Those that poopood on moving averages, thoughts on this?

[Overnight]

Don't you think it is impossible to get max profit of any trade?...

Yes, I do think it is impossible. Thus my post.

 

[ironchef]

I looked at the article on ranking technical market indicators by Annual Relative Advantage, and I suppose this is great if you just want to select indicators based on average performance and use them exactly as used in the research, but that’s not me.

I actually happened across this thread because I was looking for one that dealt with non-lagging moving averages. I was unable to find one, however. The only thing that turned up was a comment Handle123 made about the moving average created by Alan Hull in 2005.

I came up with an “instantaneous moving average” of my own several years ago with which I was relatively pleased, though it did not behave exactly as I had in mind. But I guess it must have been good/close enough nonetheless, because I never made a concerted effort to iron out its minor foibles until yesterday.

Since the Hull moving average is supposed to almost eliminate lag altogether and improve smoothing at the same time (though he recommends using it for directional signals and not for crossovers which could be distorted by the lag) I plotted both his moving average and my moving average on a one-hour chart to compare the two.

View attachment 196311

I liked mine better because there was less lag, though I have to admit, mine is not as smooth as Hull's. However, its real power is found in using directly corresponding “instantaneous” moving averages generated on lower time-frame charts to pinpoint the precise minute that entering a given intraday position is likely to result in a successful trade which fulfills its maximum potential in terms of profit. (This is just hypothetical of course, since I will not be using it until after New Year’s.)

Can I ask a stupid question?

It looks very similar to a short period EMA?

 

[ironchef]

%%
Maybe, but 80%/+ drawdowns in '29 ....could be a real gut wrencher. And i just found out GE, one of the dirty dozen ,DOW stocks has been kicked out of index 2, now 3 times..Not a stock tip

Maybe there is still hope for my GE holding then?

Another stupid question for you: Are you one of the turtles?

Happy New Year.

 

[expiated]

It looks very similar to a short period EMA?

Several years ago, when I was evaluating indicators, I was interested in exponential moving averages due to their seeming to apply an aspect of calculus that I figured would give them a certain amount of added strength.

However, following subsequent experimentation, I discovered that I could always find a simple moving average that would essentially match any given exponential moving average. As a result, I left exponential moving averages behind and just focused on finding the simple moving average that fulfilled whatever function I was trying to address.

The idea for my “instantaneous” moving average came from a completely different source. If I were to say anything more, people would probably figure it out, and perhaps at some point in the future, someone else will get the same idea. In fact, I can’t imagine why this hasn’t happened already. But in the meantime, I’ll just appreciate that the indicator is at least available for the trading I do.

 

[murray t turtle]

Maybe there is still hope for my GE holding then?

Another stupid question for you: Are you one of the turtles?

Happy New Year.

%%
Its a nickname from childhood; never lived in Chicago.I wouldnt buy GE, but being down 90%/+ from $60 area , a dead cat bounce could work?? Not a prediction; its so far below 50 dma+ most that do a dead cat bounce arent buying

 

[TommyR]

moving averages are linear transformations of the time series. if you sum/weight them over different time periods you can approximate an exponential anyway. and the time series does contain useful info in my opinion so they are probably good.

 

[ironchef]

and the time series does contain useful info in my opinion so they are probably good.

If I may ask: If the market is random, then time series would not contain any information about future stock prices, so you are saying the stock price movement is non random?

 

[murray t turtle]

If I may ask: If the market is random, then time series would not contain any information about future stock prices, so you are saying the stock price movement is non random?

%%
Same coin, but different side- some have noticed + profited off trends. NOT a prediction.

 

[bone]

Use whatever tools you can find that make the most sense to you and your own personal cognitive construct. The secret is that there is no secret. The holy grail is that there is no holy grail.

If it's MA - wonderful. If it's an oscillator - charming. If it's a Bloomberg news feed - good on you. If you're scoring the collective ratings for 100 of the top stock analysts on Wall Street - Fantastic.

I will say this much: if traders in general spent as much time on trade management as they did on trade entry, IMO a much higher percentage of them would be successful.

My 2 cents, YMMV. Happy New Year and good fortune to all !

 

[tommcginnis]

If I may ask: If the market is random, then time series would not contain any information about future stock prices, so you are saying the stock price movement is non random?

"You got it, Clyde."

Anything (pretty much) in finance has, as its best predictor, the prior observation.

Truly and completely random would have no discernible modes -- a uniform distribution. (A fair die rolled on a flat surface.)
Things that follow a Central Tendency 'subscribe' to the Central Limit Theorem -- our favorite Normal Distribution, with symmetric decreases in observations as you get further from the mean, mode, and median point. That depends on IID conditions: independent and identical distributions across the event space.

ANYTHING LINKED BY TIME violates the IID condition, by the very linkage to time (and thus the prior event). Time for Durbin & Watson! (Or, since the advent of easy computing over the last 40 years, time for ARCH, GARCH, and their friends, who really don't care about sniggling little bits of carry-over from one observation to the next, but in fact USE that very carry-over [linked error terms =/= 0] to minimize variance over non-minimized-but-controlled bias.