Fisher Transform MSI
- Experts
- Version: 1.25
- Mise à jour: 19 octobre 2026
- Activations: 5
MSI-FT Bi-Driven
What It Does and Why It Matters
The problem it solves
The Fisher Transform tells you where price sits within its recent range, and when momentum is turning. On its own it answers one question: is Fisher above or below its Trigger? Traders who rely on that alone may enter late, exit late, and cannot tell whether a signal is fresh or stale (unless manually trading on a single timeframe).
MSI-FT Bi-Driven answers the questions that come after that one.
| Trader's question | What answers it |
|---|---|
| Is a cross actually coming, or is Fisher just moving fast? | PROXIMITY |
| Is the move still building? | SLOPE |
| Is Fisher on the right side of zero? | REGIME |
| Is price at an extreme where a reversal is likely? | ZONE |
| Do the higher timeframes agree? | HTF filter |
| Is this signal still valid or has it aged out? | Signal Age |
| Which direction does the evidence actually favour right now? | Confidence meter |
| Why did the EA not take that trade? | Status line |
Five signal layers:
Each layer measures a different property of the same Fisher calculation. They are independent and can agree, disagree, or split.
-
SIGNAL - Fisher crossed its Trigger. The classic crossover.
-
PROXIMITY - Fisher is closing on its Trigger. Detects the approach, not the arrival. This is the earliest warning the EA produces.
-
SLOPE - Fisher is trending up or down over the last three bars. Confirms the move is still in motion.
-
REGIME - Fisher is above or below zero. Slowest, most stable context.
-
ZONE - Overbought, oversold, or extreme. Context for whether a reversal is likely.
Each can be toggled on or off. Each contributes a fixed weight to the confidence score when enabled.
Confluence, not coincidence
A single layer firing is one opinion. Two or more agreeing is a decision.
The confidence meter aggregates the enabled layers into two independent scores; one for BUY alignment, one for SELL alignment. The weights are fixed and disclosed:
| Layer | Weight |
|---|---|
| SIGNAL | 30 |
| PROXIMITY | 25 |
| SLOPE | 20 |
| REGIME | 15 |
| ZONE | ±10 / ±5 |
The two scores do not sum to 100. Both can be high when the layers disagree, which is exactly the moment to stand aside. Both low means no edge. The meter shows this instantly.
Entry logic: OR or AND
-
OR — any enabled layer triggers. More trades, earlier entries.
-
AND — all enabled layers must agree. Fewer trades, tighter filter.
A single input toggles between them. No code editing, no parameter rebalancing.
The HTF filter — five modes, one input
The EA reads one column across three higher timeframes and votes 2-of-3 or 3-of-3.
| Mode | What it filters on | Character |
|---|---|---|
| SIGNAL | Crossover consensus | Fastest |
| PROXIMITY | Cross-imminent consensus | Earliest |
| SLOPE | Trend consensus | Balanced |
| REGIME | Sign consensus | Slowest |
| ZONE | Extreme gate only — no directional filter | Context only |
BULLISH consensus allows BUY, blocks SELL. BEARISH does the opposite. NEUTRAL allows both. In ZONE mode the filter stops being directional and becomes a gate: extreme overbought blocks BUY, extreme oversold blocks SELL.
Event mode vs State mode — one switch
| Mode | Behaviour |
|---|---|
| STATE-DRIVEN | Re-enters on the next bar if signals stay aligned. Pyramids into a running trend. |
| EVENT-DRIVEN | One entry per signal transition. Waits for the state to flip before entering again. |
Both modes respect every filter — HTF, ZONE gate, spread, session, position limits, staleness, proximity. The switch changes how often entries occur, not whether they are filtered.
Trade lifecycle management
The EA runs the position from open to close, not just at entry.
-
Take profit and stop loss — set at entry, enforced against broker minimum distance.
-
Breakeven — moves the stop to entry plus a buffer once profit reaches a threshold. Can be set as a percentage of TP or a fixed number of points.
