Understanding Indicators | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Create MT5 indicators from MT4 indicators 4 replies

TOP 100 best indicators: post good indicators here 647 replies

Custom Indicators vs. Technical Indicators 6 replies

Lagging Indicators -vs- Leading Indicators 48 replies

Understanding indicators 2 replies

Attachments: Understanding Indicators
Exit Attachments

Understanding Indicators

  • #1
  • First Post: May 29, 2007 5:40pm May 29, 2007 5:40pm
There’s a lot of cussing and discussing about indicators on these threads, but very little understanding. It’s like the Ford/Chevy, code/no code, etc. arguments that are on every other site on the net.

I’m going to post one of these articles once in awhile when I get the urge, the time, and depending on the responses.

I'm not going to get into the value of an indicator, or how to use it. I intend just to increase understanding of the indicator itself and what it is showing.
  • #2
  • Edited May 30, 2007 7:45am May 29, 2007 5:43pm | Edited May 30, 2007 7:45am
I’ll start with an easy one, moving averages.

What is a moving average? Let’s say we have a series of numbers, and we want a 10 number moving average. For a 10 number moving average, at any point in the sequence you simply take the average of the most recent 10 numbers. When you get a new number, you then have a new set of 10, since the first, or earliest, drops out of the calculation, and the new one is factored in. This would be a simple moving average (SMA).

The problem with a simple moving average is as follows. Let’s say we’re using a 50 SMA, and 50 bars ago we had a big up move. As we move forward bar by bar with our averaging, those big move bars drop out of the calculation, and our average shifts down accordingly, and gets steeper for several bars. We look at it and say, “hmmmm” and think it means something, when actually that move 50 bars ago has little to do with the current market. More recent price moves mean more.

This is why some people like the exponential moving average (EMA). The difference with the EMA is that recent data have more effect, or “weight”, than past data. The older a datum is, the more it gets “watered down” by the EMA calculation, so that data come in strong, then fade into the past slowly.

To calculate an EMA, first you calculate the coefficient, K = 2 / (n+1) where n is the number of days in the EMA. Let’s use 5 periods, then K = 0.333.

The formula is:

EMA = (Todays price * K) + (Yesterdays EMA * (1 - K))

Today’s price has the most effect, it is multiplied by K (0.333) then added directly in. The numbers used to calculate yesterday’s EMA have already each been multiplied by K on the day they were new. But, see that now each of them is getting multiplied by 0.667 each day. The longer a number has been in the calculation, the more times it has been reduced by being multiplied by 0.667.

Day 1 0.333
Day 2 0.222
Day 3 0.148
Day 4 0.098
Day 5 0.065

Remember that with the simple moving average, we would have added the numbers and divided by 5. That's the same as dividing each number by 5 (or multiplying by 0.2) and adding the results. So any number would have a weight of 0.2 in a 5-period simple moving average. In this case, the most recent has the weight of 0.333, and then loses value each day as shown in the table above.

Most trading platforms are set up so you can choose to take the moving average of the closing prices, open, high, low, and sometimes the median price, which would be (High + Low) / 2, the typical price (High + Low + Close) / 3, or the weighted close price (High + Low + Close + Close) / 4.

I hope that clarifies moving averages.
  • #3
  • Edited May 30, 2007 5:48am May 29, 2007 5:51pm | Edited May 30, 2007 5:48am
MACD stands for “Moving Average Convergence / Divergence”. So all a MACD line really represents is how two moving averages are moving, relative to each other.

I’ll use the common settings 12, 26, 9 for the example.

To calculate MACD, first we take the 12 period moving average. Then we take the 26 period moving average. Then subtract the 26 period MA from the 12 period MA.

That’s MACD. So what is that telling us? Well, it’s the difference between the two averages, obviously. So?

Well, whenever we look at two moving averages on the same chart, what are we looking at? Many traders look for places where they cross (zero difference). When they start to spread far apart (increasing difference), it means the trend is accelerating. When they're close together, the market is flatter.

All MACD does is put that information in one line. If the fast average (12) is above the slow (26), MACD will be positive. As they get closer, or converge, MACD gets smaller. At zero difference (MA cross), MACD will cross the zero line. The farther apart the MA’s are, (greater divergence) the greater the magnitude, positive or negative, of MACD. Thus, it shows the divergence or convergence of those two MA’s.

