The difficulty I have with statements like "one move predicts a trend" is that "move" isn't well defined. So the concept of comparing previous day to current day is actually IMO much more mathematically rigorous. But it's not something I would expect to reveal trends. I would expect exactly the opposite - countertrends. Now, if you started scanning instruments for ones where the past 10-20 days were highly positively correlated with the next day, you might be on to something...
Yes, thanks for your insight. I can be a little dogmatic at times without fully considering all options and possibilities. However, with this said, I understand, which you may agree, that at lower time frames moving averages have zero testable significance. This would mean that short term or day traders using MA's to trigger entries have no greater edge than random. Surf
This could be true yet it still be positive expectation to be long, its a matter to how much it rises compared to how much it falls
I think, broadly speaking, you're right. I searched for minute-bar trending strategies in quite a few instruments, and found nothing really worth trading. Which makes sense to me - the risk inherent in gunning for stops (either from the market making or market taking side) goes up as the distance in price (and thus also time) to the hypothetical stops increases. So on short time frames I would expect a healthy quantity of false breakouts. But on long time frames generating false breakouts isn't feasible from a risk perspective. There are two caveats I can see: 1) there are a LOT of instruments in the world. Some may have easy to spot trends on short time frames. I just don't know which. 2) Adding additional sophistication to an entry signal (multiple re-enforcing averages, tick/vol charts etc.) can sometimes separate the chop from the trends. So it IS possible to have a profitable trend following system in markets that do not have a blatant bias towards trending. That said, it's painful and it would make more sense to choose a better game.
One move doesn't predict a trend. One move, any move, if defined correctly should detail a fixed point of support and resistance, period.
From a fixed point determined and defined as "support" to a fixed point determined and defined as "resistance". One half of a cycle.
Agreed. That definition isn't specific enough to take a chart and mark the "moves" even in retrospect.
I absolutely agree. Retrospect isn't good enough though, as you well know. I've defined it and I won't share. There are far smarter individuals in this forum than I am so it shouldn't be too hard for one of those learned individuals to figure it out. Hey, marketsurfer said that Covel was the "expert" maybe he can put it to pencil. Got to be worth a few bucks, huh?