Ultimate Oscillator: 57% of the Reading Is One Timeframe

Every readable description of the Ultimate Oscillator makes the same promise: three lookback windows combined, so that no single timeframe decides the reading. The formula those same pages print says something narrower. The three windows are weighted 4, 2 and 1, and those multipliers sum to seven, which leaves 57.1 percent of every reading in the shortest window and 14.3 percent in the longest.

The indicator does combine three timeframes. It does not weigh them equally, and the distance between those two statements is where most of the confusion about its settings sits.

What the Ultimate Oscillator Actually Computes

Two quantities are measured on every bar before any averaging happens. The first is the distance from the close down to the lower of that bar’s low and the previous close. The second is the full span between the higher of the bar’s high and the previous close, and that same lower point.

The second quantity is the true range, the measure Welles Wilder built the average true range around. The first is the part of that range finishing above the low point, which is why it is usually called buying pressure. A bar closing at its extreme high produces a first quantity equal to the second; a bar closing at its extreme low produces zero.

For each lookback window, both quantities are summed across the window and one sum is divided by the other. A window of 7 bars gives one ratio, a window of 14 gives a second, a window of 28 gives a third. Each ratio sits between 0 and 1 whatever its window length, because both of its terms grow with the number of bars included.

Those three ratios are then blended and multiplied by 100. That final step is what puts the result on the same 0 to 100 scale as the single-window oscillator most traders meet first, and it is the blending step, not the scaling, that this page is about.

Where the Number Comes From: 57, 29 and 14 Percent

The blend is a weighted average with multipliers of 4, 2 and 1, applied to the 7-bar, 14-bar and 28-bar ratios in that order, and divided by the sum of those multipliers.

That divisor is seven. So the 7-bar ratio enters the result as four sevenths of it, the 14-bar ratio as two sevenths and the 28-bar ratio as one seventh. Expressed as percentages, that is 57.1, 28.6 and 14.3.

This is worth stating plainly because the three ratios are already on a common scale. None of them is larger merely for covering more bars; each is a fraction between 0 and 1 by construction. Nothing in the arithmetic offsets the multipliers, so they survive intact into the final figure.

A reader who adopts the indicator to escape the jumpiness of a short oscillator has therefore bought an instrument that is more than half short-window by design. The two longer windows do not cancel the shortest one. They damp it, and they damp it by a specific and knowable amount rather than by an unstated one.

Nothing about that makes the design wrong. It makes it a particular choice, and it is a choice worth knowing before deciding whether the reading on screen is behaving oddly or behaving exactly as its weights require.

The Weights Are the Periods Upside Down

The multipliers look arbitrary until they are set beside the periods they attach to. Divide one by each period and the pattern appears: one seventh, one fourteenth, one twenty-eighth. Multiply all three by 28 and they become 4, 2 and 1 exactly.

So the multipliers are not a second set of parameters sitting alongside the periods. They are the periods inverted and rescaled, which is what a designer reaches for when the intention is to give each window the same weight per bar it observes. The shortest window watches a quarter as many bars as the longest, and it is given four times the multiplier to compensate.

Both statements about the indicator are therefore true at once, and they answer different questions. Per bar observed, the three windows contribute equally. Per reading printed, the shortest supplies 57.1 percent. A description that offers only the first leaves a reader unable to predict how the line will behave.

None of the seven readable pages on this topic sets out that relationship. The words inversely and proportional do not appear on any of them. The multipliers are printed as three numbers to be memorised, in the same way an oscillator whose colour is a comparison rather than a level is usually presented through its appearance instead of its arithmetic.

Changing the Periods Without Breaking the Weights

Custom period sets circulate for this indicator, generally organised by asset class, and one of the pages carrying them proposes 5, 10 and 20 for dollar currency pairs, 14, 28 and 56 for commodities and 3, 7 and 14 for crypto. No source is given for any of the three, and none of them mentions the multipliers at all.

