ichimoku-h4-m1-align/EXPERIMENTAL-NOTES.md

Experimental EAs

Strategies here are experimental — newer, less battle-tested than the two main alignment EAs, and shipped for research and demo testing. Backtest and forward-test on a demo account before risking capital. See the top-level Disclaimer.

All experimental EA files are prefixed experimental- so they're easy to tell apart from the two main builds at a glance. The MS-W1-D1 build (section 6) is an exception: it lives in the main README table but is new and unbacktested, so it's documented and tracked here alongside the experimental builds until it has earned main-build status.

---

1. H4-M1 PO3 Alignment EA

File: experimental-h4-m1-po3-ea.mq5 Magic number: 20260502

The same 6-timeframe (H4→M1) Ichimoku alignment entry as the main H4-M1 EA, extended with PO3 dealing-range location filters (the power-of-three price-level concept by Hopiplaka) for entry gating and tiered take-profits. Ichimoku decides when to trade; PO3 decides whether the location is worth it and how far to hold.

PO3 dealing ranges

A fixed grid of power-of-three price levels: every multiple of 3^n × InpPO3Unit. On gold with InpPO3Unit = 1.0 the base grid (3^4 = 81) is …3888, 3969, 4050, 4131…; a level whose multiple carries a higher power of 3 outranks its neighbours (3888 = 16 × 3⁵ is a 243-grade level, 4374 = 2 × 3⁷ a 2187-grade one). The dealing range containing price is floor(price / step) × step to that plus step; its midpoint is equilibrium, the lower half discount, the upper half premium.

• Bias filter (InpPO3BiasFilter): if the recent H4 extreme (InpPO3BiasBars bars, default 180) tagged or raided a major level (3^InpPO3BiasPower, default 729-grade, within InpPO3BiasTolFrac of the base rung) and price was rejected away from it, entries against that rejection are blocked until price reclaims the level or an opposite-side tag supersedes it.
• Room filter (InpPO3RoomFilter): an entry is skipped when the first strong level (3^InpPO3StrongPower, default 243-grade) ahead in the trade direction is closer than InpPO3MinRR × the ATR stop distance — no buying into a ceiling, no selling into a floor.
• Tiered take-profits: half of each order batch targets TP1, the nearest base rung worth at least InpPO3MinRR R; the other half targets TP2, the nearest strong level beyond TP1. Both are front-run by InpPO3BufferATR × ATR(M15) since price often stalls just short of a level. A tier with no qualifying level within InpPO3MaxRR R gets no TP — those orders stay runners managed by the M15 Kijun exit.

Entry alerts include the PO3 context, e.g. PO3 243[3888-4131] 39% discount | TP1 3890.12 TP2 runner. For non-gold symbols set InpPO3Unit to the instrument's convention (e.g. 0.0001 for 5-digit FX pairs so a 243 range spans 0.0243; 0.01 for a cents-based intraday grid on metals).

Exit logic

Orders that reached their PO3 take-profit close there. Everything remaining is closed when the M15 close crosses the M15 Kijun-sen against the trade's direction (long closes below the M15 Kijun, short closes above it). Independently of that signal exit, every position carries an ATR(M15) × InpATRMultiplier stop loss.

PO3 inputs

Parameter	Default	Description
InpUsePO3	true	Use PO3 dealing-range levels for take-profits and entry filters
InpPO3Unit	1.0	Price per PO3 unit (1.0 = whole dollars on gold; 0.0001 for 5-digit FX)
InpPO3BasePower	4	Base rung = 3^power units (4 → 81)
InpPO3StrongPower	5	Strong level = 3^power units (5 → 243)
InpPO3MinRR	1.5	Minimum reward:risk for a level to qualify as a TP (also the room-filter threshold)
InpPO3MaxRR	8.0	Levels beyond this R-multiple are ignored (order stays a runner)
InpPO3BufferATR	0.25	Front-run TP buffer = ATR(M15) × this
InpPO3RoomFilter	true	Skip entries without MinRR room to the next strong level
InpPO3BiasFilter	true	Block entries against a recent rejection off a major level
InpPO3BiasPower	6	Major level for the bias filter = 3^power units (6 → 729)
InpPO3BiasBars	180	H4 bars scanned for a major-level rejection
InpPO3BiasTolFrac	0.4	Rejection tag tolerance, as a fraction of the base rung

All other inputs (Ichimoku periods, risk protection, equity sizing, equity alert) are identical to the main H4-M1 EA.

