Broader Context
Used to understand higher-timeframe market direction and conditions.
GOLDBOT
A Digital AssetOps Project
Automated trading. Built through data, not promises.
A research-driven XAUUSD automated trading system being developed for MetaTrader 5.
Built to analyse. Built to filter. Built to decide when NOT to trade.
CURRENT STATUS: RESEARCH & DEVELOPMENT
GoldBot is currently an experimental trading-software research project. It is not available for purchase or live-money deployment.
- Market
- XAUUSD
- Platform
- MetaTrader 5
- Current Build
- GoldBot V15.4 — Structural Timing Audit
- Development Phase
- Historical Research & Diagnostics
- Live Trading
- Not Active
What is GoldBot?
What we're building
GoldBot is an automated trading system being developed specifically around the XAUUSD gold market.
GoldBot is being designed to analyse multiple pieces of market information before deciding whether a trading opportunity is worth taking.
Instead of entering trades because of emotion, fear, greed or impulse, GoldBot follows coded rules.
Sometimes the best trade is no trade at all. GoldBot is being built to recognise that too.
- No emotion.
- No revenge trading.
- No chasing.
- No forced trades.
Decision pipeline
How GoldBot thinks
Each stage must be satisfied before the next one is considered.
- 01
MARKET REGIME
GoldBot first tries to understand the type of market it is operating in.
- Trending
- Ranging
- Lower volatility
- Higher volatility
- Unclear conditions
- 02
HIGHER-TIMEFRAME CONTEXT
Analyse broader market direction and structure.
- 03
MARKET STRUCTURE
Evaluate price behaviour and important structural areas.
- 04
LIQUIDITY
Look for information around liquidity behaviour and possible market sweeps.
- 05
ENTRY LOCATION
Determine whether price is actually in a location where a trade makes sense.
- 06
MOMENTUM & EXHAUSTION
Evaluate whether market movement supports or contradicts the potential setup.
- 07
ENTRY QUALITY
Multiple pieces of evidence are combined rather than relying on one indicator alone.
- 08
RISK CHECK
The potential trade must satisfy the system's risk requirements.
TRADE
Conditions satisfied
NO TRADE
Equally valid outcome
GoldBot is not being designed to trade constantly. Selectivity is part of the system.
Analysis layers
Multi-timeframe analysis
Strategy Analysis
Used for structure, volatility, regime information and broader setup analysis.
Decision Timing
Used for lower-timeframe decision and entry analysis.
Research philosophy
“Research proposes.
GoldBot tests.
Unseen data decides.”
GoldBot development follows a simple principle.
An idea is not accepted simply because it improves one historical backtest.
Indicators, filters, parameters and trading rules must justify their place through testing.
The aim is not to create the most complicated algorithm. The aim is to discover the smallest robust rule set capable of surviving different market conditions.
RESEARCH PROPOSES
Ideas and hypotheses are investigated.
GOLDBOT TESTS
The system measures what actually happens across historical market data.
UNSEEN DATA DECIDES
Changes ultimately need to survive data they were not designed around.
Transparency
We don't hide the bad tests.We investigate them.
Many trading systems are marketed by showing only their strongest periods.
GoldBot development takes the opposite approach.
Different years, market environments and volatility conditions are deliberately investigated.
If the strategy fails during a period, the objective is not to hide that period. The objective is to understand:
- What changed?
- What market condition caused the problem?
- Did the entry logic fail?
- Did the market regime change?
- Was the rule genuinely useful?
- Did a successful historical result simply come from overfitting?
Failure is data.
A bad backtest can teach us more than a carefully selected winning screenshot.
Build V15.4
Development Status
Current phase: Historical Research & Diagnostics.
The current research build is measuring momentum strengthening, stability and rollover characteristics while keeping the underlying trading rules unchanged.
Historical Research
ACTIVEStrategy Diagnostics
ACTIVERobustness Testing
UPCOMINGOut-of-Sample Testing
UPCOMINGWalk-Forward Testing
UPCOMINGExecution Stress Testing
UPCOMINGDemo Forward Testing
LOCKEDLive-Money Validation
LOCKEDCommercial Release
LOCKEDHistorical testing progress
- Historical test periods
- 2021–2025 + 2026 YTD
- Full years tested
- 5
- Baseline simulated trades
- 1,148
The 1,148 figure represents baseline simulated trades across the six historical test periods. It is not a live-trading result and should not be interpreted as profitability evidence.
Gold = active · Grey = upcoming · Locked = not yet reachable
Development pipeline
The road GoldBot has to earn
PHASE 1
Core Strategy Development
Building the trading engine and initial rules.
PHASE 2You are here → Historical Research & Diagnostics
Historical Diagnostics
Understanding where the strategy wins and loses.
PHASE 3
Robustness Testing
Testing whether results survive parameter changes and different conditions.
