Awesome Oscillator: What the Colour of Each Column Means

The Awesome Oscillator draws a histogram under the price chart, one column per bar, in two colours.

Almost every published account of it reads those colours as the message. Green is described as buying pressure and red as selling pressure, and the reader is left with the impression that the colour of the newest column says which way the market is leaning.

It does not. The colour is decided entirely by a comparison with the column immediately before it, which makes it a statement about change and not about level. A column can be green while the reading is deeply negative, and red while it is strongly positive.

This page separates the colour from the level, states what the indicator computes and on which price, and sets out the zero-line conditions that the three documented signal definitions carry and that the ranking accounts leave out.

What the Awesome Oscillator Computes

MetaQuotes states the construction directly in the MetaTrader 5 documentation. A 34-period simple moving average is subtracted from a 5-period simple moving average, and both averages run on the median price of each bar rather than on the close.

Written out, the median price is the high plus the low divided by two, and the oscillator is the 5-period average of that quantity minus the 34-period average of it. The result is plotted as a histogram around a zero line, positive when the short average sits above the long one and negative when it does not.

That is the whole definition. Everything else attached to the indicator is interpretation layered on top of a difference between two averages, and the interpretation is where the published accounts start diverging from the documentation.

The relationship to the moving average behind it is worth holding on to, because a reader who understands that one average reacts faster than the other already understands why the histogram crosses zero when it does.

Why Each Column Is Coloured Against the One Before It

The colouring rule is mechanical. A column that stands higher than the column before it is drawn in one colour, and a column that stands lower than the one before it is drawn in the other. The comparison is with the neighbour, and with nothing else.

Two consequences follow, and neither is usually stated.

The first is that colour and sign are independent. A reading well below zero that has risen slightly from the previous bar is drawn in the rising colour, even though the short average is still under the long one. A reading well above zero that has eased is drawn in the falling colour while remaining positive. Colour tells you the last change; the position relative to zero tells you the state.

The second is that a run of one colour is a run of consecutive changes in the same direction, which is a different observation from the histogram being high or low. Three rising columns near the bottom of a deep trough and three rising columns just under a peak look identical in colour and mean different things.

The practical effect of collapsing the two is that a reader watching colour alone is watching a first difference and believing they are watching a level. That mistake is available on any histogram drawn this way, and it is why the sections below treat the sign of the reading as the primary fact and the colour as secondary.

Median Price Is Not the Close

The choice of input is stated in the documentation and is almost never repeated in the accounts built on it. Both averages are taken over the midpoint of each bar, not its closing price. The same input is used by the Williams Alligator, which builds three lines on it rather than a difference between two.

On a bar with a small range the two are close enough that the distinction rarely shows. On a bar with a long wick they are not. A bar that trades far in one direction and closes back near its open has a midpoint pulled toward the extreme, so the median-price average moves where a close-based average would barely register the bar at all.

This matters most where long wicks cluster, which is around scheduled releases and session opens. An oscillator built on midpoints treats the extremes of those bars as real, because for its purposes they are the data.

Midpoint inputs are not unusual once you look for them. Several volume-weighted and range-based measures start from the same quantity rather than the close, and VWAP is the most widely used example of a study that begins by reducing each bar to a representative price.

What separates this case is that the choice is invisible in the output. The histogram gives no indication of which price it was built from, so a reader has to know.

It also means the indicator cannot be reproduced from a close-only series. Anyone recomputing it in a spreadsheet from closing prices will get a curve that resembles the platform version and does not match it, and the mismatch will be largest on exactly the bars that attract the most attention.

The Zero-Line Condition the Saucer Description Usually Drops

Three shapes are defined on this histogram in the MetaTrader 5 documentation, and each carries a condition about the zero line that the readable published accounts omit.

The saucer is defined as a reversal in the histogram from falling to rising, formed across at least three columns, and the documentation is explicit that every column involved must sit on the same side of the zero line for the shape to count. A three-column reversal straddling zero is not the defined shape.

The zero line crossing needs only two columns, one on each side, and the documentation notes that the buy and sell versions cannot be generated at the same moment.