Status & caveats

• Distinct magic number (20260502, vs. 20260501 for the base H4-M1 EA) so the two can run on the same account/symbol without colliding.
• The PO3 grid is a fixed, hand-picked price scale — it needs to be tuned per instrument (InpPO3Unit) and re-checked if a symbol's price range shifts materially over time.

---

2. H1-M1 Time-Theory Reversion EA

File: experimental-h1-m1-reversion-ea.mq5 Magic number: 20260722 (runs independently of the other EAs)

A mean-reversion EA — the logical opposite of the trend/alignment builds. It fades an over-extended H1 Ichimoku trend back to a flat Kijun, timed by Ichimoku time theory and triggered by lower-timeframe momentum or a liquidity-raid rejection candle. Run it on its own instance; don't mix it with the alignment logic on the same symbol (the two would fight each other).

Core idea

After price has trended away and stayed off the H1 Kijun for one of the Ichimoku time cycles (9, 17, 26 or 33 bars, ±2) while the Kijun has gone flat, the extended move is "due" to snap back. The flat Kijun becomes a magnet; the trade is taken toward it, against the prevailing trend.

---

Entry — all gates must pass (checked per new M1 bar, per symbol)

Evaluated in CheckReversion(). The trade bails the moment any gate fails.

#	Gate	Rule	Input(s)
1	Extension	Last H1 close is ≥ InpFarATRMult × ATR(H1) from the Kijun. Above ⇒ sell back down; below ⇒ buy back up (sets direction).	InpFarATRMult
2	Trend to fade	The reversion fires only against an established H1 Ichimoku trend: a sell needs a bullish H1 trend (close above the Kumo and Tenkan > Kijun); a buy needs the bearish mirror.	InpUseTrendFilter
3	Time theory	Consecutive H1 candles since the last Kijun touch (the "break away") must equal an Ichimoku cycle — 9, 17, 26, 33 ± tolerance. A "touch" = the Kijun inside a candle's high–low; the count resets to 0 on any touch, so streaks between cycles (13, 30, …) don't qualify, and a streak longer than every window is rejected too.	InpTimeCycles, InpTimeTol
4	Flat Kijun	Kijun's move over the last InpFlatBars H1 bars ≤ InpFlatATRMult × ATR(H1).	InpFlatBars, InpFlatATRMult
5	A trigger fires	Either trigger below (both configurable).	see below

Trigger A — M5 Kijun cross (InpUseM5Cross) A fresh M5 close cross of the M5 Kijun in the reversion direction (the H1 "breakout close" confirmation on the lower timeframe). For a sell: the prior closed M5 bar was at/above the M5 Kijun and the last closed M5 bar is below it.

Trigger B — Rejection candle raiding liquidity (InpUseRejection) The last closed H1 candle is a long-wicked, small-body candle (wick ≥ InpRejWickFrac of range, body ≤ InpRejBodyFrac) whose wick raids an unliquidated fractal swing and closes back inside it. Swing liquidity is mapped across three timeframes, checked Daily → H4 → H1; a raid of any one qualifies:

Timeframe	Bars scanned	Input
Daily	50	InpRaidBarsD1
H4	300	InpRaidBarsH4
H1	500	InpRaidBarsH1

• A swing point is a fractal high/low with InpSwingWing lower/higher bars on each side.
• Raid + reject: the H1 rejection wick pokes beyond the level and the close comes back inside it.
• Unliquidated (InpRequireUnraided, default on): no more-recent closed bar on that timeframe has exceeded the level since it formed — the resting liquidity is still there. Set a timeframe's bar count to 0 to disable it.