The MACD SIGNAL line is an average of the MACD line. Taking a moving average tends to smooth out the data, and also creates a “lag”. Why does it create a lag? Because recent data are averaged with old data. That slows the action and indicates the trend.

MACD HISTOGRAM is the difference between MACD and the signal line. It is not a histogram in the pure sense of the word. The lines merely show the magnitude of the difference between the two lines.

Now we go one further. There’s a thing called MACD divergence, which means that MACD and price are not tracking each other. Don’t confuse this divergence with the word divergence that the “D” stands for in “MACD”, they are not related. You can have divergence between any indicator and price. Below is an example.

Note that at (1), price goes down, while MACD goes up. This is MACD divergence. MACD is not following price. At (2), price goes up while MACD doesn’t. OK, well, it went up a little, but you get the idea. If MACD had gone down at (2) it would be a more significant divergence.

I hope you now understand MACD.
Attached Image (click to enlarge)
Click to Enlarge Name: macd divergence.jpg Size: 72 KB
  • #4
  • May 29, 2007 6:47pm May 29, 2007 6:47pm
I think it has the potential to end up as a Sticky in the Freshman Forum.
  • #5
  • May 30, 2007 7:22am May 30, 2007 7:22am
According to Dr. Alex Elder, in “Trading for a Living”:

“Professionals divide indicators into three groups: trend-following indicators, oscillators, and miscellaneous. Trend-following indicators work best when markets are moving but give bad and dangerous signals when the markets are flat. Oscillators catch turning points in flat markets but give premature and dangerous signals when the markets begin to trend. Miscellaneous indicators provide special insights into mass psychology. The secret of successful trading is to combine several indicators from different groups so that their negative features cancel each other out while their positive features remain undisturbed.”

OK, he was doing great until he told “the secret of successful trading”. No comment on that, since this is the stuff indicator flame wars are made of, and I don’t have my nomex underwear on.

Trend-following indicators would also be categorized as “lagging or coincident”. Moving averages (SMA, EMA, MACD) are all lagging or coincident indicators.

Oscillators are “leading or coincident” indicators. Examples would be Stochastic, Percent R, RSI, CCI, Momentum. These often turn before price.

Pivot points, trendlines, support/ resistance, and fibs are considered to be leading indicators – they give points where price is more likely to turn.

Miscellaneous indicators can also be leading or coincident. Anything that isn’t an oscillator or trend-following indicator would probably go into this category.
  • #6
  • May 30, 2007 7:39am May 30, 2007 7:39am
Williams Percent R is an oscillator which was invented by Larry Williams, the famous futures trader.

The “R” stands for “range”, so you could call this indicator “percent range”.

The word “oscillator” refers to anything that repeats a movement or vibration.

In trading, oscillators are leading or coincident indicators. This means they turn ahead of, or at the same time as price.

Percent R is calculated as follows (from Larry Williams’ “The Money Tree”):

“Value 1 is Highest High – Lowest Low of the period under study.
“Value 2 is Close – Lowest Low of time period.
“Divide Value 2 by Value 1 to arrive at %R.”

Simply put, you find the range over the time period, then express the closing price as a percentage of that range, that’s all. You would have to multiply by 100 to get percent.

Let’s say that over the time period we have a range of 500 pips, highest high to lowest low. The last close is 100 pips from the low. That’s one-fifth of the range, and %R would be 20. If the last close were 50 pips from the high, that would be one-tenth, and %R would be 90 (close would be nine-tenths of the way from low of range to high of range).

The logic of this is best seen by considering the daily chart for a stock. The first hour or so after the open of the stock market represents the activity of amateurs. The close represents the activity of professionals. The close price, relative to recent price range, is significant in this situation since it reflects what the professional traders are doing.
  • #7
  • May 30, 2007 8:17am May 30, 2007 8:17am
I will say one thing regarding the use of indicators. When I was first told this it knocked my socks off, and since then I've seen it expressed in varying degrees by many on this forum. Here it is:

"Indicators, trendlines, fibs, support/resistance, etc., all give you DECISION POINTS. Price action confirms what you should do."
  • #8
  • May 31, 2007 6:58pm May 31, 2007 6:58pm
blueruby,