-
Trailing stop — activates at a defined profit level and trails by a fixed distance with a step filter to prevent churn. Can be held back until breakeven is reached.
-
Partial close — closes a portion of the position at defined profit levels. Can be a fixed number of points or a percentage of TP.
-
Reversal — when an exit signal fires and the opposite direction is now valid, the EA closes and reverses in one operation.[Subject to Z mode; if False only closes trades in profit, if True closes trades regardless of profit check]
Every management step respects the broker's stops level and freeze level.
Risk management
-
Position limits — separate caps for BUY and SELL. Set either to zero to run long-only or short-only. Set both positive for two-way operation.
-
Lot sizing — fixed lots, percent of balance risked against the stop, or a fixed percent of equity. The chosen method scales the lot automatically.
-
Margin safety — before every order, the EA checks that required margin stays within a fraction of free margin. If it does not, the lot is reduced to fit, or the trade is skipped. No margin-call-by-EA.
-
Spread filter — blocks entries when the spread exceeds a threshold. The threshold is yours to set.
-
Session filter — restrict trading to chosen hours, or leave it open 24/7.
-
Entry spacing — optional minimum distance between same-direction entries, to prevent clustering.
The panel — what the trader sees
One panel, read top-down.
-
Multi-timeframe table — the five layers for each of the three HTFs, colour-coded, with ages.
-
HTF consensus — current vote.
-
Configuration summary — what is enabled, in what logic.
-
Status line — one of three states, shown only when relevant:
-
⛔ TRADE BLOCKED: <reason> (red)
-
✕ TRADE CLOSED: <reason> (sky-blue)
-
✅ TRADE ALLOWED (green)
-
-
PROXIMITY footer — names the timeframes where a cross is imminent. Shown only when applicable.
-
Confidence meter — two independent bars for BUY and SELL alignment.
Nothing is shown unless it means something. The panel is quiet when nothing is happening and informative when something is.
The PROXIMITY layer — the edge
The classic Fisher crossover plots on the closed candle. By the time it appears, the move has begun.
PROXIMITY measures whether Fisher is closing on its Trigger. It detects the approach and reads the direction from the sign of the gap:
-
Fisher above Trigger and closing → down-cross imminent → SELL precursor
-
Fisher below Trigger and closing → up-cross imminent → BUY precursor
The distance is normalised by Fisher's own recent volatility, so the same thresholds work on gold, FX, indices, and crypto without re-tuning. Two safeguards filter false starts:
-
A touch band — when the gap is inside it, the state goes neutral because direction is ambiguous.
-
A minimum convergence count — require N consecutive converging bars before reporting.
The result is a signal that arrives before the plotted crossover without firing on every Fisher wiggle.
Configuration
Eight inputs cover the entire signal engine:
| Input | What it does |
|---|---|
| EventDrivenMode | State or event entry behaviour |
| EntryLogic / ExitLogic | OR or AND |
| UseEntry / UseExit | Which layers participate |
| HTFColumn | Which column the HTF filter reads |
| UseZoneGate | Extreme-zone block |
| ProximityTouchNorm / ProximityVeryCloseNorm | Touch band and footer threshold |
| ProximityATRLen / ProximityMinConvergeBars | Normalisation length and confirmation |
Everything else is risk, display, or notification, and all of it has a working default. The EA trades out of the box with no configuration change. Tuning is optional, not required.
What it does not claim
It does not predict price. It does not guarantee profit. It does not publish a win rate. It aggregates Fisher state across five independent properties, weights them, filters them, and acts when the aggregate meets your configured logic. Whether that produces an edge on your instrument and timeframe is a question only your own testing can answer.
Who it is for
The trader who already uses the Fisher Transform (or wants to learn to use it) and wants it to answer more than one question at a time. The trader who wants a multi-timeframe view without maintaining three charts. The trader who values a status line that says why the EA did nothing, as much as one that says what it did. The trader who wants the signal engine disclosed, weighted, and visible, not hidden behind a black box. More detailed how to use guide available on request upon purchase.