---

Exit & stop management

• Take profit — the H1 Kijun (the reversion target), fixed at entry, attached to the order. A setup whose Kijun is inside the broker's minimum stop distance is skipped.
• Initial stop (Stage 1) — the H1 signal candle's own extreme (its high for a sell, low for a buy) + InpSLBufferATR × ATR(H1), widened to the broker minimum if needed. A tight stop just past the rejection/raid wick ⇒ small risk.
• M15 Kijun trail (Stage 2) — once a closed M15 candle prints clearly beyond the M15 Kijun in the trade direction ("clearly" = ≥ InpM15ClearATR × ATR(M15) past it), the stop is moved to the M15 Kijun itself, padded by InpM15SLBufferATR × ATR(M15). It re-evaluates each new M15 bar — following the Kijun as it drifts — only ever tightens (a short's stop only moves down, a long's up), and never sits inside the broker minimum.

---

Risk sizing

Identical equity-scaled sizing to the alignment EAs — GetEquityRisk() picks order count and lot size from account equity; above $8000 RiskBasedLots() sizes so the initial stop risks InpHighEquityRiskPct% of equity across the batch. Because the initial stop is tight, the same % risk buys a larger position than a wide swing stop would.

---

Full setup at a glance (sell example)

Trend up (price above Kumo, Tenkan > Kijun)
   └─ price ≥ 2·ATR(H1) above a FLAT Kijun
        └─ 9 / 17 / 26 / 33 H1 candles (±2) since last Kijun touch
             └─ trigger: M5 Kijun cross down  OR  H1 rejection candle
                         raiding an unraided D1/H4/H1 swing HIGH
                  └─ SELL
                     • TP  = H1 Kijun
                     • SL  = H1 signal-candle high (tight)
                     • then M15 closes clearly below M15 Kijun
                          → trail SL to the M15 Kijun (padded)
                          → keep tightening each M15 bar as it drifts

Buy setups are the exact mirror (downtrend, price below the Kijun, swing lows, M5 cross up, M15 closes above the M15 Kijun).

---

Key inputs

Group	Parameter	Default	Purpose
Setup	InpFarATRMult	2.0	"Far from Kijun" threshold (× ATR H1)
Setup	InpUseTrendFilter	true	Only fade an established H1 Ichimoku trend
Setup	InpTimeCycles / InpTimeTol	9,17,26,33 / 2	Ichimoku time cycles (bars since last touch) ± tolerance
Setup	InpFlatBars / InpFlatATRMult	5 / 0.25	Flat-Kijun window and tolerance
Triggers	InpUseM5Cross	true	Enable the M5 Kijun-cross trigger
Triggers	InpUseRejection	true	Enable the liquidity-raid rejection trigger
Triggers	InpRejWickFrac / InpRejBodyFrac	0.55 / 0.35	Rejection candle wick/body shape
Triggers	InpRaidBarsD1/H4/H1	50 / 300 / 500	Swing-liquidity scan depth per timeframe (0 = off)
Triggers	InpSwingWing	2	Fractal half-width for raid swings
Triggers	InpRequireUnraided	true	Require resting (unliquidated) liquidity
Stops	InpSLBufferATR	0.10	Initial-stop padding (× ATR H1)
Stops	InpM15ClearATR	0.1	"Clearly beyond M15 Kijun" buffer (× ATR M15)
Stops	InpM15SLBufferATR	0.1	Trailed-stop padding beyond the M15 Kijun (× ATR M15)
Risk	InpATRPeriod	14	ATR period (H1 and M15)
Risk	InpMaxSpreadPoints	60	Spread filter (0 = off)
Risk	InpHighEquityRiskPct	1.0	% equity risked per trade above $8000

The equity/alert inputs (InpMinProfitTrigger, InpWithdrawProfitPct, InpCheckDay, InpResetBaseline, InpSendPush) are the same as the main EAs — see the README.

---

Status & caveats

• Not yet compiled/backtested here — needs an F7 compile in MetaEditor and a Strategy-Tester + demo run before live use.
• "Unliquidated" is judged at each timeframe's own resolution — a daily level is "untouched" if no later daily bar's high exceeded it; an intraday poke within the still-forming daily bar isn't captured at daily resolution.
• Trend filter vs. flat Kijun — compatible by design (a strong extension where price ran away and the Kijun flattened underneath), but if live setups get filtered out because the Kijun is still gently sloping, loosen InpFlatATRMult.
• M15 trail — the Stage-2 stop trails to the M15 Kijun (padded), re-evaluated each M15 bar and tightening only; it does not use M15 fractal swings.