PHASE 4
Unseen Data Testing
Testing data that was not used to develop the rules.
PHASE 5
Demo Forward Testing
Allowing GoldBot to operate in current markets without risking real money.
PHASE 6
Controlled Live Validation
Only considered after earlier stages are successfully completed.
PHASE 7
Potential Early Access
Only considered if GoldBot earns its way through the previous stages.
Building in public
GoldBot Development Journal
A growing timeline of what is being researched, tested and changed.
Research Build
GOLDBOT V15.4
Structural Timing Diagnostic Audit
GoldBot V15.4 added diagnostic-only measurements designed to investigate whether GoldBot is sometimes entering otherwise valid directional moves too late. The audit measures bars since the relevant structural break, ATR distance travelled since the break, entry distance from the broken structure level, initial displacement strength, whether the broken level was retested, and entry timing classification. No trading rules, entry rules, scoring rules, stop-loss logic, target logic or trade-selection rules were changed.
Status: Diagnostic audit complete
Research Finding
V15.4 Structural Timing Audit — Complete
No universal “late entry = bad” rule was found.
Some historical periods showed substantially later entries performing worse, while other periods showed the opposite behaviour. This means structural timing cannot currently be justified as a simple standalone trade filter. The stronger research hypothesis is that structural timing becomes useful when interpreted alongside market context. Potential future combinations include structural timing + momentum state, structural timing + VWAP context, structural timing + EMA extension, structural timing + pullback/retest quality, and structural timing + structure location. These combinations have not yet been optimised or validated.
Status: Research finding
Research Build
GOLDBOT V15.3
Momentum & Exhaustion Diagnostic Audit
Diagnostic research examining momentum strengthening, stability and rollover behaviour while preserving the existing trading rules.
Status: Completed diagnostic audit
Entry Quality Research
Entry Quality Research
Investigating characteristics that separate stronger entries from weaker entries.
Status: Research ongoing
Market Regime Analysis
Market Regime Analysis
Investigating how GoldBot behaves across different market environments.
Status: Research ongoing
Robustness Programme
Robustness Programme
Parameter-neighbourhood, unseen-data and walk-forward testing.
Status: Planned
View Full Development History
Development archive coming soon.
Built in public.
GoldBot development is being documented rather than appearing online one day with an unexplained profit chart and a sales page. The aim is to create a visible development history covering:
- What was tested
- What worked
- What failed
- What changed
- What was rejected
- What is being investigated next
Watch the system evolve from research code to validated trading technology.
One trading engine.Different risk environments.
Future GoldBot architecture is being designed with multiple account environments in mind.
Planned ArchitectureNot Yet Validated
PERSONAL
For personally funded trading accounts.
Potential future controls may include:
- User-defined risk
- Account-level protection
- Maximum exposure controls
- Emergency shutdown controls
CHALLENGE
Future configuration intended for prop-firm evaluation environments.
Potential future protections may include:
- Daily drawdown restrictions
- Maximum overall drawdown limits
- Trading-session restrictions
- News-event controls
- Account-rule protection
FUNDED
Future configuration intended for funded-account environments.
Potential future architecture may prioritise:
- Capital preservation
- Strict drawdown management
- Exposure limits
- Rule compliance
- Risk consistency
GoldBot does not currently claim to pass any proprietary trading firm's challenge or funded-account requirements.
Risk comes before opportunity.
A trading algorithm should not only know when it wants to enter. It must also know how much it can risk and when trading needs to stop.
Planned protections — not yet complete
- Defined stop-loss protection
- Position-risk limits
- Exposure limits
- Daily-loss protection
- Maximum drawdown protection
- Trading shutdown controls
- Abnormal-condition protection
These are development priorities. Research-stage work and planned development are deliberately kept separate.
Survival first.
Opportunity second.
What GoldBot is NOT
- A guaranteed-income system
- A promise of passive income
- A 100% win-rate strategy
- A signal-selling Telegram group
- A martingale gambling system
- A finished commercial product
- A substitute for understanding financial risk
GoldBot has to earn the right to progress.
Evidence policy
Evidence before marketing.
When GoldBot reaches later development stages, results should be clearly separated into categories.
Category 1
Historical Simulation
Category 2
Out-of-Sample Testing
Category 3
Demo Forward Testing
Category 4
Live Validation
Never mix these categories. Historical simulations must never be presented as though they were live trading results.
Verified Performance Dashboard
Status: Locked until validation
Performance information will only appear here when the relevant testing stage has genuinely been completed.
Early access
Follow the GoldBot journey.
GoldBot is not currently for sale. Not available for purchase.
Follow development to receive future information about:
- Major development milestones
- Research findings
- New GoldBot versions
- Demo-testing progress
- Validation stages
- Potential future early access
GOLDBOT
A Digital AssetOps Project
Digital AssetOps is the parent technology brand behind GoldBot and future automation, research and digital-finance software projects.