Enkan Usdjpy Mean Reversion
- 专家
- 版本: 1.110
- 更新: 24 十月 2026
- 激活: 5
ENKAN USDJPY MEAN REVERSION
Precision entries. Adaptive baskets. Account-aware sizing.
Enkan seeks reversals after extended USDJPY price moves, using completed M5-bar signals and a maximum six-level basket. Version 1.110 introduces account-aware compounding as the Standard setting: reference base volume 0.06 lot with a gentler 1.3x volume progression. Entry signals, grid spacing and exit rules are unchanged. No external AI service, DLL, WebRequest or external indicator is required.
RECENT MT5 REAL-TICK TEST — NEW STANDARD
- Period: 1 August 2024 – 1 September 2026
- Broker/data: Phillip Securities Japan; USDJPY.ps01, M5
- History quality: 100% real ticks
- Initial deposit: JPY 1,000,000; leverage 1:25
- Compounding: reference base 0.06 lot, progression 1.3x, reference balance JPY 1,000,000
- Net profit: JPY 422,698 (+42.27% over the full period)
- Final balance: JPY 1,422,698
- Profit factor: 2.64
- Maximum equity drawdown amount: JPY 367,015
- Maximum relative equity drawdown: 30%
- Position trades: 506
These are historical Strategy Tester results for the selected sizing configuration, not live or forward results. The setting was selected from four sizing configurations on this same period; this is not an independent out-of-sample test. Maximum drawdown amount and percentage are separate statistics and may occur at different points. The previous fixed-0.04-lot version's long-period results do not represent this new Standard.
HOW ENKAN TRADES
Selective entries: completed USDJPY M5 bars are evaluated using SMA20 displacement, RSI14, Bollinger Z-score, ATR, trend strength, bar range and tick-volume activity. A qualifying extreme triggers a counter-trend entry on the next eligible M5 bar. New signals are limited to UTC 23:00–05:59. Actual fills depend on available quotes.
Adaptive baskets: the EA can add same-direction positions when price moves against the basket, up to six levels including the first entry. Initial spacing adapts to volatility and activity. Later spacing expands by 1.5x; the new volume progression is 1.3x.
Exits: with two or more levels, the complete basket seeks an exit approximately 2 pips beyond the volume-weighted average entry. A one-level position has no fixed take-profit. A common server-side stop is placed 750 pips from the initial entry-bar BID-open reference. A time exit is attempted after 720 hours; market closure or execution failure can delay it. New baskets are suspended from December 20 through January 7 UTC; existing baskets remain managed.
ACCOUNT-AWARE COMPOUNDING
Basket base lot = InpBaseLot × min(account balance, account equity) / InpReferenceBalance
This unrounded base is calculated at the start of a new basket and frozen until that basket closes. Deposits, withdrawals and other strategies' profits or losses affect the next basket, not the existing one. Broker minimums, maximums and lot increments affect executable volumes.
At JPY 1,000,000 balance and equity, reference balance 1,000,000 and a 0.01-lot step, the illustrative ladder is 0.06 / 0.08 / 0.10 / 0.13 / 0.17 / 0.22 lot — total 0.76 lot. This is an example at the reference capital, not a fixed maximum exposure. Future baskets can grow as capital grows. Minimum-lot rounding can prevent small accounts from scaling down proportionally. Compounding is not a fixed-percentage loss limit.
IMPORTANT FOR NON-JPY ACCOUNTS
InpReferenceBalance is in the account's deposit currency. There is no automatic currency conversion. The default 1,000,000 is intended for the JPY test account. On USD, EUR or another currency, manually set the reference to the equivalent of JPY 1,000,000 in that currency before use, then test the broker's actual lot step. Do not assume 1,000,000 is always interpreted as JPY.
SETUP AND KEY INPUTS
- USDJPY M5; one Enkan instance per account, symbol and Magic Number.
- InpSizingMode=1: compounding. 0 selects fixed sizing.
- InpBaseLot=0.06: reference initial lot, not an unconditional fixed order size in compound mode.
- InpReferenceBalance=1000000: reference capital in ACCOUNT CURRENCY; adjust for non-JPY accounts.
- InpLotMultiplier=1.3; InpMaxLevels=6.
- InpSpacingATRMultiplier=1.25; InpSpacingAcceleration=1.5.
- InpBasketTargetPips=2; InpTargetMinLevels=2; InpSoloTakeProfitPips=0.
- InpBasketStopPips=750; InpMaxHoldHours=720.
- InpDeviationATRThreshold=2.5; InpRSIThreshold=27.5; InpTrendMaxATR=4; InpShockMaxATR=3.
- InpGateBBZThreshold=3.8; InpGateDirectionalTrendThreshold=-3; InpRSIReplayBars=4096.
- InpMagicNumber=26090986; InpMaxDeviationPoints=20; InpAllowNewEntries=true; InpDiagnostics=true.
- InpProductOrderComment=MQL5:ENKAN_USDJPY.
- Broker time: Market default European DST, UTC+2 winter / UTC+3 summer. Fixed-offset mode is available, with InpBrokerUTCOffsetHours=9. The published Phillip test uses fixed UTC+9.
- InpPauseYearEndNewEntries=true; pause dates December 20 through January 7.
UPDATING FROM AN EARLIER VERSION
Update only when the old basket is completely closed and no related pending or unresolved orders remain. Let the previous version finish managing its basket. Do not change version, sizing or Magic Number mid-basket. Incompatible legacy recovery data is not migrated automatically.
MT5 may retain old inputs when an EX5 is replaced. Explicitly check base0.06, multiplier1.3, mode1 and the correct reference balance after updating. Do not load the old fixed-0.04 preset as the new Standard. Test on demo first and keep the terminal and connection running while a basket is open.
IMPORTANT RISK AND TEST CONDITIONS
Enkan uses grid / martingale-style averaging. Six levels limits the number of additions, not monetary loss. Persistent directional moves can increase exposure, margin requirements and floating losses together. Margin limits may prevent additions or cause liquidation before the intended exit. A 750-pip stop does not guarantee a maximum loss: gaps, slippage, reduced liquidity and broker conditions can produce worse fills.
The test uses historical variable spreads and broker conditions, including swap; no separate extra slippage penalty was added. Different fees, spreads, swap, server time, volume increments or execution can materially change results. The initial deposit is a test condition, not a guaranteed sufficient balance. Historical results do not guarantee future performance.
Basket targets, additions and time exits require the EA and terminal to remain operational. The server-side stop normally remains registered while offline; client-side management does not run while stopped.
Long-period sizing test
2016-03-07 to 2026-09-01, PhillipSecuritiesJP-PROD USDJPY.ps01 M5, initial JPY 1,000,000, leverage 1:25, 100% real ticks. The D compounding research build returned net profit JPY 2,030,059, final balance JPY 3,030,059, profit factor 2.40, maximum equity drawdown JPY 789,119 (30%), and 2,393 position trades. This is a historical sizing test, not live performance or an independent out-of-sample result. The production release adds restart-state persistence and account-currency reference-balance input. Older fixed-lot long-period reports used a different tick-data revision and are not a like-for-like comparison.
