How Acumen Works

A complete guide to the analysis methods, indicators, and pattern detection algorithms that power your trading insights.

The Analysis Pipeline

Every analysis in Acumen follows a four-stage pipeline. Raw market data (OHLCV candles) flows through technical indicator calculations, then pattern detection algorithms scan for known formations, and finally a multi-factor confidence scorer generates a weighted trading signal.

Market DataOHLCV CandlesIndicatorsEMA, RSI, MACDPatterns24 DetectorsTrading SignalConfidence + Direction

Technical Indicators

Indicators transform raw price data into actionable signals. Acumen calculates five core indicators in real-time, each revealing a different dimension of market behavior.

Exponential Moving Average (EMA) gives more weight to recent prices using a multiplier of 2 / (period + 1). Acumen uses EMA-21 (fast) and EMA-50 (slow).

Simple Moving Average (SMA) calculates the arithmetic mean of the last N prices. Used as the middle band in Bollinger Bands.

When EMA-21 crosses above EMA-50 ("golden cross"), it signals a potential uptrend. The reverse ("death cross") signals a potential downtrend.

EMA 21EMA 50Price

Pattern Detection

Acumen runs 24 pattern detection algorithms based on Thomas Bulkowski's "Encyclopedia of Chart Patterns"and quantitative research from NBER papers. Patterns are grouped into chart patterns (multi-candle formations) and candlestick patterns (1-3 candle signals).

Head & ShouldersBullish Engulfing

Chart Patterns (9)

Head & Shoulders
Bearish89%
Inverse H&S
Bullish89%
Double Bottom
Bullish78%
Double Top
Bearish76%
Ascending Triangle
Bullish76%
Descending Triangle
Bearish73%
Cup & Handle
Bullish72%
Bullish Flag
Bullish67%
Bearish Flag
Bearish67%

Candlestick Patterns (15)

DojiNeutral
HammerBullish
Inverted HammerBullish
Hanging ManBearish
Shooting StarBearish
Bullish EngulfingBullish
Bearish EngulfingBearish
Morning StarBullish
Evening StarBearish
Three White SoldiersBullish
Three Black CrowsBearish
Piercing LineBullish
Dark Cloud CoverBearish
Tweezer TopBearish
Tweezer BottomBullish

Support & Resistance Detection

Acumen identifies key price levels where buying or selling pressure has historically been strong enough to reverse price direction.

Step 1 — Pivot Detection: Scans for pivot highs and lows using a 20-candle lookback. A pivot high is a price point higher than all surrounding points within the lookback window.

Step 2 — Clustering: Nearby pivot points (within 2% threshold) are grouped together. The more touches a cluster has, the stronger the level.

Step 3 — Classification: Levels below the current price are classified as support; above as resistance. Results are sorted by proximity (closest first), limited to the top 5 each.

ResistanceSupport

Confidence Scoring System

The final trading signal is generated by a multi-factor weighted scoring system. Four independent factors are evaluated and combined into a single confidence percentage that determines signal strength (Strong, Moderate, or Weak).

Pattern 35%Volume 20%Trend 20%Indicators 25%Confidence Score
Pattern Confidence35%

How well the detected pattern matches its ideal geometric form, including depth, symmetry, and spacing.

Volume Confirmation20%

Whether volume behavior aligns with the expected pattern. Recent volume 1.2x+ above average increases the score.

Trend Alignment20%

Checks if price, EMA-21, and EMA-50 align. Price > EMA-21 > EMA-50 scores 0.9 (strong uptrend).

Indicator Confluence25%

Checks RSI zone (oversold/neutral/overbought) and MACD histogram direction and momentum.

Strong Signal

Confidence > 75%

Moderate Signal

Confidence 55% - 75%

Weak Signal

Confidence < 55%

AI-Powered Analysis

Beyond algorithmic detection, Acumen can send chart data to AI models (Claude 3.5 Sonnet, Grok) via OpenRouter for a second opinion. The AI performs two types of analysis:

Pattern Confidence Scoring

When a pattern is detected algorithmically, the AI reviews the same data and provides an independent confidence assessment, agreeing or disagreeing with the algorithm's finding.

Comprehensive Chart Analysis

Even without a detected pattern, the AI can analyze the full OHLCV dataset to identify trends, provide market outlook, and suggest entry/exit strategies with detailed reasoning.

Plain English

Explained Simply

No finance degree required. Here's what each concept means in everyday language.

What are Moving Averages?

Imagine tracking the average temperature over the last 3 weeks vs. the last 3 months. The 3-week average reacts quickly to a sudden heatwave, while the 3-month average barely budges. Moving averages do the same thing with prices. When the fast average crosses above the slow one, it's like the weather just turned warm — prices might keep going up.

What is RSI?

Think of RSI as a "tiredness meter" for price movements. If a stock has been going up for days straight, the meter reads high (overbought) — it might need a rest. If it's been falling non-stop, the meter reads low (oversold) — it might be ready to bounce. The sweet spot is in the middle.

What is MACD?

MACD is like comparing the speed of two runners. One runner represents the short-term trend, the other the long-term trend. When the fast runner starts pulling ahead, momentum is picking up. The histogram bars show how much gap there is between them — taller bars mean stronger momentum.

What are Bollinger Bands?

Picture a rubber band stretched around the price. When the market is calm, the band tightens. When it gets volatile, the band stretches wide. If the price touches the top of the band, it might snap back down. If it touches the bottom, it might bounce up. A tight squeeze often comes right before a big move.

What are Support and Resistance levels?

Think of support as a floor and resistance as a ceiling. If you bounce a ball in a room, it keeps hitting the floor (support) and ceiling (resistance). In markets, certain price levels act the same way — buyers step in at support and sellers step in at resistance. The more times a level is tested, the stronger it is.

What are Chart Patterns?

Chart patterns are like recognizing shapes in price data — similar to spotting animal shapes in clouds, except these shapes have been statistically proven to predict what happens next. A "Head & Shoulders" looks like a person's head between two shoulders and often signals a price reversal. These patterns work because thousands of traders see the same shapes and act on them.

How does the Confidence Score work?

Imagine four experts each grading a trade idea: one checks the pattern quality (35% weight), one checks trading volume (20%), one checks the overall trend direction (20%), and one checks other indicators (25%). Their weighted average becomes the final score. Above 75% is a strong signal, 55-75% is moderate, and below 55% is weak.

Disclaimer

Acumen is an educational and analytical tool. Pattern detection and technical indicators are based on historical statistical research and do not guarantee future results. Always do your own research and consider consulting a financial advisor before making trading decisions. Past performance does not indicate future results.

Ready to see it in action?

Launch Acumen