Can linear regression analysis really predict the future?

[dtrader98]

I considered all the prior questions in my response, including yours. A word of thanks means a lot around here, and goes a long way towards picking brains.

Here's another way to think about how and how our minds perceive information (and there are plenty of papers related under the ideas of 'visual interpolation' and 'the binding problem') in a regressive manner. It gives ideas as to why we can also be fooled by things like optical illusions.

So much information is discrete in the world that our brains must find a way to interpolate/stitch the information in a continuous form recognizable to us.
Take video for instance, it is not continuous but is comprised of dynamic frames being refreshed at a rate that appears continuous to us. Photographs are comprised of discrete pixels,not continuous lines.=)

 

[Rocko Bonaparte]

Quote from wutang:

If what you're saying is true why do we need computers to do regression analysis with stock prices? Can't one just become accustomed to the "inertia" of price and other traders and trade by eye or by "feel" as we do when we drive cars? We don't need computer models or high level math in order to drive, why do we need it to trade?

I suspect this has something to do with all the posts on here I see about using price action, and that the only indicators people need are price and volume at a maximum.

I'm wondering what kind of sophisticated stuff you all are doing with linear regression. I had dabbled in doing polynomial curve fitting, and tried to use that to extrapolate the next value. It wasn't profitable, but I didn't really build a strategy around the indicator to try to better account for its behavior in any way.

 

[wutang]

Quote from dtrader98:

I considered all the prior questions in my response, including yours. A word of thanks means a lot around here, and goes a long way towards picking brains.

At the moment I think I'll just be silently following threads like this as I know I've got a lot to teach myself before I could really ask any questions of value. All I can add is to let you guys know that these threads are not unappreciated. I find it interesting that the few people on this site who really seem to know what they're talking about are also the ones who are most willing to help. Alright, I won't interrupt this thread again.

 

[MAESTRO]

Quote from cfd_trader:

What is the probability that the "random walk" of index/stock/futures prices is just that, "random", i.e. 100% random, 100% of the time?

IMHO: 0 - null, zero. ...

Well, like many others on this site you are mixing up two things: Random Processes and Uniformly Distributed Random Generators that are usually used to form a Random Walk. Many random processes have very stable and predictable distribution patterns thus enabling us to efficiently deal with them. As for your statement it is, unfortunately, 100% wrong. Random nature of price fluctuations is the only insurance of the market efficiency.

 

[MAESTRO]

Quote from wutang:

If what you're saying is true why do we need computers to do regression analysis with stock prices? Can't one just become accustomed to the "inertia" of price and other traders and trade by eye or by "feel" as we do when we drive cars? We don't need computer models or high level math in order to drive, why do we need it to trade? Doesn't this describe what traders already know about mean reversion and breakout trading strategies, buying when traders are anxious to buy and selling when buyers dry up and feel nervious holding?

There is a huge difference between your observations as a car driver and the information that your gauges are telling you. Any successful driver, pilot etc. is a result of training on how to use the driver's skill and his/her abilities to quickly process the information that is presented by the gauges on the cockpit.

 

[cfd_trader]

MAESTRO, with such a pseudo you must be a master in all things therefore I shall not dare contradict your superior sayings.

More seriously though, you're not answering my point about specific "patterns" (eg double/triple tops) that do repeat over and over (sometimes down to a single tick accuracy - though not all the time oviously, otherwise there would be no market.

If as you say randomness is a prerequisite for the markets to be "efficient", what if the markets after all are not efficient all the time?

Does someone truly believe that the markets are "efficient" 100% of the time? That there's never any anomaly any time in any market? What about front-running, and all other sort of "manipulations" of the markets by the Big Boys? Are these part of the process that guarantees that markets are "efficient" at all time??

 

[MAESTRO]

Quote from cfd_trader:

MAESTRO, with such a pseudo you must be a master in all things therefore I shall not dare contradict your superior sayings.

More seriously though, you're not answering my point about specific "patterns" (eg double/triple tops) that do repeat over and over (sometimes down to a single tick accuracy - though not all the time oviously, otherwise there would be no market.

If as you say randomness is a prerequisite for the markets to be "efficient", what if the markets after all are not efficient all the time?

Does someone truly believe that the markets are "efficient" 100% of the time? That there's never any anomaly any time in any market? What about front-running, and all other sort of "manipulations" of the markets by the Big Boys? Are these part of the process that guarantees that markets are "efficient" at all time??

Markets are not efficient at all times, however, they ARE most of the time and with the development of faster and faster electronic arbitrage tools they exhibit less and less opportunities to make reliable profits. If you run the true Random Walk you will surprise yourself by spotting all the usual TA patterns (i.e. Head and Shoulders, Support/Resistance, trends etc). All of those things are reactions of our visual cortex to the presented data set. Unfortunately, as the result of our evolution process our brain is constantly looking for patterns even when the patterns do not exist.

 

[TrueRange]

Best ET thread ever. Thanks guys.

 

[themainevent]

Interesting discussion all around.

My feeling about indicators on a whole is that they are all just reinterpretations of the same data in a different visual format.

Ultimately, trend and price action are all I need, but I always like to learn about what others are using to maybe add to my trading.

 

[MAESTRO]

Quote from themainevent:

Interesting discussion all around.

My feeling about indicators on a whole is that they are all just reinterpretations of the same data in a different visual format.

Ultimately, trend and price action are all I need, but I always like to learn about what others are using to maybe add to my trading.

Trading is a probabilities game. I personally believe that placing bets on very frequent events that exhibit stable distribution patterns is the only reliable way of making money on the markets. Many traders like “Set-Ups". Although there is nothing wrong with that one thing needs to be considered is the frequency of those setups. If they are not very frequent then reliability of the decisions that are based on them is poor. Any stable pattern can only reliably exhibit itself through quite large sample set. Think of a coin tossing game. We know that the distribution should be 50/50 but in order to enjoy this prediction one needs to toss the coin quite a few times. If, on another hand, the sample set is limited to, let us say, 10 tosses then the pattern could be very unreliable as the coin might land on heads all 10 times. So, the key to success in trading is to find a very frequent, stable pattern and exploit it based on its known distribution.