Session Breakout Research

EURUSD M5 Session Breakout: cost-research result

Profitability is not established. Across ten independently reset USD 10,000 segments selected for available EURUSD coverage from 2018 to 2026, the frozen base scenario closed 1,498 positions and lost USD 3,013.00 before unverified financing; the stress scenario lost USD 4,239.41. The pooled profit factors before unverified financing were 0.8759 and 0.8301. These are separate reset segments, not a continuous ten-year equity curve or a holdout.

How the corrected preset works

The preset forms an opening range from 08:00 to 09:00 London local time, then checks two completed M5 bars from 09:00 until before 12:00. London uses UTC in winter and UTC+1 during EU daylight saving time; this preset converts from a broker server clock of UTC+2 in winter and UTC+3 during US daylight saving time, never the computer clock. A close crossing above the range high plus its five-point buffer can request a buy; a close crossing below the range low minus that buffer can request a sell. On five-digit EURUSD, five points equal 0.5 pip. The source skips that evaluation while all twelve completed range bars are unavailable, and can evaluate again on a later M5 bar. It calculates the stop at the opposite range edge and a planned two-R target from the captured signal quote. A changed quote or delayed fill can change the filled R relationship; the evidence does not prove full source-quote or execution causality.

The preset risk is 0.25% of current equity. Volume uses the estimated stop loss plus the USD 6.50 per round-trip-lot sizing estimate, then floors conservatively to the broker volume step; a result below the broker minimum is skipped. It rejects a spread above 30 points. One filled entry per London calendar day consumes that EA symbol/magic quota; a rejected request does not. An open position owned by this EA on the same symbol blocks another request; on netting accounts, any same-symbol position blocks it, while a foreign-magic position on a hedging account does not. Under the chosen policy, there is no trailing stop, break-even move, or time exit; positions may remain open overnight, while tester-end closes at independent segment boundaries are reported separately.

The corrected preset sets CampaignSession=0. It removes the extra CampaignSession 2 overlap filter that was incompatible with the strict London-before-12:00 entry geometry, while retaining the London range and entry hours above. The correction is supported as a functional preset by five February 2024 accounting fixtures and a bounded exact-public-EX5 zero-fee ledger comparison. A final observed-entry audit checked all 2,996 campaign entries for weekday clock, daily quota, protective SL/TP and spread; all 1,498 base entries were within one point of two R at the logged request quote. All matched base/stress pairs retained identical SL/TP; 622 delayed stress entries differed by more than one point from their logged quote, up to 48 points. This is bounded observed evidence, not proof of the original calculation quote, M1 range/cross causality, risk sizing, restart behavior, or every order-request path.

The campaign applies USD 3.50 per filled lot-side with no added delay for base, and USD 7.00 per filled lot-side with 200 ms delay for stress. Its source value of USD 6.50 per round-trip lot is a stop-risk sizing estimate, not a second cash charge; cash scenarios therefore model USD 7 and USD 14 per round trip respectively. Price result, modeled cash withdrawal, and native tester swap are separate. MetaTrader 5 documents that execution price can change between request and execution. Historical financing remains unverified.

What the frozen campaign observed

Observed outcomes include a low win rate (475 winners and 1,023 losers before unverified financing) and meaningful modeled fees. Those observations are consistent with the negative aggregate outcomes, but this evidence does not isolate a causal explanation between range construction, crossover timing, costs, or other execution conditions.

Download the corrected London preset package or download the frozen 20-case facts (JSON). Extract the package, copy the original TFSessionBreakoutResearch_Sessions.ex5 to MQL5/Experts, and load corrected-london.set in Strategy Tester. The original executable remains tester-only and no account orders are part of these instructions. This research does not validate full risk handling, restart behavior, causal range/crossover claims, all order requests, partial fills, live execution, or future performance.

Session Breakout status

Frozen 20-case EURUSD M5 cost research: profitability is not established. Each base and stress scenario closed 1,498 positions; net before unverified financing was USD −3,013.00 and USD −4,239.41, with pooled PF 0.8759 and 0.8301.

The corrected CampaignSession=0 preset retains the London opening range and entry geometry. It is a tester-only research configuration. The public EX5 comparison is bounded to a selected February 2024 zero-fee ledger; the multiyear campaign uses a cost wrapper.

Library strategy 13: London Opening-Range Breakout has its own EA download and evidence. It uses M15 pending breakout triggers, range-size/spread filters, staged management and a 16:00 London exit, so it remains a distinct implementation.