How 12 Indicators Build GodzillaBTC BUY & SELL Signals

See how EMA, RSI, MACD, ADX, ATR, Bollinger z-score, ROC, stochastic, slope, volatility, range position and breakouts combine with 1D/4H/1H validation.

GodzillaBTC does not use a single “magic” indicator. Its signal engine combines trend, momentum, volatility, market structure and historical outcome evidence, then checks whether the live quote is fresh enough to publish an actionable trade. The result can be CONFIRMED BUY/SELL, STANDARD BUY/SELL, TACTICAL BUY/SELL, DEVELOPING bias or WAIT.

The important detail is that the 12 indicators/features do not all vote equally. Some create directional evidence, some decide whether a trend is strong enough, some define volatility and stop distance, and some classify the market regime. Treating them as twelve identical votes would double-count related information.

Current architecture: 1D trend context → 4H setup engine → 1H timing check → technical quality gate → target-specific historical validation → fresh-quote check → Kelly-inspired risk overlay → forward signal tracking.

From market data to BUY, SELL or WAIT

GodzillaBTC refreshes recent 1D, 4H and 1H OHLC candles and a live quote.
The 4H series is the primary setup timeframe; 1D supplies higher-timeframe trend context and 1H helps with timing.
Twelve technical features are calculated from closed candles.
A directional score produces an initial BUY, SELL or neutral bias.
Technical gates test alignment, momentum, trend strength and overextension.
Historical cohorts test the exact target used by each signal tier versus a -1R stop.
Freshness rules reject actionable signals when the quote is stale or the market is closed.
A Kelly-inspired model sizes risk only after the signal qualifies.
Published signals are frozen into the forward track-record database.

The 12 indicators and market features

#	Feature	Role in the engine
1	EMA 20 / EMA 50	Defines medium-term trend structure. EMA20 above EMA50 supports bullish conditions; the reverse supports bearish conditions.
2	RSI 14	Measures momentum. The engine looks for momentum above or below a neutral zone rather than treating every overbought reading as an automatic sell.
3	MACD	Adds a second momentum/trend comparison using faster and slower exponential averages. Its sign helps confirm directional momentum.
4	ADX 14	Measures directional trend strength. Weak ADX can block a signal even when the raw direction looks attractive.
5	ATR 14	Measures recent trading range/volatility. ATR defines the risk unit used for the model stop and target spacing; it is primarily a risk feature, not a directional vote.
6	Bollinger z-score	Measures statistical distance from the recent mean. Extreme extension can reduce signal quality or block chasing a move that has already stretched too far.
7	ROC 14	Rate of Change measures the direction and speed of price movement over the lookback period and contributes to momentum agreement.
8	Stochastic 14	Locates the close within its recent high-low range. It helps detect stretched short-term conditions and mean-reversion pressure.
9	Normalized slope	Measures the direction of the recent price path. Slope alignment is especially important for STANDARD and TACTICAL signals.
10	Volatility percentile	Compares current realized movement with its recent distribution. It helps classify normal versus high-volatility regimes rather than forcing one rule set onto every condition.
11	Range position	Shows whether price is near the upper, middle or lower part of its recent range. It supplies market-structure context.
12	Breakout state	Checks whether the close is breaking a recent structural high or low. A breakout can strengthen directional evidence when other features agree.

Why 1D, 4H and 1H have different jobs

4H is the setup timeframe. Most directional scoring and the target-specific historical cohort are built from 4H bars when enough history exists. 1D is context. It helps prevent a medium-term signal from fighting a stronger higher-timeframe trend without justification. 1H is timing. It can strengthen or weaken the effective score when short-term momentum agrees or opposes the 4H setup.

This multi-timeframe design is why a market can show a bearish bias but still remain DEVELOPING: the 4H direction may be negative while 1D context or 1H timing is not yet aligned.

How individual features become a directional score

The score uses trend, momentum, mean-reversion and structure checks. EMA structure, RSI, MACD, ADX when trend strength is adequate, Bollinger z-score, ROC, stochastic, slope, range position and breakout state can contribute positive, negative or neutral evidence. ATR and volatility percentile have different jobs: they help determine risk distance and regime quality instead of simply adding another directional vote.

