picks-up again, you could then re-buy (with a bit of a loss) in the hope it carries on rising. But in either case you won't "gain a much larger amount if it crashes hard enough."
If you don't already own the stock and you're asking about short selling (selling something you don't have at one price and buying it back later [to "repay" the ones you sold] at a hopefully lower, but possibly higher price) then that's different. If it's currently $1,000 and you decide to "short-sell" if it hits $900, then two things can happen:
It continues to fall in price. The longer it falls, the cheaper it will be when you buy it back and repay the "loan" you short-sold. This could potentially gain quite a lot, up to a maximum just shy of $900 if it were to drop to a dollar a share.
If the drop to $900 was a blip, and it climbs back up again, then if you manage to buy-back at $1,000 then you will only have lost $100. However, the price could – at least in theory – climb much higher before you are able to buy (see the story of how to turn $37k into a $106k loss) and you could potentially end up losing an unlimited amount (quite easily more than the original $1,000 price).
Setting a stop loss limit is a good policy. Two reasons:
you might be otherwise occupied when a fall occurs (vacation, whatever)
you might not be 100% rational when seeing your investment evaporate (see Kahneman & Tversky) and it is good idea to set your limits for cutting losses ahead, in cool mind and not in an adrenaline rush.