---

3. M1-M5 Breakout Alignment EA

File: experimental-m1-m5-breakout-ea.mq5 Magic number: 20260717

A fast, 2-timeframe variant of the alignment idea: instead of requiring agreement across six timeframes down to H4, it only requires M5 and M1 to align, but demands each close clear the cloud by a minimum ATR distance so marginal breakouts that merely graze the cloud edge don't qualify. Designed for quicker, more frequent signals than the main EAs at the cost of a much shorter-term (and noisier) trend anchor.

Entry logic

Runs on every new M1 bar close, per symbol. CheckAlign() on each of M5 and M1 requires price and Chikou above/below Tenkan, Kijun, and the cloud — same rule table as the main EAs (see README) — plus a breakout-strength buffer: the close must clear the cloud by at least InpMinBreakoutATR × ATR(tf) on that timeframe (each of M5 and M1 uses its own ATR), not merely sit on the other side of it. Set InpMinBreakoutATR = 0 to disable the buffer and accept a bare cloud break. If the buffer is enabled but the ATR value isn't available, the signal is skipped entirely rather than trading unfiltered. A trade opens only when both M5 and M1 agree on direction and no position is already open on that symbol.

Exit logic

Every position carries an ATR(M1) × InpATRMultiplier stop loss (note: ATR is computed on M1 here, not M15 as in the main EAs). Once a trade is in profit by at least InpTrailActivateATR × ATR(M5), an ATR chandelier trailing stop takes over: the stop is re-computed every new M1 bar as highest high since entry − InpTrailATR × ATR(M5) for longs (lowest low + InpTrailATR × ATR(M5) for shorts), using the extreme of the M5 bar that is still forming so a peak is locked in before it retraces. The trail only ever tightens and never sits inside the broker's minimum stop distance. Its behavior is set by InpTrailMode: 0 disables the trail entirely (the original Kijun-only exit), 1 trails every profitable trade, and 2 (default) trails only when the market is choppy — ADX(M5) below InpChopADXLevel (default 22) — standing down in trending markets so the M5 kijun-cross close rides the trend. The M5 kijun-cross close remains as the final fallback exit for trades that never arm the trail.

Risk protection & equity sizing

Identical InpUseStopLoss / InpMaxSpreadPoints / InpHighEquityRiskPct risk protection and the same GetEquityRisk() equity-tiered position sizing as the main EAs — see README.

Inputs

Parameter	Default	Description
Symbols	GOLDm#	Comma-separated list of symbols to watch (up to 60)
Tenkan / Kijun / SenkouB	9 / 26 / 52	Ichimoku periods
Slippage	30	Maximum allowed slippage, in points
InpMinBreakoutATR	0.5	Min close distance beyond the cloud, in ATR multiples per timeframe (0 = off)
InpUseStopLoss	true	Attach an ATR(M1)-based stop loss to every entry
InpATRPeriod	14	ATR period, computed on M1
InpATRMultiplier	3.0	Stop distance = ATR(M1) × multiplier
InpMaxSpreadPoints	60	Max spread (points) to allow an entry; 0 disables
InpHighEquityRiskPct	1.0	% of equity risked per trade once equity exceeds $8000
InpTrailMode	TRAIL_CHOPPY	0 = off, 1 = always, 2 = choppy-only via ADX(M5)
InpTrailATR	2.0	Chandelier trail distance = ATR(M5) × multiplier
InpTrailActivateATR	1.0	Arm the trail once profit ≥ ATR(M5) × multiplier
InpADXPeriod	14	ADX period for choppy-market detection (M5)
InpChopADXLevel	22.0	ADX below this = choppy → trail on in auto mode

The equity/alert inputs (InpMinProfitTrigger, InpWithdrawProfitPct, InpCheckDay, InpResetBaseline, InpSendPush) are the same as the main EAs — see the README.

Technical notes

• Magic number: 20260717 — independent from the other EAs, so it can run alongside them without interfering.
• State recovery: SyncStateFromPositions() rebuilds per-symbol direction state from open positions filtered by magic number on every tick, same as the main EAs.
• Per-symbol M1 gating: each symbol only re-evaluates entry/exit logic once per newly closed M1 bar.

