Settings

Every input on this pane is a deliberate choice. That framing matters more here than on a single-timeframe stoch, because the instrument gives you two smoothing passes per slot and then asks you to blend three slots across timeframes. A knob turn that is mild on a single stoch compounds across three slots and a blend into a reading you did not quite intend. Moving K Length from 14 to 9 on one slot, for example, does not just "speed up slot 01" — it speeds slot 01's raw %K, which is then fed into a smoothing pass, which is then fed into a D pass, which is then blended with slots 02 and 03 at whatever weight you gave it. Five downstream effects from one knob. That is the instrument; it is not a bug.

Use this page two ways:

• If you are mid-configuration, jump to the decision map below and use it as a triage list. Come back and read the surrounding notes when the knob matters.

• If you are reading linearly, the primary inputs are grouped by slot, then by oscillator display, then by the optional master pass, then by the global repaint switch at the bottom.

Nothing here is a recommendation. Each input is walked with a cost in both directions so you can make a choice you can defend under your own reasoning. The defaults are a sensible starting place and nothing more. If a default fits what you are trading, keep it; if it does not, change it for a reason you can state.

Decision map — what to touch in what order

Per-slot:

1. Pick a Source:. close is the textbook baseline and the default. hlc3 and ohlc4 change the raw %K on sessions with long wicks. Important: the source only substitutes for the close term of the stochastic formula. The high and low come from the slot's timeframe context, not from this input. (See MTF and Repainting for the full note.)

2. Pick a TimeFrame: that is at or above the chart timeframe. A slot set below the chart timeframe raises a runtime error naming the slot.

3. Pick a K Length: you can defend on the instrument you trade. Shorter is quicker and noisier. Longer is calmer and lags more. There is no fast/slow pair on this pane, so both smoothing passes compound on top of this one lookback.

4. Pick a K Smoothing: length and K Type: family. Together these define the first smoothing pass — how close the slot's K line stays to raw %K.

5. Pick a D Length: and D Type:. Together these define the second smoothing pass, which sits on top of K. A change in D length is felt more when K smoothing is short than when it is long.

6. Pick a Blended Weight:. Weight does not affect the slot's own line or the slot's alerts. It only affects how much this slot steers the blend. Weight 0 excludes the slot from the blend but keeps everything else running — see C-3 below.

Global:

1. Leave Enable Master Smoothing off unless you have a named reason to turn it on. It is not a correction layer. It is a lag-added view of the blend.

2. Decide On Bar Close?. ON is stable. OFF is earlier at the cost of intra-bar revision. This is one switch for all three slots at once. Per-slot repaint control lives on CTX.

3. Leave reference lines at 80/20 unless your instrument's session-to-session behavior tells you something different is useful. The pane drives no logic or alerts on those lines — they are reference zones on the display.

That is the order the rest of the page walks in.

Per-slot primary inputs

Slot 01 is documented as the archetype. Slots 02 and 03 carry the same inputs with different TimeFrame: defaults (15 and 60). Labels match the source script exactly.

Enable Stoch 0N

Boolean. Default true on slots 01–03. Turns the whole slot on or off. A disabled slot does not compute, does not plot, does not contribute to the blend, does not feed alerts, and does not count in the alignment tally.

This is the only knob that fully takes a slot out of every downstream behavior. Zeroing a weight does not. Hiding the plot does not. If you want a slot to be gone, disable it.

Hide Stoch 0N Plot

Boolean. Default false. Hides only the slot's plotted K line. The slot still computes. Its K and D still exist in memory. If its weight is non-zero, it still contributes to the blend. Its per-slot alerts still fire. It still counts in the alignment tally.

Useful when you want a slot to feed the blend or drive alerts without adding another line to the display. Trap: hidden slots are easy to forget about. If a day later the blend moves in a way the visible slots cannot explain, the first thing to check is whether a hidden slot is still pulling its weight. (See Troubleshooting.)

Source:

Series. Default close. The series used as the close argument of the stochastic formula. close is the textbook baseline. hlc3 and ohlc4 can shift the raw %K meaningfully on bars with long wicks.

The important detail, worth naming once so you do not have to guess later: the stochastic formula takes three inputs, not one — a close, a high, and a low, over the K Length lookback. Source: only substitutes for the close. The high and low come from the slot's timeframe context (a 5m slot sees 5m highs and lows, a 60m slot sees 60m highs and lows), not from whatever you picked here. So if you swap to hlc3, the slot's raw %K changes because the close term changed — but not because the whole measurement was recomputed from scratch.

