## Intro to Linear Regression

By JohnLast 4458 days ago Comments (5)

A really brief introduction to the "best fit" line through X:Y data.

Watch this video. As a matter of fact all the basic technical analysis are a variation on the theme of the Linear Regression Analysis.

- Trend classification: The trend line is going up - up trend; or the Trend line is going down - down trend

- Finding an entry within the trend: two basic entries:

a. Buyung the dips when in up trend, or Selling the Highs in a Down trend

b. Buying the break - outs over the previous highs or the previous lows.

- Finding when a trend reaches its end. Counter trend entries when the trend - line is violated.

For example Head and Shoulders. And of course there is no reversal pattern unless there was a previous

Trend.

The trend is the technical term very similar with the Linear regression analysis. Of course it is not exactly the same when you form a trend line you are looking to find the lows (uptrend) and the highs (downtrend).

However the linear regression channels are quite popular within Metatrader and are a part of some very popular templates of indicators. Consider also the Andrew's Pitchfork (linear regression again).

So the idea here is to see that what may looks like some kind of mystical technical analysis tools are in fact an application of  basic statistical tools.

And when you form trend lines you are doing that manually and as you will see in the video the linear regression line is the line that minimizes the error.

• JohnLast 4419 days ago

And yet once again the theme for the linear regression came from behind the scenes.

However the idea's of Darvas boxes was something more sophisticated that the simple linear regression, because his approach was incorporating the idea of the changes in volatility. As the volatility contracts, the market starts to move somewhat sideways. And those moments are creating the so called boxes.

From a standard technical analysis perspective those boxes are the so called continuation patterns. However Mr. Darvas was really not bothering with names. They were just boxes of volatility (not penants or triangles etc.)

What is interesting is that the same ideas comes again and again under different envellopes and names.

According to some commentators those ideas worked because they were perfectly adapted to the market state at that historical period. Clear Bull trends lasting for years.

Even more Mr. Darvas was doing a perfect fundamental analysis, considering to invest in the things that were at that point the future: technology companies.

• JohnLast 4402 days ago

And yet Mr. Darvas made money on the market. Most of the new commers do not know that the so called technical analysis was created with the very long frame in mind, like the daily.

The technical analysis from the books is exactly that.

The technical analysis, the so called 'price action', is intended to make money on the short term, and very short term using highly leveraged accounts.

The main results of this is providing liquidity for the market makers.

• JohnLast 4402 days ago

I was thinking to add r-squared as a filter in the spinal implant, however the results are not confirming by now its usefulness as a filter.

I use two ways as a positive and as a negative condition.

• jaguar1637 4401 days ago

Yes, Mr Darvas makes money with chart on daily basis, as many other traders are able to do it

I would like to watch him in smaller tf's, means 5MN or 15MN

• JohnLast 4401 days ago

That is a HUGE question here.

Because you may think that the small time frames have different characteristics than the big time frames and are funfamentally different.

Or on the other hand you may say that they are fundamentally the same, but on the big time frames you have a small sample.

Basically you can say that Darvas made money because he was lucky, that the market behave at that time according to its trading plan but on bigger time frame he would loose it.

That is q big question here, becaue you may even say that Warren Buffet is very rich because he was lucky. And according to the probabilities at least several Buffets would exist. But on longer time periods he would loose it all (200 years of time horizon or so LOL).