Status & caveats

• Shortest-timeframe anchor of all the EAs in this repo (M5, vs. H1 or H4 for the others) — expect more signals, more noise, and a stop/exit cadence tuned for fast moves rather than sustained trends.
• Not yet extensively backtested here — run it in the Strategy Tester and on demo before considering live capital.

---

4. M30-M1 Breakout Alignment EA

File: experimental-m30-m1-breakout-ea.mq5 Magic number: 20260723

A shorter-anchor clone of the main H1-M1 alignment EA. It runs the exact same 4-of-4 Ichimoku alignment entry and M5 Kijun exit, but drops the top timeframe: instead of aligning H1→M1 it aligns M30→M1. This is a trend/breakout alignment build (all timeframes must agree in one direction) — not a reversion EA.

Entry logic

Runs on every new M1 bar close, per symbol. CheckAlign() on each of M30, M15, M5, M1 requires price and Chikou above/below Tenkan, Kijun, and the cloud — the same rule table as the main EAs (see README). A trade opens only when all four timeframes agree on direction and no position is already open on that symbol. Because the highest anchor is M30 rather than H1, setups form and clear faster than the main H1-M1 build.

Exit logic

Every position carries an ATR(M15) × InpATRMultiplier stop loss. Once a trade is in profit by at least InpTrailActivateATR × ATR(M5), an ATR chandelier trailing stop takes over: the stop is re-computed every new M1 bar as highest high since entry − InpTrailATR × ATR(M5) for longs (lowest low + InpTrailATR × ATR(M5) for shorts), using the extreme of the M5 bar that is still forming so a peak is locked in before it retraces. The trail only ever tightens and never sits inside the broker's minimum stop distance. Its behavior is set by InpTrailMode: 0 disables the trail entirely (the original Kijun-only exit), 1 trails every profitable trade, and 2 (default) trails only when the market is choppy — ADX(M5) below InpChopADXLevel (default 22) — standing down in trending markets so the M5 kijun-cross close rides the trend. The M5 kijun-cross close remains as the final fallback exit for trades that never arm the trail.

Risk protection & equity sizing

Identical InpUseStopLoss / InpMaxSpreadPoints / InpHighEquityRiskPct risk protection and the same GetEquityRisk() equity-tiered position sizing as the main EAs — see README. ATR is computed on M15 (as in the H1-M1 EA), not M1.

Inputs

Parameter	Default	Description
Symbols	GOLDm#	Comma-separated list of symbols to watch (up to 60)
Tenkan / Kijun / SenkouB	9 / 26 / 52	Ichimoku periods
Slippage	30	Maximum allowed slippage, in points
InpUseStopLoss	true	Attach an ATR(M15)-based stop loss to every entry
InpATRPeriod	14	ATR period, computed on M15
InpATRMultiplier	3.0	Stop distance = ATR(M15) × multiplier
InpMaxSpreadPoints	60	Max spread (points) to allow an entry; 0 disables
InpHighEquityRiskPct	1.0	% of equity risked per trade once equity exceeds $8000
InpTrailMode	TRAIL_CHOPPY	0 = off, 1 = always, 2 = choppy-only via ADX(M5)
InpTrailATR	2.0	Chandelier trail distance = ATR(M5) × multiplier
InpTrailActivateATR	1.0	Arm the trail once profit ≥ ATR(M5) × multiplier
InpADXPeriod	14	ADX period for choppy-market detection (M5)
InpChopADXLevel	22.0	ADX below this = choppy → trail on in auto mode

The equity/alert inputs (InpMinProfitTrigger, InpWithdrawProfitPct, InpCheckDay, InpResetBaseline, InpSendPush) are the same as the main EAs — see the README.

Technical notes

• Magic number: 20260723 — independent from the other EAs, so it can run alongside them without interfering.
• State recovery: SyncStateFromPositions() rebuilds per-symbol direction state from open positions filtered by magic number on every tick, same as the main EAs.
• Per-symbol M1 gating: each symbol only re-evaluates entry/exit logic once per newly closed M1 bar; the weekly equity alert is gated on a new M30 bar (the highest timeframe present).

Status & caveats

• Shorter trend anchor than the main H1-M1 EA (M30 vs. H1) — expect more frequent signals and a faster exit cadence, at the cost of a noisier top-timeframe trend filter.
• Not yet extensively backtested here — run it in the Strategy Tester and on demo before considering live capital.

