The System Explained
How ETH Core AI Works
A plain-language breakdown of how the system reads the market, filters setups, and produces clear trade decisions.
The Scan Cycle
From Raw Data to Decision
1
Scan ETH Market Data
Every few minutes, ETH Core AI fetches live OHLCV data across 1m, 5m, 15m, 1h, and 4h timeframes. It calculates RSI, EMA, ATR, and identifies market structure — trend direction, higher highs/lows, and key levels.
2
Build Multi-Timeframe Bias
Each timeframe casts a vote: BULLISH, BEARISH, or MIXED. The votes are weighted and combined into an overall MTF alignment score. This tells the system whether conditions favor LONG, SHORT, or WAIT.
3
Detect Chart Patterns
The system scans for tradeable patterns: double bottom, double top, bull/bear flag, breakout, breakdown, reclaim, and rejection. A confirmed pattern becomes the trigger candidate.
4
Check External Context
BTC trend, funding rate, open interest, liquidity levels, Fear & Greed index, macro event countdown, news sentiment, and AI reasoning are all pulled in and scored.
5
Calculate the VPF Score
Five components combine into a single 0–100 VPF Score (Validated · Filtered · Executable): pattern strength (blended context + execution + raw pattern, max 30), level confluence (max 30), regime & structure alignment (max 25), derivatives delta (max 10), and session quality (max 5) — minus a conflict penalty when signals disagree. A tier badge is assigned: HIGH CONFIDENCE (85–100), GOOD (76–84), or BELOW THRESHOLD (<76). Anything below 76 is NOT_EXECUTABLE and filtered out automatically.
6
Run the Five VPF Gates
The signal must pass all five VPF gates for EXECUTABLE status: pattern strength (≥12/30), level confluence, regime & structure alignment, VPF Score ≥76/100, and risk-to-reward ≥1.3. The Regime & Structure gate runs three internal checks — pattern matches regime, MTF alignment is not NEUTRAL, and in a RANGING regime the pattern is at a zone boundary.
Macro & News runs as an independent veto layer — full blackouts hard-block execution; medium-impact events apply ADJUSTED MODE (score −15 and SL/TP widened 20%).
Open Signal Conflict is a second independent veto layer — an opposite-direction signal is blocked while a previous executable signal is still in EXECUTABLE / TRIGGERED / IN_ZONE state. The cooldown auto-releases the moment that earlier signal hits SL, TP1, TP2, or expires, so legitimate regime-flip trades aren't permanently delayed.
Any gate failure or veto marks the signal NOT EXECUTABLE.
Macro & News runs as an independent veto layer — full blackouts hard-block execution; medium-impact events apply ADJUSTED MODE (score −15 and SL/TP widened 20%).
Open Signal Conflict is a second independent veto layer — an opposite-direction signal is blocked while a previous executable signal is still in EXECUTABLE / TRIGGERED / IN_ZONE state. The cooldown auto-releases the moment that earlier signal hits SL, TP1, TP2, or expires, so legitimate regime-flip trades aren't permanently delayed.
Any gate failure or veto marks the signal NOT EXECUTABLE.
7
Produce the Trade Plan
If all gates pass, the system calculates and outputs: entry zone, stop loss, TP1, TP2, RR ratio, and a filter reason summary — ready for the trader to review.
8
Log and Track
Every scan result is logged with a unique trade ID. The outcome tracker monitors open trades and records TP/SL hits, max moves, and generates loss diagnoses for closed trades.
Key Concepts
Terms Every User Should Know
A triggered trade means the system detected a LONG or SHORT signal with a confirmed pattern. It does not mean the trade is safe to enter — it just means a directional setup was found. Triggered trades include both executable and non-executable signals.
Executable means all five VPF gates passed (pattern strength, level confluence, regime & structure alignment, VPF Score ≥76, RR ≥1.3) and no macro/news blackout or opposite-direction open-signal conflict is active. The system considers the setup ready to enter. Only executable signals should be used for live entries — non-executable signals show market context only.
Triggered win rate counts all trades that got a direction — including those that failed risk filters. Executable win rate counts only trades that passed all gates. Executable win rate is the more meaningful number because it reflects the quality of filtered decisions.
Direction (LONG/SHORT/WAIT) and execution readiness are separate judgments. The market may favor LONG, but if the entry trigger is not confirmed, the RR is too low, a news blackout is active, or the AI conflicts — the trade is blocked. Context is useful; execution requires full alignment.
Even a system with 40% win rate can be profitable with 2:1 RR. Trading with poor RR is one of the most common reasons traders lose over time even when they "win" individual trades. ETH Core AI requires minimum RR before marking any trade executable.
Yes. The VPF Score is one of five executable gates. Even with a HIGH CONFIDENCE (85+) score, if a macro blackout is active, a news event is imminent, or the AI reasoning directly conflicts, the trade will be blocked. A high score does not override safety gates.
The system analyzes trade outcomes and identifies failure patterns. When enough evidence accumulates, it generates a structured recommendation — a specific parameter change with expected impact. The trader reviews this in the dashboard and approves or rejects it. No change happens automatically.