MetaTrader 5
Experts
지표
스크립트
라이브러리
보관함당사 팬 페이지에 가입하십시오
- 조회수:
- 90
- 평가:
- 게시됨:
-
이 코드를 기반으로 한 로봇이나 지표가 필요하신가요? 프리랜스로 주문하세요
프리랜스로 이동
Gold has a daily maintenance break on the CME. The first hour of trading after it rises more than chance explains, and it does so on every calendar year of an eleven-year sample — while every other hour of the day measures flat.
This Expert Advisor buys at that reopen, holds for a fixed time with a volatility-scaled server-side stop, and does nothing else. One trade per session. No averaging, no grid, no martingale, no recovery.
The number, and then the number after I found my cost model was wrong
Measured over 11 years of hourly data, entering at the reopen and holding two hours:
| as the research charged cost | at the cost the tester measured | |
|---|---|---|
| per trade | +3.34 bps | +1.60 bps |
| t statistic | 7.40 | 3.42 |
| win rate | 59.3% | 50.8% |
| profit factor | 1.63 | 1.30 |
| positive years | 11 of 11 | 10 of 11 |
The left column is what I would have published if I had stopped where most people stop. Then a real-tick backtest in MetaTrader's own Strategy Tester showed the true round-trip cost at the reopen is about 60 points, not the 19 my research had charged — because the M1 bar's spread field is a per-bar summary and understates the reopen spread by roughly three times.
The edge survives that correction, at less than half its original strength, and the win rate becomes a coin flip carried by a favourable payoff rather than by frequency. The right column is the real one, and it is the one this EA is built on.
What it is not
It is not a living, and the header of the source file says so before any parameter is described. On capital sized for a 20% drawdown it measures about +3.3% a year with 4.6% maximum drawdown, Sharpe 1.23, roughly 200 trades a year — which also means roughly one year in nine is negative. At 0.01 lots that is about 130 currency units a year, and that figure measures the position size, not the strategy.
A 2.5-year window of it, taken alone, gives t = +1.12 — not significant. A profitable two-year backtest of this proves nothing by itself; the eleven-year sample is what carries it.
What the result survived before I trusted it
Each of these could have killed it, and the reason they are listed is that a result nobody tried to break is not a result:
- It is not a gap. The recorded gap across the break averages −0.011 — the move happens with the market open.
- It is not a bar-construction artifact. Rebuilding the hour from independently fetched M1 bars agrees with the H1 series to 0.0000 on all 73 overlapping days.
- It is not outliers. The 1%-trimmed mean is +0.70 against a full-sample +0.75; the median is +0.48.
- It is not the spread normalisation that would be the obvious trap, since MT5 bars are bid-side and the reopen spread is wider: the spread stays flat across the whole hour, so it explains 0% of the effect.
- It does not need the exact reopen tick. Entering 15 minutes late still pays, 30 minutes late still pays, and it dies at 45.
- The weekly reopen is NOT part of it. Sunday measures +1.11 bps with t = 0.88 — indistinguishable from noise — against +3.90 bps and t = 8.32 for the daily ones. That is why InpSkipMonday exists and defaults to true.
Inputs, and why they are what they are
| Input | Meaning |
|---|---|
| InpEntryHour | The server hour of the reopen. Working in server time is deliberate: this broker's clock follows US daylight saving, so the reopen sits at a fixed server hour all year, while a UTC trigger would drift by an hour twice a year. |
| InpEntryWindowMin | Only enter within this many minutes of the hour. |
| InpHoldHours | Hours to hold before closing on time. |
| InpSkipMonday | Skip the weekly reopen. See the Sunday figures above. |
| InpLots | Fixed lot. No scaling, ever. |
| InpStopAtrMult | Stop = this × the mean hourly range. Default 2.0, not the 1.5 the first version used — see below. |
| InpRangeBars | H1 bars in the mean-range unit. 480 is about 20 sessions. |
| InpMaxSpreadPts | Refuse to enter if the spread is worse than this. |
On the stop multiplier. At the corrected cost, 1.5× gives +1.60 bps (t 3.42) while 2.0× gives +1.91 (t 4.07) and is hit on 1.3% of trades instead of 2.6%. Wider still keeps improving — and the stop stays, because "remove the stop and the backtest improves" is exactly the reasoning that makes a martingale look profitable. The stop is what bounds a tail event; the time exit bounds the rest.
Test it before believing any of the above
And test it with every tick based on real ticks. On other code of mine, the same rules over the same period produced −15,627 with 1-minute bar modelling and +79,000 with real ticks. A backtest that does not state its modelling mode is not telling you what it measured.
Trade Excursion Report: MFE and MAE of every closed trade, to CSV
Exports one CSV line per closed trade with the numbers a statement does not carry: the best unrealised profit the trade showed (MFE), the worst unrealised loss it survived (MAE), how much of the MFE was actually kept, how many bars it spanned, and how long it was held. Same measurement as the Trade Excursion indicator, written to a file instead of drawn, because the questions worth asking of this data are ones a chart cannot answer: does the capture ratio depend on the hour of entry? Are the trades I close early the ones I hold longest? Do losers on Mondays reach a different MAE than on Fridays? Those are one pivot table away once the CSV exists. Output is semicolon separated with a header row, in MQL5\Files; opens in any spreadsheet. Measurement limit: excursions come from bar highs and lows on the chosen timeframe; the path inside a bar is unknown. The bars column reports the resolution each trade got, and the summary counts single-bar trades separately instead of averaging them in.
Trade Excursion: how much of each winner you actually kept (MFE/MAE from your own history)
Your statement tells you what you KEPT. It does not tell you what you HAD. This indicator reads your own closed trades from the account history and draws, for each one, the full range it travelled: the best unrealised profit it ever showed (MFE) and the worst unrealised loss it survived (MAE). The gap between what a trade showed and what it paid is where two common problems hide, and neither is visible in a balance curve: winners closing far below their best point (the exit leaves money behind), and losers whose worst excursion barely passed the stop (the stop sits inside the noise). The number to look at is the CAPTURE RATIO: of the profit a winner showed at its best, how much did you take home? 90% says the exits are well placed; 30% says the market kept offering and the exit kept refusing. Limit, stated up front: excursions come from the bars each trade spanned, on the chart timeframe; the path inside a bar is unknown. Predicts nothing; measures trades that already happened.
OHLC Reality Check - at what stop distance does your backtest start lying?
Opens a virtual bracketed trade on every M1 bar, walks it forward on the real tick history, and reports for a sweep of stop distances how often 1-minute OHLC modelling would score the trade the wrong way round.
CGoldSymbol - find the broker's gold and size the position correctly on it
Detects the gold symbol whatever the broker calls it, reads the contract specification from the terminal instead of assuming it, and turns a risk in account currency into a lot size that is correct for that broker.