It's create loss not sl. But people wont tolerate collapse of probability, and will cling to any ideology they fancy. Morons everywhere.
I don't like to use Standard Deviation myself simply because it changes according to changes in price which in turns changes volatility. How would be able to place stops on something that changes according to new information? It's like trading based on indicators that repaints. Second Standard Deviation based stops are very hard to control in terms of risk exposure of your capital. Say you have had large price swings in recent tf periods, then that's going to spike up volatility which will directly affect the perceived standard deviation of price. Let's say a price swing of 4 cents in either direction used to be perceived as within 1 sd, now with the large swings, that same 4-cent price swing is only 0.6 standard deviation. If you have placed a stop based on 1 standard deviation, then your trade will be stopped lot later, lot farther away and after suffering lot larger losses simply because the price behaviour has changed, exposing your trading capital to more losses. On the reverse, when the perceived becomes bigger for the same price change due to flatter market, the stops become tighter and your trades get taken out more often, robbing you of potential profits, now this time under-utilizing your trading capital.
In those circumstances (falling volatilty) you should reduce your stop, and increase your position size to reflect the lower risk. When vol rises you should increase your stop and reduce your position size. I note in passing that this would also be a problem with ATR or any stop based on recent price changes. GAT
Your post contradicts itself -- you blame std deviations both for changing and for not changing -- for responding to new info, and for being unresponsive. And as Globalarbtrader noted, this would be an issue, whether you used σ or ATR. No matter the tool used to quantify your intuition, *you* first have to decide where to go, and *you* have to be comfortable with the idea that the risk capital you have put on the (proverbial) table is fair in relation to the reward sought. *That's* your first task.
Not too sure about the above comments … here’s the S&P 500 over the last several months using a 100 day lookback for both measures …
6 mos. is 120 trading days. That's like using a 100-Minute Moving Average to find trends on 2 hours of trading. The total period tested/trained and displayed should be many, many times that lookback parameter.
Thanks, KCOJ. Monday might be nasty for me, but I'll see if I can (soon) put up a current chart to show what I mean.