---

5. H4-M15 Alignment EA

File: experimental-h4-m15-align-ea.mq5 Magic number: 20260724

A trimmed clone of the main H4-M1 EA that keeps the same H4 top anchor but stops the alignment at M15 — it aligns H4, H1, M30, M15 and disregards M5 and M1. The idea is to keep the multi-hour trend context of the H4-M1 build while cutting out the two lowest, noisiest timeframes, so entries fire on a cleaner 4-timeframe agreement rather than waiting for a full 6-timeframe stack down to M1.

Entry logic

Runs on every new M15 bar close, per symbol. CheckAlign() on each of H4, H1, M30, M15 requires price and Chikou above/below Tenkan, Kijun, and the cloud — the same rule table as the main EAs (see README). A trade opens only when all four timeframes agree on direction and no position is already open on that symbol. Because M5 and M1 no longer have to line up, setups clear the entry filter sooner than the full H4-M1 stack.

Exit logic

Identical to the H4-M1 EA: all positions close when the M15 close crosses the M15 Kijun-sen against the trade's direction (long closes below the M15 Kijun, short closes above it). Independently, every position carries an ATR(M15) × InpATRMultiplier stop loss — the same SL logic as the main H4-M1 build.

Risk protection & equity sizing

Same risk as the H4-M1 EA — identical InpUseStopLoss / InpMaxSpreadPoints / InpHighEquityRiskPct risk protection and the same GetEquityRisk() equity-tiered position sizing (see README). ATR is computed on M15, exactly as in the H4-M1 EA.

Inputs

Parameter	Default	Description
Symbols	GOLDm#	Comma-separated list of symbols to watch (up to 60)
Tenkan / Kijun / SenkouB	9 / 26 / 52	Ichimoku periods
Slippage	30	Maximum allowed slippage, in points
InpUseStopLoss	true	Attach an ATR(M15)-based stop loss to every entry
InpATRPeriod	14	ATR period, computed on M15
InpATRMultiplier	3.0	Stop distance = ATR(M15) × multiplier
InpMaxSpreadPoints	60	Max spread (points) to allow an entry; 0 disables
InpHighEquityRiskPct	1.0	% of equity risked per trade once equity exceeds $8000

The equity/alert inputs (InpMinProfitTrigger, InpWithdrawProfitPct, InpCheckDay, InpResetBaseline, InpSendPush) are the same as the main EAs — see the README.

Technical notes

• Magic number: 20260724 — independent from the other EAs, so it can run alongside them (including the base H4-M1 EA) without interfering.
• State recovery: SyncStateFromPositions() rebuilds per-symbol direction state from open positions filtered by magic number on every tick, same as the main EAs.
• Per-symbol M15 gating: each symbol only re-evaluates entry/exit logic once per newly closed M15 bar (the lowest timeframe in the alignment set, vs. M1 in the H4-M1 EA); the weekly equity alert is still gated on a new H4 bar.

Status & caveats

• Fewer timeframes to satisfy than the main H4-M1 EA (4 vs. 6) and a coarser entry cadence (M15 bars vs. M1) — expect somewhat earlier, more frequent entries, without the fine M5/M1 timing confirmation.
• Not yet extensively backtested here — run it in the Strategy Tester and on demo before considering live capital.

---

6. MS-W1-D1 Alignment EA (and companion Python monitor)

File: ichimoku-ms-w1-d1-ea.mq5 Magic number: 20260806

The slowest, rarest member of the family. Instead of aligning down to M1, it requires only three timeframes — MS (monthly) → W1 (weekly) → D1 (daily), highest to lowest — to agree. It targets multi-week/month trend trades, so lower timeframes (H4 down to M1) are deliberately dropped: they would veto almost every valid signal. Expect a handful of signals per year per symbol.

Because the signal is so rare, the companion Python monitor below is the recommended way to watch for it (no VPS, no chart, no EA running) — use the EA itself for backtesting in the Strategy Tester and for automated execution once you trust the signal.