TimeFrame:

String. Defaults: "5" on slot 01, "15" on slot 02, "60" on slot 03. The higher timeframe this slot runs on. Must be at or above the chart timeframe. A slot set below the chart raises a runtime error that names the slot — for example, Stoch 02 timeframe cannot be lower than the chart timeframe. if you set slot 02 to 1 on a 5m chart.

The error is the indicator refusing a nonsensical configuration. It is not an indicator fault. Fix the slot or raise the chart.

K Length:

Integer, default 14, minval 1. The lookback passed to ta.stoch. Shorter reacts faster and is noisier. Longer is calmer and lags more. Because both smoothing passes below compound on this one lookback, a long K Length with a long smoothing pair produces a slot that can fall well behind the underlying price story.

14 is textbook default for a reason — it is a reasonable compromise across a lot of timeframes and instruments. Worth saying out loud because the first thing some readers want to do is "optimize" it. You can. Understand that changing it cascades through both smoothing passes and therefore shapes everything else on the slot.

K Smoothing:

Integer, default 3, minval 1. Length of the first smoothing pass. Applied to raw %K to produce the slot's K line. Short values stay close to raw %K — you will see more bar-to-bar movement. Long values pull the K line toward 50 and slow its response.

Together with K Type: (the MA family), this is where the slot's first real character decision gets made. On Base, you cannot tune inside the family. The family choice plus the length is the whole surface on this pass.

K Type:

Enum, default SMA. MA family used for the K smoothing pass. Drawn from the Axiom Moving Average Library Lite (the library the Base trim imports).

SMA is the uniform-weight baseline. Recent bars and older bars in the lookback count the same. EMA is more reactive to the most recent bar. Other Lite-library families have their own lag and noise profiles. The Lite-library table below covers the families that actually ship on Base in the current source.

The Base trim does not expose the wider Pro-library families or their inside-family knobs. On Base, the Lite family choice plus the length field is the tuning surface for this pass. That is a real constraint, not a temporary one.

D Length:

Integer, default 3, minval 1. Length of the second smoothing pass. Applied to the K line to produce the D line.

Because this pass sits on top of K rather than on top of raw %K, the effect of a given D Length: change depends on how short or long K Smoothing is. Short K smoothing plus a long D length gives you a K line that moves faster than its own D — the slot is more likely to show a K-vs-D flip. Long K smoothing plus a long D length gives you two very slow passes stacked on top of each other, and the slot can become unresponsive enough that you stop learning from it.

One edge case is worth naming because the input allows it: with length 1 on a length-based MA such as SMA or EMA, D can collapse onto K instead of becoming a slower comparison line. That does not create better flips. It can erase the separation the color and alerts depend on. Use length 1 only if you deliberately want that near-no-smoothing behavior.

D Type:

Enum, default SMA. MA family for the D smoothing pass. Can differ from K Type:, and sometimes it should — you can use a reactive K family and a conservative D family to exaggerate the K-vs-D relationship, or vice versa. Same Lite-library constraint applies.

Line Width:

Integer, default 2, minval 1. How thick the slot's K line draws. Cosmetic. Does not affect computation, colors, or alerts. Useful mainly if you want one slot to dominate visually while you are learning to read its behavior.

Blended Weight:

Float, default 33.3, minval 0. How much this slot steers the blend. The blend is a weighted average across enabled slots with non-zero weight and non-na K values — the weights do not have to sum to 100; the blend normalizes.

Weight does not affect the slot's own K or D line. It does not affect the slot's own alerts. It does not affect whether the slot counts toward alignment. All it changes is how much the slot moves the blended pair. A weight of 0 is not the same as disabled. It excludes the slot from the blend and nothing else. (See C-3 in the claims register, and Limitations and Trust Boundaries for why this is worth remembering.)

Lite-library MA family table

Every MA family below is available for K Type:, D Type:, and Master MA Type in the current Base source. Base does not expose HMA, ALMA, Jurik/JMA, KAMA, or FRAMA.

If HMA, ALMA, Jurik/JMA, KAMA, FRAMA, or inside-family knobs are load-bearing for how you want a slot to behave, Base is not the trim doing that work. It is not pretending those controls exist here.

