GodzillaBTCGodzillaBTC v4 separates market-state classification from setup selection. TREND conditions use pullback/momentum logic, RANGE conditions can use a separately validated mean-reversion fade, and CHAOS forces WAIT. CONFIRMED, STANDARD, TACTICAL, EARLY PAPER, DEVELOPING and WAIT describe evidence quality—not guaranteed outcomes.
Trend-pullbacks retain larger +1.25R to +2R targets. RANGE_FADE setups can use +0.50R to +0.75R targets because their break-even hit rate is much higher. Every target profile is validated separately; DEVELOPING is watch-only and CHAOS/insufficient evidence remains WAIT.
GodzillaBTC does not treat every market as independent. The USD-direction guard suppresses duplicate long-USD or short-USD FX bets, then a second factor layer compares recent 4H return correlations across FX, gold and Bitcoin. Highly overlapping trade P&L is suppressed and at most three open paper signals are allowed across all markets.
A chronological holdout is converted to a Beta posterior. Publication requires strong probability that the setup clears its target-specific break-even hurdle plus a cost buffer, together with positive holdout expectancy. Separately, a rolling likelihood health heuristic watches the current model’s paper WIN/LOSS/AMBIGUOUS record and can pause new signals when evidence of model decay becomes strong enough.
The signal engine watches relevant high-impact macro releases for the currencies in each market. Critical releases such as rate decisions, CPI/inflation and major employment events use a wider blackout window; other high-impact events use a shorter one. During the blackout, a technical BUY/SELL setup is shown only as a paused bias. Earlier caution windows can keep the direction but reduce Kelly/tier risk. The engine does not assume that an “above forecast” or “below forecast” result is automatically bullish or bearish.
GodzillaBTC now records every published CONFIRMED, STANDARD and TACTICAL signal from v3.1 onward, then marks the fixed target/stop outcome as WIN, LOSS, TIMEOUT, AMBIGUOUS or CANCELLED. The public win rate uses only decisive WIN/LOSS outcomes so unresolved or ambiguous trades are not quietly counted as wins.
View wins, losses & win rateFor actionable signals, the API also returns a conservative Kelly-inspired risk cap derived from the target-specific historical probability and payoff. It is a position-sizing overlay only; it does not make the direction more accurate. Read the full 12-indicator and risk-model guide.
Core coverage includes EUR/USD, GBP/USD, USD/JPY, USD/CHF, AUD/USD, USD/CAD, NZD/USD, XAU/USD and BTC/USD. Current coverage includes EUR/USD, GBP/USD, USD/JPY, USD/CHF, AUD/USD, USD/CAD, NZD/USD, XAU/USD and BTC/USD. The terminal uses a primary market-data feed with provider fallback and last-good D1 snapshots; availability and data state are shown transparently.
Confirmed, Standard and Tactical signals require fresh enough market data plus target-specific historical quality gates. Developing bias is watch-only. Historical hit rates are conditional research statistics, not guarantees. See current validation evidence and limitations.