Indicator hygiene: settings drift, signal spam, and honest confirmation.

The synchronization rule

Paired tools must agree, or they produce contradictions by construction.

The EW Convergence system ships as two scripts — a lower-pane HUD and a price overlay — and the manual opens its settings guidance with a blunt instruction: keep the timeframe toggles and convergence thresholds aligned across both, to avoid mixed signals. If the HUD includes the 1-minute chart and the overlay doesn't, the counts driving each will differ; if MinN is 2 in one and 3 in the other, the overlay's labels will stop corresponding to the HUD's events. The result isn't an error message — it's two tools quietly disagreeing while both look correct, which is the most expensive kind of configuration failure because nothing announces it.

Early versus confirmed

One toggle decides whether you're seeing signals or structure.

The overlay's pivot gating is the system's clearest expression of a real tradeoff. With gating off, labels fire on alignment alone — more signals, earlier, including ones that won't hold. With gating on and pivot lines enabled, labels fire only when alignment and a fresh pivot confirmation coincide — fewer signals, later, higher quality. The manual frames it exactly as a switch between early signal and confirmed structure, and both settings are legitimate. What isn't legitimate is switching between them based on whether the current mode is producing the answer you want, which converts a quality dial into a rationalization dial.

Clutter control

Spam filters that protect judgment, not just the screen.

The overlay carries deliberate throttles — labels only on a new pivot, only when the text changes, a minimum bar gap between labels, and print-on-close only — with the manual recommending a starting point of a 20-bar gap with change-only filtering. These read as cosmetic and aren't. A chart printing a label every few bars trains the operator to ignore labels, which defeats the tool entirely; worse, a dense signal field lets an operator find support for any decision by waiting a few bars. Throttling is what keeps each print meaningful enough to actually inform a decision — the same logic as the cadence discipline in the analytics layer, applied to a chart.

• Intraday safe mode exists for symbols without deep fast-timeframe history — enable it and drop the fast TFs rather than trusting thin counts.
• Signals that arrive too frequently get ignored; signals that get ignored may as well not exist.
• Lowering the bar gap 'temporarily' to see more prints is exactly the drift the throttles were set to prevent.

The honest-confirmation test

Configured before the setup, or it isn't confirmation.

The unifying discipline behind all of these guards is simple: indicator settings are decided in advance and left alone through the trade. A threshold adjusted while a setup is forming isn't a filter, it's a search for permission — and it produces exactly the contaminated evidence the journal doctrine quarantines elsewhere. This matters beyond the individual trade because indicator context enters the record: ATR state, regime, and setup notes all get logged and later sliced by the analytics layer, and slices built on inconsistently-configured tools describe nothing. Settings changes are legitimate, documented, and made between trades — never during one.

The key idea

A tool only informs if it means the same thing every time.

Every hygiene rule in the indicator manuals — synchronized toggles, deliberate gating modes, spam throttles, safe-mode for thin data — protects one property: consistency of meaning. An indicator whose configuration drifts produces signals that can't be compared to last month's, can't be evaluated in the review, and can't build the operator's genuine pattern knowledge. Configure once, for a reason, and let the record judge it.