Entry logic

CheckAlign() on each of MS, W1, and D1 requires price and Chikou above/ below Tenkan, Kijun, and the cloud — the same rule table as the main EAs (see README). A trade opens only when all three timeframes agree on direction, no position is open on the symbol, and the live spread passes InpMaxSpreadPoints. The stack and gating:

Index	Timeframe	Role
0	MS (PERIOD_MN1)	Highest — trend anchor
1	W1	Intermediate
2	D1	Lowest — bar-gating and exit reference (default)

Exit logic

All positions close when the close crosses the Kijun-sen against the trade direction on InpExitTF (default D1; set it to W1 to give trends more room). Independently, every position carries an ATR(InpATRTF) × InpATRMultiplier stop loss — InpATRTF defaults to D1 (an M15 stop is meaningless for a multi-week hold).

Inputs

Parameter	Default	Description
Symbols	GOLDm#	Comma-separated list of symbols to watch (up to 60)
Tenkan / Kijun / SenkouB	9 / 26 / 52	Ichimoku periods
Slippage	30	Maximum allowed slippage, in points
InpUseStopLoss	true	Attach an ATR-based stop loss to every entry
InpATRPeriod	14	ATR period, computed on InpATRTF
InpATRTF	PERIOD_D1	ATR timeframe for the stop distance
InpATRMultiplier	3.0	Stop distance = ATR × multiplier
InpExitTF	PERIOD_D1	Exit when close crosses kijun on this TF (D1 or W1)
InpMaxSpreadPoints	60	Max spread (points) to allow an entry; 0 disables
InpHighEquityRiskPct	1.0	% of equity risked per trade once equity exceeds $8000
InpReentryCooldownSec	0	Min seconds after an exit before re-entering the same symbol

The equity/alert inputs (InpMinProfitTrigger, InpWithdrawProfitPct, InpCheckDay, InpResetBaseline, InpSendPush) are the same as the main EAs — see the README.

Technical notes

• Magic number: 20260806 — independent from the other EAs.
• State recovery: SyncStateFromPositions() rebuilds per-symbol direction state from open positions filtered by magic number on every tick, same as the main EAs.
• Per-symbol D1 gating: each symbol only re-evaluates entry/exit logic once per newly closed D1 bar (the lowest timeframe in the alignment set); the weekly equity alert is gated on a new D1 bar.
• Backtesting: test per-symbol in the Strategy Tester (set Symbols to a single symbol). Verify the monitor and EA agree on historical signals before trusting either.

Status & caveats

• Very rare signals by design — a few per year per symbol. Do not judge it by trade count; the edge is a small number of large multi-week/month winners.
• The strict chikou condition (chikou must clear the reference candle's high/low 26 bars back) filters heavily even during clean trends — expect the monitor to read unaligned on many days where price looks "obviously" trending.
• ATR computed on D1 by default; if backtests show stops being hit in normal pullbacks, raise InpATRMultiplier before loosening the exit timeframe.

Companion monitor (Python + GitHub Actions, no VPS)

A daily, free monitor that computes the exact same MS→W1→D1 alignment from independent daily OHLC data and pushes a Telegram message when it fires — no VPS, no chart, no EA needed. Located in monitor/, documented in the README.

• Symbols: BTC/USD, ETH/USD, XAUUSD, XAGUSD, US100, US30, EURUSD, GBPUSD, USDJPY, AUDUSD, USDCAD (edit monitor/config.py). The FX list is limited to common trending majors — high-volatility crosses like GBPJPY are deliberately excluded.
• Data: Yahoo Finance daily bars via yfinance; metals use the COMEX futures (GC=F, SI=F) as proxies for the XM spot symbols because Yahoo's spot symbols are delisted.
• Logic: monitor/ichimoku.py is a faithful port of CheckAlign() in the EA, including the chikou-offset handling — so the monitor and EA should agree on the signal.
• Dedupe: state/state.json remembers the last notified direction per symbol, so a signal that persists for weeks won't spam you daily. It only notifies on new alignments, direction flips, and clears.
• Scheduling: .github/workflows/ms-w1-d1-monitor.yml runs it daily at 22:30 UTC on GitHub Actions for free, persisting the dedupe state between runs as a workflow artifact. Needs TELEGRAM_BOT_TOKEN and TELEGRAM_CHAT_ID secrets.
• Status: the Python port was verified against real data (it correctly flags fully-aligned trends and clears partial ones) but has not yet been cross-checked against the EA's backtest output — validate both before relying on either for execution decisions.