I have a trading method that I have been using for several years with the indicators you are writing about, (macd histogram, cci, pivot points, ema's and fibonacci's)
may I post them on your thread?... I am not sure if that's what you intended or if this was a thread you wanted to post without others jumping in.
I guess I could start my own thread...lol
L.C.
  • #9
  • Jun 1, 2007 5:09am Jun 1, 2007 5:09am
Reading this thread I come to realize how utterly little I know about indicators. I have never used them. To us price action traders indicators seem like a foggy mystery
  • #10
  • Jun 1, 2007 5:34am Jun 1, 2007 5:34am
Quoting L.C.
blueruby,

I have a trading method that I have been using for several years with the indicators you are writing about, (macd histogram, cci, pivot points, ema's and fibonacci's)
may I post them on your thread?... I am not sure if that's what you intended or if this was a thread you wanted to post without others jumping in.
I guess I could start my own thread...lol
L.C.
L.C.,

Please start a new thread in the "trading systems" forum. Your system will get better exposure that way and discussions about it will all be in that one thread.

For this thread I intended only understanding, not usage of indicators.

Thanks. I'll look forward to reading about your system!
  • #11
  • Edited 5:15pm Jun 4, 2007 10:24am | Edited 5:15pm
Stochastic Oscillator

Webster’s defines Stochastic as “involving a random variable” or “involving chance or probability”. Either could apply here, even though they are contradictory.

Dictionary.com’s definition states, “of or pertaining to a process involving a randomly determined sequence of observations each of which is considered as a sample of one element from a probability distribution.” I like that better. So we’re saying we have watched something for awhile, and we are considering each data point as part of a distribution. It doesn’t say it’s a normal distribution, but it does imply that probability is being derived from seemingly random data.

Calculation of the stochastic oscillator is as follows:

First, you find %K, the formula for which is: 100% * (Close – Lowest low of time period ) / ( Highest high – Lowest Low of time period).

Does this look familiar? It is the same formula as for Williams %R! Percent K which is not smoothed is the same as %R, and is what you get if you use a Fast Stochastic.

% K is then smoothed to get % D. The most common way of smoothing %K is to simply take the 3-period simple moving average.

So, to recap so far:
1. %K for Fast Stochastic is the same as Williams %R.
2. %D for Fast Stochastic is a 3-period MA of %K.

When you apply a Stochastic, usually you do not get a Fast Stochastic. You get Slow Stochastic.

1. %K for Slow Stochastic is the same as %D for Fast Stochastic.
2. %D for Slow Stochastic has been smoothed again.

In either case, %D is usually called the signal line.

Stochastic is considered to be a leading or coincident indicator.
  • #12
  • Jun 4, 2007 5:03pm Jun 4, 2007 5:03pm
RSI stands for Relative Strength Indicator. Put simply, RSI is a function of the amount of change in closing prices. It is a leading or coincident indicator, never lagging.

To calculate RSI, first you calculate RS, or Relative Strength.

Let’s use 10 periods as an example. To calculate RS, you would look at closing prices for the last 10 periods.

For every up close, you add up the amount of change from the previous day's close and divide by 10 to get the average net up close over the 10 days.

Do the same for the down closes.

Then, divide the average up close by the average down close.

RS = average net up close / average net down close

So if the average up close is greater than the average down close over the period, then RS will be greater than one. If the average up close is the same as the average down close, RS will equal one, and, if the down close average is larger, RS will be a fraction less than one.

Calculate RSI as follows:

RSI = 100 – (100 / 1 + RS)

Think about this. If the down closes are larger, RS will be a fraction, and

RSI = 100 – (100 / x) where x is less than 2.

Therefore, RSI will always be less than 50 when down closes dominate since you'll be subtracting a number greater than 50.

If the average of the net up close is larger than that of the down close, then RS > 1. That makes the denominator of the fraction larger, so the dividend will be smaller, therefore a smaller number will be subtracted, and RSI gets larger.

In fact, whenever the up closes dominate, the denominator (x) will be greater than 2, and therefore RSI will always be greater than 50.
  • #13
  • Edited 9:56am Jun 6, 2007 7:25am | Edited 9:56am
Quoting WTB
Reading this thread I come to realize how utterly little I know about indicators. I have never used them. To us price action traders indicators seem like a foggy mystery
Funny that you mention that. The more I understand them, the less I look at them.