An effective score around ±4 creates a directional bias. Stronger evidence is required for the highest tier. The score by itself is not a signal: it still has to pass technical gates and historical target validation.

How the signal tiers differ

Tier	Target	What it demands	Risk cap
CONFIRMED	+2R	Strongest model score, 4H trend/slope alignment, 1D agreement, 1H timing support, adequate ADX and strong historical evidence.	Up to 1.00% account risk before the Kelly overlay reduces it further.
STANDARD	+1.5R	Balanced score, aligned 4H trend, at least two of three momentum checks, fresh data and positive target-specific expectancy.	Up to 0.75%.
TACTICAL	+1.25R	Aligned slope, momentum agreement and contextual support with a higher required target-hit floor because the target is smaller.	Up to 0.50%.
DEVELOPING	None	Directional bias exists, but at least one publication rule is missing.	0% — watch only.
WAIT	None	No sufficiently qualified directional setup.	0%.

Historical validation uses the target actually shown to the trader

GodzillaBTC does not validate every signal against one generic target. CONFIRMED is tested against +2R before -1R, STANDARD against +1.5R before -1R and TACTICAL against +1.25R before -1R. Comparable historical setups are grouped by direction, score strength and, when the sample permits, market regime. The engine reports sample size, target hits, stops, timeouts, average R and a smoothed target-hit estimate.

That historical estimate is not the same as the public forward track record. It is a qualification tool derived from past candles. Once a live signal is published, its real forward outcome is tracked separately.

Where the Kelly-inspired risk model fits — and what it cannot do

The Kelly Criterion is a position-sizing framework. It does not predict market direction and therefore does not make RSI, MACD or the BUY/SELL classifier more accurate. Its purpose is to scale exposure when a validated setup has an estimated edge.

For a simple win/loss payoff, full Kelly can be written as f* = (b·p − (1−p)) / b, where p is win probability and b is the payoff multiple. Full Kelly is often too aggressive when probabilities are estimated rather than known. GodzillaBTC therefore uses a Kelly-inspired conservative overlay: it blends the smoothed target-hit probability with part of the lower uncertainty bound, applies only one quarter of Kelly, shrinks further when the sample is small, and then applies a hard tier risk cap.

Example: if a STANDARD setup has a +1.5R target and only a modest edge, the Kelly overlay may recommend far less than the tier's 0.75% maximum. If the historical edge is not positive, the Kelly recommendation falls toward zero. The signal direction itself is unchanged.

Important: Kelly is only as good as the probability and payoff estimates fed into it. Estimation error can make full-Kelly sizing dangerously aggressive. Fractional Kelly and hard caps are risk controls, not guarantees.

Worked example: from raw conditions to a STANDARD BUY

EMA20 is above EMA50 and 4H slope is positive.
RSI is above the neutral zone, MACD is positive and ROC is positive: at least two momentum checks agree.
ADX is strong enough for the STANDARD gate.
The z-score is not so extreme that the engine considers price overextended.
1D context is supportive or, at minimum, 1D and 1H are not both fighting the 4H setup.
The historical STANDARD cohort has at least 20 comparable setups, a smoothed +1.5R-before--1R estimate above the required floor and positive average R.
The current live quote is fresh and the market is open.
The engine publishes STANDARD BUY with entry, ATR-based -1R stop and +1.5R final target.
The Kelly-inspired overlay converts the target-specific probability into a conservative maximum account-risk percentage; it does not change BUY into SELL.
The published signal is written to the forward database so its eventual WIN/LOSS/TIMEOUT can be independently displayed.

Why more indicators do not automatically mean more accuracy

Many indicators are transformations of the same price series and can be correlated. Adding another oscillator can create the illusion of confirmation without adding new information. GodzillaBTC therefore gives different features different jobs and uses historical target-specific evidence as a second gate. The goal is not to maximize the number of indicators or signals; it is to make each published signal explainable and testable.

Related GodzillaBTC resources

Research note: Kelly sizing is intended to manage capital allocation, not to improve forecast accuracy. Full Kelly can create severe drawdown when inputs are uncertain, which is why GodzillaBTC uses a fractional, capped variant. Further reading: Stanford: Kelly gambling and position sizing.