Oscillator group

Overbought Level

Float, default 80, step 0.1. Draws the upper dashed reference line on the pane. No internal logic fires on this level. No alert is attached to it.

If you want an alert when a line crosses this level, you wire it yourself in TradingView on top of the plotted lines. The indicator does not pretend it has a threshold alert. See Alerts for the full list of what is and is not wired.

Oversold Level

Float, default 20, step 0.1. Same design — a visual reference line with no attached logic. The 80 and 20 lines are zones on the pane. They are not triggers.

Display group

Plot Blended K/D

Boolean, default true. Toggles the blended K line, the blended D line, and the tinted fill between them, as a group.

Important: this is a display toggle, not a logic toggle. Turning it off hides the three blend visuals. The blend still computes. The two blend-based alerts (Blended Stoch Is Bullish, Blended Stoch Is Bearish) still evaluate on every confirmed chart bar. If you hide the blend for visual reasons and then get a blend alert, that is the design working exactly as intended — display and logic are separated deliberately.

Blended Line Width:

Integer, default 3, minval 1. How thick the blended K and blended D draw. Cosmetic.

Master Smoothing group

Enable Master Smoothing

Boolean, default false. Runs one more MA pass over the blend before plotting. The per-slot K and D lines are untouched. The per-slot alerts are untouched. What changes is the blended K, the blended D, the tinted fill, and the two blend-based alerts, which will fire later than they would on the unsmoothed blend.

Worth being honest about what this does and does not do. It makes the blend visuals calmer and adds lag. It is not a noise removal in any meaningful sense — the noise is still there, on the slots, reporting what the stack is doing. Master smoothing hides it. If the blend looks choppy and you are tempted to enable master smoothing to "fix" it, a more useful first move is to revisit the slot configurations, because choppy blend behavior usually has its cause on one of the slots and calming the blend at the display layer just pushes the cause further from where you can see it.

A subtle misread worth inoculating against: a reader who enables master smoothing because it "looks cleaner" can start treating the smoothed blend alerts as more reliable than the unsmoothed version. That is the relationship inverted. The smoothed blend is a later version of the same information, not a better one. The extra pass does not filter out noise that did not matter; it delays every flip equally. If you are tempted to conclude the smoothed version is more trustworthy, pause and ask whether you are confusing "easier on the eyes" with "more accurate." Those are not the same thing.

The case where master smoothing earns its keep: you want a deliberately slower regime-level view of the blend alongside the per-slot evidence, and you are comfortable with blend alerts firing later — because you have chosen to act on confirmed regime-level state rather than on faster blend changes. That is a defensible use. Using it to hide chop is not.

Master MA Type

Enum, default EMA. MA family for the master pass. Note the shift: per-slot defaults are SMA at both passes; master defaults to EMA. That is intentional. The per-slot cascade is a classical slow-stoch pair; the master pass is a single smoothing over an already-bounded blend, and a reactive family tends to read better there. Not a recommendation — a named design choice you can change.

Master Length

Integer, default 3, minval 1. Length of the master pass. Higher values calm the blend more and slow it more.

PU Settings group

On Bar Close?

Boolean, default true. The repaint switch. One switch, all three slots at once, on the Base trim.

When ON: each slot's request.security call returns the previous higher-timeframe bar's values. The reading is confirmed — it does not revise as the next higher-timeframe bar forms. Slower to update. Stable across chart bars inside a single higher-timeframe bar.

When OFF: each slot returns the live higher-timeframe bar's values. The reading can revise until the higher-timeframe bar closes. Faster to react. Repaints by construction inside the open higher-timeframe bar.

Both modes are honest if you know which mode you set. Neither is "better" — they are different tradeoffs for different reading contexts. The MTF and Repainting page treats this switch as its own teaching surface and walks a 1m verification. If you are about to turn it off and have not been here before, go there first.

Per-slot bar-close switches are not available on Base. If you want On Bar Close? = true on slot 03 and false on slot 01 simultaneously, CTX is the trim that exposes that.

Cross-references

• The Source: + slot-timeframe high/low relationship gets its full treatment in MTF and Repainting.

• The observer-slot pattern (weight 0 with the slot still plotting) lives inside Workflows.

• The master-smoothing misread — treating the smoothed blend alerts as "earlier" — lives in Limitations and Trust Boundaries.

• The family-choice-as-tuning-surface frame gets reinforced in For the Geeks alongside the overall pipeline.