Verified Results (Revised): FXEA365 Tested All Strategies on MT5 with 132 Files — 5 Variants Passed (1 Core + 2 Support + 2 Martingale)

This is the FXEA365 operator.

This article is a revised edition correcting errors discovered on the same day as the original publication on 2026-05-28.

The original version reported "1 out of 25 passed (4%)," but after fixing a bug in our HTM report parsing script that miscalculated DD%, a recount revealed that 5 variants (1 core + 2 support + 2 martingale) actually passed live performance thresholds (PF≥1.5 / trades≥50 / DD≤30%). We are issuing this official correction.

How the Correction Came About

• During the initial verification, the Python parser failed to correctly extract the DD% from the MT5 Japanese locale format "残高最大ドローダウン: 594.58 (5.67%)". As a result, EAs that should have been filtered out by the DD≤30% criterion were incorrectly evaluated, leading to the conclusion that "only BLAZE GOLD v2 qualifies with PF≥1.5."
• Re-parsing with a separate implementation (UTF-16 LE + Japanese labels プロフィットファクター: / 残高最大ドローダウン:) revealed that AUSSIE BB REVERT v2.1 and EURJPY ASIA BREAKOUT 10-year edition had in fact passed with PF=1.52 / 50-79 trades / DD≤4%.
• Simultaneously, we confirmed that the martingale EA retest sessions (AUDUSD/EURJPY KAMIKAZE) had produced outstanding figures of PF=1.81/2.09 with 994/1731 trades.

Final Verification Results — MT5 Live Strategy Tester 132 HTM Files

Passed Live Thresholds (PF≥1.5 & trades≥50 & DD≤30% & net profit positive)

Failed (Major Categories)

Why These 5 Variants Passed

BLAZE GOLD v2 (XAUUSD)

• Entries only during the NY session at UTC 15:00, with an asymmetric design of SL=0.5 ATR / TP2=7.0 ATR
• Gold has low correlation with other assets and exhibits strong trends during NY hours — this is the source of the strategy's edge

AUSSIE BB REVERT v2.1 (AUDUSD)

• Trade frequency is low at 50 trades over 5 years, but with the v21_explicit settings (best value from the g2_04 grid), it achieved DD=2.92% / PF=1.52
• A Bollinger Band touch-reversal strategy optimized for AUDUSD's range-reversion characteristics
• ⚠ 50 trades is right at the lower limit for sample size; a 4-week demo forward test may only generate 2-4 trades

EURJPY ASIA BREAKOUT 10-year edition

• 10-year backtest from 2016-2025 with PF=1.52 / DD=4% / 79 trades
• Note: The 4-year version shows PF=2.96 / 38 trades, but insufficient sample size — the 10-year version was selected
• Captures EURJPY trend momentum as London session opens

Two Martingale EAs (AUDUSD/EURJPY KAMIKAZE)

• 994/1,731 trades over 4 years, with win rates of 85.81%/78.39% — typical of the high win rates seen in martingale strategies
• ⚠ Martingale strategies are designed for eventual account blowout. Even though the retest succeeded, the good results during the past 4 years were simply a tailwind from market conditions. Losses during sudden market shifts can be orders of magnitude larger
• Mandatory display of "Bankruptcy-risk strategy + daily withdrawal recommended" upon publication

Industry-Level Lessons (Unchanged from Original)

1. Python BT / author-published figures ≠ live MT5 results (verified across gogojungle, GitHub, and our own EAs)
2. "Thorough optimization" backfires (settings that passed 162K grid + WF + Monte Carlo performed worst on live testing)
3. EAs that work depend on asset-specific characteristics (BLAZE GOLD works on XAUUSD, AUSSIE BB on AUDUSD, EURJPY ASIA on EURJPY — all of them failed when applied to other assets)

Important New Lessons from the Correction

4. Parser bugs distort the entire evaluation — in this case, failure to extract DD% from a localized format led to the incorrect conclusion of "only 1 EA passes." Always cross-verify with an independently implemented parser.
5. Fewer than 50 trades means no reliable conclusion — AUSSIE BB's 50 trades and EURJPY ASIA's 79 trades have low statistical reliability; additional verification with 4 weeks of demo forward testing is mandatory
6. Never take past summaries at face value — summaries in notes or reports represent conclusions at the time of verification. When new parsers or new verification data emerge, re-evaluate immediately

FXEA365 Policy (Revised)

1. Expanded public lineup: In addition to the core BLAZE GOLD v2, AUSSIE BB REVERT v2.1 and EURJPY ASIA BREAKOUT 10-year edition are now published as live-verified EAs
2. Two martingale EAs reinstated: AUDUSD/EURJPY KAMIKAZE re-published under the bankruptcy-risk category
3. 4-week demo forward testing conducted simultaneously for all 5: core + 2 support + 2 martingale
4. Sales message corrected to: "We live-tested 132 HTM files from strategies around the world, and only 5 variants passed (1 core + 2 support + 2 martingale)"
5. Parser double-verification: Going forward, all verification data will be re-parsed by an independent implementation before finalizing figures

A Note to Our Readers: Correction and Apology

The original article stating "only 1 EA passed (4%)" may have been read by multiple people before we could correct it. That conclusion was an error caused by a parser bug. The actual live verification data shows that AUSSIE BB REVERT v2.1, EURJPY ASIA BREAKOUT 10-year edition, and the two martingale EAs all met the passing threshold. We are issuing this official correction.

When purchasing or using any EA, please continue to verify the following:

• MT5 Strategy Tester live data is publicly disclosed
• PF≥1.5 maintained across multiple periods (5yr/10yr)
• Forward-tested on a demo account for 4 weeks to confirm reproducibility
• At least 50 trades in the sample
• DD≤30% (recommended ≤15%)

Public Verification Data

• BLAZE GOLD v2 details: https://fxea365.com/ea/blaze-gold
• AUSSIE BB REVERT v2.1 details: https://fxea365.com/ea/aussie-bb-revert
• EURJPY ASIA BREAKOUT details: Coming soon
• Martingale EAs: https://fxea365.com/ea/ranking (martingale category)
• Forward test progress: https://fxea365.com/forward-test

FXEA365 — An MT5 EA distribution service differentiated by honest, empirical data. All figures in this article are live measurements from MT5 Strategy Tester across 132 HTM files.