The multipliers matter here more than anywhere else on the page. Two of those three sets are doubling series in the same way the defaults are: 5, 10, 20 doubles and 14, 28, 56 doubles, so inverting and rescaling them returns 4, 2 and 1 unchanged. Substituting either set leaves the design intact.

The third does not double. Moving from 3 to 7 is not a doubling step, and inverting 3, 7 and 14 gives multipliers of 4, roughly 1.71 and roughly 0.86 rather than 4, 2 and 1. Keeping the published multipliers over those periods hands the two longer windows more weight than the per-bar logic allows, and the indicator quietly stops being the thing its formula describes.

The test is arithmetic and takes a moment. Each period should be double the one before it. If a proposed set passes, the standard multipliers still apply. If it fails, either the multipliers need recomputing or the set needs adjusting, and a platform that exposes only the three periods will not do either. That constraint is unusual among indicators, and it is worth carrying into any decision that puts more than one timeframe on the same chart.

MetaTrader Does Not Ship It

Every readable page on this indicator documents it on a platform aimed at share and futures traders. Not one of them mentions MetaTrader, which is the platform most retail currency traders actually open.

The official MQL4 and MQL5 technical indicator references list the functions each terminal provides for built-in indicators, and both include the stochastic oscillator, Williams percent range, the relative strength index, the Chaikin oscillator, DeMarker, the force index and the awesome oscillator. Neither reference contains the word ultimate anywhere. The MetaTrader 5 oscillators help page lists the same family and likewise omits it.

The practical consequence is that a trader who reads about this indicator and then searches the oscillator menu on a MetaTrader chart will not find it, and the absence is a property of the terminal rather than a fault in the installation. Using it there means adding a custom indicator, which introduces a question the vendor documentation for a built-in tool would normally settle: whose calculation is running.

That question has weight because the inputs are not self-describing. A custom build can expose three periods and hard-code the multipliers, or expose neither, and two builds bearing the same name can differ on both. Verifying the arithmetic against a single bar is the only reliable way to establish which one is on the chart.

Thirty and Seventy Are Not the Signal

The levels 30 and 70 appear on every page consulted. What varies is the job those levels are given.

The fuller accounts describe a buy setup with three separate requirements: price makes a lower low while the oscillator does not, the oscillator drops under 30 at some point during that disagreement, and the line subsequently rises above the high it reached inside the disagreement. The mirror image, using 70, describes the sell setup. On that reading, 30 is one condition of three, and on its own it is not a signal at all.

Other pages present the same number as the trigger, so that a line falling through 30 is described as the oversold reading to act on. Both accounts are published, they are not compatible, and neither acknowledges the other exists.

The disagreement is not settled by the arithmetic on this page, because a threshold is a convention rather than a consequence of the formula. It is settled by knowing which definition a given description is using before adopting its rules. The divergence half of the question is a topic of its own, and how a gap between price and an indicator is read is covered separately rather than restated here.

Which Thresholds Have No Published Basis

The numbers 30, 70 and 50 are repeated across all seven pages read for this article, and not one of them attributes any of the three to a source. Alternative period sets of 4, 8, 16 and 20, 40, 80 appear the same way.

No exchange, regulator or standards body defines this indicator or publishes levels for it, so there is no official document against which those figures could be checked. Where several pages repeat a number that nothing authoritative supports, it is left off this page rather than passed along, which is why no threshold above is presented as a level meaning anything in itself.

The same applies to reliability: none of the sources read offers evidence for how often any of these setups is followed by a move.

Which Setting Applies to You

The defaults suit a reader who wants a momentum reading dominated by the last week of a daily chart, with two slower windows trimming its extremes. Anyone wanting the longer windows to matter more needs different multipliers, not different periods, and most platforms do not expose them.

A shorter or longer set is safe to substitute provided it doubles at each step, and worth avoiding otherwise. Where the reading itself is the question rather than the settings, the distinction between an indicator that leads price and one that follows it is the more useful place to start.