5 ms·

As I understand the article the "robots" would have done more or less ok. It was human intervention that caused more losses. This is exactly what not should h

As I understand the article the "robots" would have done more or less ok. It was human intervention that caused more losses. This is exactly what not should happen when investing longterm: Being driven by emotions rather than rational facts.

adwn 10y ago
> This is exactly what not should happen when investing longterm: Being driven by emotions rather than rational facts. Putting a limit on one's potential losses is pretty rational, not emotional. Models are not perfect, and you do not want to find yourself with high paper losses when they fail. Of course, if your models don't work well together with stop-loss limits, then that's a problem.
- binaryanomaly 10y ago
  > Putting a limit on one's potential losses is pretty rational, not emotional. Models are not perfect, and you do not want to find yourself with high paper losses when they fail. At least here they say that it is not a good idea with a quant approach: http://en.swissquote.com/epb/support/faq#node-301 http://en.swissquote.com/epb/support/faq#node-301 The BlackRock example plus the fact that many ETFs beat managed Funds in terms of performance seem to support the position that human interference is in most cases for the worse.
- ohyes 10y ago
  But you do that by properly sizing your positions, not by panic selling.
CoolGuySteve 10y ago
Adding a stoploss is common practice. It prevents bugs, over-fitting, and adverse selection from ruining you. You'll have a bad time if you think your model is infallible.
- mrchicity 10y ago
  It gives a false sense of security in cases like this. The biggest losses you want to liquidate will always happen in cases where it's impossible or extremely costly to do so: company has merger announcement overnight while you're short, accounting scandal, bankruptcy, earnings surprise, fat finger order takes all liquidity out of the market. You will end up losing many multiples of your designated loss limit in the bad cases. The only way to prevent massive losses is to size your bets sensibly or insure against tail risk using options ahead of time, of course that comes at a cost. By bugs and adverse selection I'm assuming you're talking about something like automated market making. For traders like that, a loss limit makes sense because of technology risk. If your order router has a bug like Knight's did, or your data feed gets stuck, you can lose a lot of money very quickly.

adwn 10y ago

> This is exactly what not should happen when investing longterm: Being driven by emotions rather than rational facts. Putting a limit on one's potential losses is pretty rational, not emotional. Models are not perfect, and you do not want to find yourself with high paper losses when they fail. Of course, if your models don't work well together with stop-loss limits, then that's a problem.

binaryanomaly 10y ago
> Putting a limit on one's potential losses is pretty rational, not emotional. Models are not perfect, and you do not want to find yourself with high paper losses when they fail. At least here they say that it is not a good idea with a quant approach: http://en.swissquote.com/epb/support/faq#node-301 http://en.swissquote.com/epb/support/faq#node-301 The BlackRock example plus the fact that many ETFs beat managed Funds in terms of performance seem to support the position that human interference is in most cases for the worse.
ohyes 10y ago
But you do that by properly sizing your positions, not by panic selling.

CoolGuySteve 10y ago

Adding a stoploss is common practice. It prevents bugs, over-fitting, and adverse selection from ruining you. You'll have a bad time if you think your model is infallible.

mrchicity 10y ago
It gives a false sense of security in cases like this. The biggest losses you want to liquidate will always happen in cases where it's impossible or extremely costly to do so: company has merger announcement overnight while you're short, accounting scandal, bankruptcy, earnings surprise, fat finger order takes all liquidity out of the market. You will end up losing many multiples of your designated loss limit in the bad cases. The only way to prevent massive losses is to size your bets sensibly or insure against tail risk using options ahead of time, of course that comes at a cost. By bugs and adverse selection I'm assuming you're talking about something like automated market making. For traders like that, a loss limit makes sense because of technology risk. If your order router has a bug like Knight's did, or your data feed gets stuck, you can lose a lot of money very quickly.