The problem with crossover systems
- Thread starter 1a2b3cppp
- Start date
What you said.
What does that mean... "duped"?
(I can't start a PM with you, but you can with me if you choose.)
The "problem" with crossover systems is they don't do what you want. That is, they don't "confirm a new trend so that it's safe to enter the water" on a trade. And they don't reduce or eliminate whipsaws.
That's kind of correct, but also a bit unfair on this useful pattern. No crossovers indicate a trend is about to start. But how could they?
But they are pretty good at confirming a trend might have started some time ago. Price isn't necessarily in a trend - rising prices don't automatically mean there is an uptrend. All trends have a MA crossover but not all crossovers indicate a trend has started. And none f them indicate a trend is going to start tomorrow.
In the K.I.S.S. frame of trading..
1. Crossovers dont' do what you want.
2. Even when crossovers appear to be "good", they're lagging with price. If you're a sharp trader, you should have seen the trade before the cross.
3. Some crossover params may actually accerate your losses. Trading on a daily average crossover is weak. Trading the MACD is weaker still.
If you cared about my view... I'd say, "forget the notion of daily price, moving average crossovers". There is better to be traded. (Interestingly, my first exposure to TA was the "20 day MA crossover".)
Last edited:
1. Crossovers do what they do. Traders run into fatal losses when they try to make them do something they can't. Plus two other scenarios - one, if traders blindly enter on a cross: without some filter this is random trading; the other, is if traders buy on a cross and hold until there is an opposite cross.
2. Yes, there is always a trade before the crossover. The crossover confirms a trend might have already developed: if you only want to trade maturing trends, that is not a problem.
3. It works, its always worked, it will always work. It isn't perfect but the very poor reputation is down to very poor use, not very poor characteristics.
Congrats. You've successfully argued against your own case. Good for you.
Trend line / regression line is ‘quicker’ combined with historical intraday timepoints where orderflow is known to cause the largest historical price variance.