I'm sure they never tell you anything that you can't see by "eyeballing" price action.

However, for automated trading signals, they are necessary. Note that I said "signals". Signals can be a huge help. You can let the custom indicator watch the charts and sound an alert when you need to pay attention. Pretty handy in a round-the-clock market.

As for automated trading, I have to say that doesn't work. IMO, there always has to be a human making the call.
  • #14
  • Jun 6, 2007 9:53am Jun 6, 2007 9:53am
CCI compares typical price (H+L+C / 3) difference from an SMA to the average deviation from that SMA over the same time period.

The CCI calculation starts out with a Simple Moving Average of Typical Price. If you aren’t clear on exactly what that is, go read the second post of this thread, the one on moving averages.

First subtract that SMA from the current typical price. This difference expresses how far price is from the SMA, and in which direction.

A number has no meaning by itself. It has to be compared to something else to give a concept of price movement. CCI compares that price difference to what price has been doing relative to that SMA during the period we are interested in.

To do this, we divide by the average absolute deviation. How is average absolute deviation calculated?

Let’s use 10 periods as an example. We would look at the last 10 typical prices, and add up the differences between each price and the SMA of typical price, dropping all minus signs – we’re not interested in direction here, just in how much typical price has deviated from the SMA of typical prices. Once they are totaled, divide by 10 to get the average absolute deviation. This indicates how much price has varied around the moving average during the 10 periods, which is a type of measure of volatility. Think about it, if price is moving rapidly away from the MA, or varying greatly above and below, the deviation number will come out greater than if it is placidly tracking along the MA.

Divide the price difference now by the average deviation. The 1/0.015 is an arbitrary scaling factor. The formula is (ignore the ````` used as spacers):

. ````````````1 ````````````Typical Price – (SMA of Typical Price)
CCI = --------- ------------------------------------------------------
```````````0.015 ````````Average Absolute Deviation of Typ Price


Why is the scaling factor used? Since the difference of today’s price and the SMA is not going to be greatly different than the average deviation, the non-scaled CCI would usually wind up with values between roughly -2 and +2. Nothing wrong with that, but we like to see indicators that go something like -200 to +200.

Why was 0.015 chosen as a scaling factor? Apparently Donald Lambert felt that if the price were 1.5 times the average absolute deviation, this was a significant point, because he chose a scaling factor that would give a CCI of 100 when this case occurred. This was probably based on empirical data or his own experience.
  • #15
  • Jun 6, 2007 5:46pm Jun 6, 2007 5:46pm
The Awesome Oscillator is not really an oscillator. Sorry.

I don't find anything particularly awesome about it either.

Awesome Oscillator is a MACD which uses 5 and 34 SMA of median price, (H + L) / 2.

If a bar is higher than the previous, it's green; if lower, it's red.

If you put up a MACD and an Awesome Oscillator, you'll see that they are very close. The small difference is because MACD uses an EMA, and, unless you set it differently, it takes that EMA of the close.
  • #16
  • Jun 7, 2007 12:45pm Jun 7, 2007 12:45pm
blueruby. i don't know why you took the effort to do all this.but its helpful
  • #17
  • Jun 7, 2007 1:28pm Jun 7, 2007 1:28pm
Thanks for these explanations!
  • #18
  • Jun 7, 2007 1:37pm Jun 7, 2007 1:37pm
Blueruby

Thank you so much for your explanations. I find them more on my level of understanding. The math is sometimes over my head, but that's ok. I wish i had these explanations in one folder for printing. Thank you again.
  • #19
  • Jun 7, 2007 2:11pm Jun 7, 2007 2:11pm
Quoting blueruby
The Awesome Oscillator is not really an oscillator. Sorry.

I don't find anything particularly awesome about it either.
Best description of the Awesome Oscillator... EVER!!
  • #20
  • Jun 7, 2007 2:23pm Jun 7, 2007 2:23pm
Quoting R Richards
Blueruby

Thank you so much for your explanations. I find them more on my level of understanding. The math is sometimes over my head, but that's ok. I wish i had these explanations in one folder for printing. Thank you again.
You can have that, too! Just click on "thread tools" and select "show printable version".

You're welcome, glad I could help.
0 traders viewing now