Bitcoin Price Gauge – Logic and Methodology

This app answers: Is Bitcoin currently dipping or pumping relative to its trend? It does so by comparing price to trailing moving averages, and by ranking today’s deviation against history.

Horizons and Data Aggregation

• 30 days (short‑term): computed on hourly averages
• 365 days (long‑term): computed on daily averages

Historical raw ticks are downsampled to 10‑minute buckets for repository size, then aggregated to hourly/daily series for analysis and charts.

Data files in data/

• bootstrap_10min.csv – 10‑minute ticks used as the raw source to build hourly/daily series
• history_30d_hourly.csv – precomputed hourly series with derived metrics
• history_365d.csv – precomputed daily series with derived metrics

Columns in precomputed history files:

time,price,sma,multiple,percentile,volAdjPercentile

• time: ISO UTC timestamp for the sample (start of hour/day)
• price: aggregated average price at time
• sma: trailing Simple Moving Average (window matches the horizon; current period excluded)
• multiple: Price Multiple R = price ÷ sma
• percentile: percentile rank of current R within the winsorized historical R distribution (baseline since 2015‑01‑01)
• volAdjPercentile: percentile rank of z = log(R) ÷ rolling σ(log R) (σ window = SMA window), baseline since 2015, winsorized

Core Algorithms

1. Simple Moving Average (SMA)

• Window lengths: 30 days (hourly samples → 30×24), 365 days (daily samples → 365)
• Excludes the current sample when computing the trailing average

2. Price Multiple (R)

• R = price ÷ SMA
• Interprets the magnitude of deviation from trend (e.g., R = 1.10 ≈ 10% above trend)

3. Winsorization

• To reduce outlier impact, historical distributions are winsorized at the 1st and 99th percentiles before ranking

4. Percentile (Raw)

• Baseline population: historical R values since 2015‑01‑01 for the selected horizon
• Output: percentile of the current R within that winsorized distribution

5. Volatility‑adjusted Percentile

• Compute log(R) for each historical point
• Compute rolling σ(log R) over the same window as the SMA (30d or 365d)
• z‑score: z = log(R) ÷ rolling σ(log R)
• Rank today’s z within the winsorized z distribution since 2015
• Interpretation: makes different volatility regimes comparable

Labels

Labels are assigned from percentile bands of the selected analysis type (Raw or Vol‑adjusted):

• 0–10%: Extreme dip
• 10–20%: Very big dip
• 20–30%: Big dip
• 30–40%: Dip
• 40–50%: Small dip
• 50–60%: Around average
• 60–70%: Small pump
• 70–80%: Pump
• 80–90%: Big pump
• 90–100%: Extreme pump

API

GET /api/summary

Current analysis for both horizons. Includes price, SMA, R, raw percentile, historical average, counts over the recent window, and (when available) the latest vol‑adjusted percentile (via history cache).

GET /api/history

Historical series for both horizons. Prefers precomputed CSVs; if absent, computes on the fly from bootstrap_10min.csv.

Response shape (per horizon):

{
  horizons: {
    "365d": [ { t, price, sma, multiple, percentile, volAdjPercentile }, ... ],
    "30d":  [ { t, price, sma, multiple, percentile, volAdjPercentile }, ... ]
  }
}

GET /api/health

Simple health status.

Price Source

• Live price: Bull Bitcoin Index USD (fetched server‑side)
• Historical: repository CSVs listed above

Frontend Visualization

• Dual‑axis charts per horizon:
  • Price (left axis) and Multiple (right axis)
  • Price (left axis) and Percentile (right axis; Raw or Vol‑adjusted via UI)
• Interactive zoom presets (All time, 4 years, Last year, Last month) and synchronized panning/zooming
• Reference lines for current Multiple and current Percentile
• Download button to export the currently selected horizon’s history as CSV

Architecture

• Node.js + Express + TypeScript backend
• Singleton services:
  • DataService – loads CSVs, builds hourly/daily aggregates
  • AnalysisService – computes SMA, R, percentiles, and vol‑adjusted metrics
• Frontend: vanilla JS with Chart.js + chartjs‑plugin‑zoom

Repository Size Considerations

GitHub’s 100 MB file limit is respected by downsampling raw ticks to 10‑minute (bootstrap_10min.csv) and by publishing compact precomputed history files. The app prefers the precomputed files for fast startup and consistent results.

Disclaimer

This tool analyzes historical deviations from trend. It is not a prediction model and is not financial advice.