Interesting thread. Indeed, you are right, except acts of God (Fukushima for instance), before any crisis, charts give lots of alerts. This is 100% true for any stocks because insiders always buy / sell before the news. I will recommend you to look at Bollinger Bands and volumes (rather than Moving Averages). They are both really useful. CM
If I would have done that, and re-entered long because it didn't happen... than I would have lost about 8% in that time period from Feb'16 until now. So.... do that 4 years... that's significant. And a blackswan event of 20-30% drop doesn't cover it. What you fail to grasp, and what most doom-traders/perma-bears don't acknowledge is that even a significant drop might not be enough to not enter a bull market. Black swans are usually hard to predict. And even if they can be predicted... the bull run might outweigh it. Nothing wrong with trying to find out a way to predict it though, but I don't believe something as simple as a EMA crossover is going to help.
What I'm saying is that the most significant price falls occur when price is already falling, as demonstrated by the location of price and the 20EMA below the 50EMA. Therefore, in order to avoid serious loss from a price fall Black Swan, don't be long when the market is bearish. Step two would be to consider being short, but that should not be an automatic decision based purely on those criteria. There is no way to predict on what day a Black Swan will occur or how deep it will be. But that doesn't mean it can't be planned for.
Here is a further study of the 20 worst 1-day price falls in the Dow since 1900. The object is to establish – “Would a trader have avoided these events by exiting long positions when price closed below the 50EMA and/or the 20EMA crossed below the 50EMA?” 19/10/1987: -22.6%: Yes 14/12/1914: -20.5%: Yes 28/10/1929: -13.5%: Yes 29/10/1929: -11.7%: Yes 05/10/1931: -10.7%: Yes 06/11/1929: -9.9%: Yes 04/12/1933: -9.1%: No 12/08/1932: -8.4%: No 14/03/1907: -8.3%: Yes 04/01/1932: -8.1%: Yes 26/10/1987: -8.0%: Yes 21/07/1933: -7.9%: No 15/10/2008: -7.9%: Yes 16/06/1930: -7.8%: Yes 01/12/2008: -7.7%: Yes 09/10/2008: -7.3%: Yes 01/02/1917: -7.2%: Yes 18/10/1937: -7.2%: Yes 27/10/1997: -7.2%: Yes 05/10/1932: -7.2%: No That’s 16 out of the 20, reducing exposure to a price-negative Black Swan by a factor of 5 times. Even if there was no justification for being short on these dates, that’s a massive level of easy risk management.
I've noticed that, unlike other ETers, you trade somewhat like I do. So I looked up threads you've posted. On this one I ask, since you are a long-term trader, in your 20 and 50 EMAs are the periods days? And do you confine this Black Swan-watchout process to a single market or two that you follow or do are you watching out for a global stock market crash? If the latter, what market is your indicator, the S&P500?
Hi Steve - Yes, these are from daily charts, I only use daily EMA's - 20, 50 and 200. In any market I was long in, if price + 20EMA closed below the 50EMA I would exit immediately. That includes any individual equities. But if I was long in equities, and this pattern showed in that day's chart for the index the equity belongs to, I would also close immediately, even if the equity didn't make the pattern. I haven't looked at using the reverse signal as an exit to cash for short positions, but I can't see why it wouldn't work. The pattern is most useful when it comes up in the index chart - like the Dow or S&P - when it could actually be giving an early warning signal for some equities. This pattern in a single market might best be called the last resort - the chart has probably already got so bearish you'd be out of that market before this printed anyway.
I suppose you are in cash lately because the 20 would have crossed below the 50 a while ago and hasn't rise above it again since.
I went into cash well before the 20/50 dead cross but, while they are inverted, we're clearly not in a consistent uptrend. Not that I suggest the use of a golden cross of these MA's as a buy signal.
I don't use the MA crossover. Rather I quantify trend as a slope (or regression line) and use that as one of several factors in asset selection. I use MAs, in part, to calculate my stop loss order. But our approaches are similar enough that I'd like to talk to you at length. Are you amenable to email correspondence or a phone call? I don't know which of us is more advanced, but we could probably learn from each other.
Hi Steve - Cheers mate. I've started something I haven't tried on ET before - a private thread between just the two of us. So we should be able to get comfortably into some detail. Hoping to hear from you soon, Tom