景氣對策訊號燈用自己來預測自己，無意義

mikeon 發表於 2012-1-26 21:22

法人喜歡用景氣對策訊號燈來預測股市，

這也不對，

因燈號裡的變數之一即包含股價，

用自己來預測自己，無意義。

Institutional investors like to use economic monitoring indicator to predict stock market.

This is not right, because one of variables in the indicator contains stock price.

There is no point in using yourself to predict yourself.

燈號包括M1b、直接及間接金融、股價指數、工業生產指數、非農業就業人數、海關出口值、機電設備進口值、製造業銷售值、批發零售及餐飲業營業額指數等 9 項。

The indicator includes 9 variables including M1b, direct and indirect finance, stock price index, industrial production index, non-agricultural employment, customs export value, import value of mechanical and electrical equipment, manufacturing sales value, wholesale and retail and catering industry turnover index.

mikeon 發表於 2012-1-28 19:12

李建德君說股價在燈號裡的變數才占1/9而已

即便包含自己，影響沒那麼大 

這顯然是連最基本的數學都沒學好

Jiende Li said that stock price only accounts for 1/9 of signal variables.

Even if it contains itself, the impact is not great.

Obviously, he has not even learned the most basic mathematics.

單單看迴歸公式 y = bx + a

等式兩邊分別設成 y 跟 x ，

表示兩邊為不同變數，

這是小學生即懂的道理。

Just look at regression model y = bx + a

Set both sides of equation to y and x respectively

indicates that the two sides are different variables.

This is the fact that elementary school students understand.

Suppose there is a variable x that is constant 10, x = 10

stock price    x

  5.8             10

  6.1             10

  6.5             10

  7.4             10

Correlation coefficient between stock price and x = 0

現在拿股價和 x 來組成一個燈號

燈 = 股價 + x

Now take stock price and x to form a light signal

Light = stock price + x

stock price   light(= stock price + x)

  5.8                     15.8

  6.1                     16.1

  6.5                     16.5

  7.4                     17.4

Correlation coefficient between stock price and x = 1

Stock price only accounts for 1/2 of light signal variables, but correlation coefficient depends entirely on it.

weyzhiro 發表於 2012-1-29 01:54

很好理解，不愧是麥可大，

講得很簡單就懂了 ! !

簡單是最高級的複雜阿 ! ! !

It's easy to understand. Michael, good job !

It's easy to understand !

Simplicity is the highest complexity ! ! !

mikeon 發表於 2012-1-31 01:28

相關係數是指線性相關，x 跟 y 構成一條斜線

Correlation coefficient refers to linear correlation, x and y form a diagonal line.

若 x 跟 y 構成底下3種狀況相關係數為0

1. 橫線或垂直線

2. 散成一團

3. 弧線或曲線

If x and y form the three conditions below, correlation coefficient is 0

1. Horizontal or vertical lines

2. Scattered around

3. Arc or curve

y = x 能畫出一條斜線，

y = 100x 和 y = 0.01x 也能畫出一條斜線，相關係數都是1。

相關係數跟變數的數字或權數的大小無關。

y = x can draw a diagonal line

y = 100x and y = 0.01x can also draw a diagonal line, and correlation coefficient is both 1.

Correlation coefficient has nothing to do with the size of variables and weightings.

mikeon 發表於 2012-2-1 18:53

奇怪的是，那些跟我辯論景氣對策訊號燈很有用的人，

怎麼從來不曾看過他們在追蹤或討論過這些指標。

The weird thing is that those people who debated with me about economic monitoring indicator is useful,

I never see them tracking or discussing these variables.

為何不追蹤 ? 因跟股價的關係度不高。

平常沒在看的東西把它們綁在一起，怎麼突然變重要了呢 ?

無非跟股價指數合在一起的結果，

問題是這樣有意義嗎 ?

Why not track it ? Because the relationship with stock prices is not high.

Why do things you don’t follow suddenly become important ?

It is nothing more than the result of combination with stock index.

But does it make sense ?

mikeon 發表於 2012-2-2 08:04

Onion桑把 x 變數設成跟股價方向相反的例子來反駁，

說景氣燈號不受股價影響那麼大，

這才是錯誤的類比。

Reader Onion refutes the case by setting x variable in the opposite direction to stock price and said that economic monitoring indicator is not affected by stock price. This is a wrong metaphor.

股價跟這 8 項變數不是負相關，而是相關係數不高，所以才不見有人在觀察它們。

平常沒在看的東西把它們綁在一起，

怎麼突然就變重要了呢 ?

因為跟股價合在一起的結果。

Stock prices and these 8 variables are not a negative correlation, but correlation coefficient is not high, so no one follows them.

Things you don't care about bound together suddenly become important.

It is nothing more than the result of combination with stock index.