The twin peaks definition carries the strictest condition of the three. Two troughs form on the same side of the zero line, the second shallower than the first, and the histogram must remain on that side of zero throughout the section between them. If it crosses zero between the two peaks the shape does not hold at all, and the documentation says a crossing signal is what has been produced instead.

None of this says a shape predicts anything, and the documentation makes no claim of that kind either. It is a specification of when a named pattern exists, which is the part a reader needs before any question about usefulness can even be asked.

Reading the Histogram Without Reading the Colour

Once colour is set aside, three facts remain on the chart and all of them are unambiguous.

The sign says which average is on top. Positive means the 5-period average of median price is above the 34-period average; negative means it is below. That is a statement about the recent midpoint range relative to a longer one, and nothing more.

The distance from zero says how far apart the two averages are. A histogram far from zero is describing a wide separation, which is what a fast move produces and what a drift does not.

The change from one column to the next says whether that separation is growing or shrinking. This is the same quantity the colour encodes, read as a number rather than as a hue, and read that way it can be compared with the column three bars ago instead of only with the one immediately before.

A reader who wants a wider frame for where this sits among indicator families will find it in how technical indicators are grouped.

Where It Repeats What a MACD Histogram Already Tells You

Both instruments are a difference between a faster and a slower average, drawn as a histogram around zero. The family resemblance is not a coincidence and it has a practical consequence.

The differences are in the inputs and in what the histogram is a difference of. This one uses simple averages of the bar midpoint. A MACD histogram is conventionally built from exponential averages of the close and plots the gap between the MACD line and its own signal line, which is a second layer of smoothing this indicator does not have.

What that means on a chart is that placing both under the same price series produces two curves that turn at similar moments for similar reasons. Reading agreement between them as confirmation is double counting, because the second reading is largely a restatement of the first with different smoothing constants.

An instrument that measures something structurally different, such as where volume actually traded on the volume profile indicator, adds information that a second average-difference oscillator cannot.

Claims About This Indicator That Carry No Source

The two period lengths are stated as settled fact by every published account on this subject, and not one of them attributes them to anything. They are documented by MetaQuotes as the platform implementation, which is the reason this page can state them, and that is a different claim from saying where the numbers originally came from.

The performance material is weaker still. One of the readable accounts works through a sequence of crossing signals and reports how many of them made money, with specific entry and peak prices and a percentage move, none of it sourced to any data set a reader could check. The same page states its author experience as two different numbers in two different places.

The dates are worth checking too. Of the accounts readable at the time this page was written, one carried a publication and review date in 2026, one was dated June 2022 with no update marker, and one showed no author and no date of any kind. A page with no date is not necessarily wrong, but it removes the reader’s only way of asking whether it has been looked at since it was written.

No success rate, win rate or expected move appears above. A figure that a publisher states and no primary source supports is not evidence, and the fact that several publishers repeat the same figure only shows that it travels well.

When This Oscillator Will Not Help You

A reader who already runs a MACD histogram on the same chart is unlikely to gain much, for the reason set out above. Two average-difference measures moving together is not two pieces of evidence.

A reader working on very short timeframes where bar ranges are a few pips wide will find the median-price input barely distinguishable from the close, which removes the one property that makes this indicator different from a plainer construction.

And a reader looking for a rule that says when to enter will not find one here, because the documentation defines shapes rather than outcomes, and the published accounts that go further do so without evidence anyone can inspect. Understanding what the histogram measures is the useful part; treating any of the three shapes as a forecast is not supported by anything cited on this page.

Four Things to Confirm on the Chart Itself

Check the sign of the reading before the colour of the newest column, since the two answer different questions and only the first describes the state of the two averages.

Count the columns in any saucer you think you have found and confirm every one sits on the same side of zero, because a straddling reversal is a different shape with the same appearance.

Trace the section between two troughs before calling them twin peaks and confirm the histogram never crossed zero in between. Then check what else is already on the chart, and drop whichever of two average-difference instruments is telling you the same thing twice.