Methodology: How ChipEye Builds Its Data
Every number on ChipEye comes from public Binance market data and a documented model. This page explains what is measured, what is estimated, and where the limits are.
Data sources
- Universe: Binance USDT-margined perpetual contracts, ranked by rolling 24-hour quote volume. The scanner takes the top 200.
- Candles: Binance spot and futures klines. Chip detail loads up to 400 candles for the selected timeframe.
- Open interest: Binance futures open interest history, currently limited to about 30 days.
- Funding: Binance futures funding-rate history, aggregated per candle and normalized to an 8-hour equivalent.
- Market cap and radar quotes: CoinGecko for market cap, Binance for price, volume, volatility and BTC ratio.
Chip distribution model
- The high-low range of each candle is split into 160 logarithmic price bins. Log bins give low-priced and high-priced coins comparable resolution.
- Each candle's volume is distributed across its range with a triangular weighting centered on the typical price.
- Existing chips decay with turnover. The default assumes a 20-day average holding cycle; the detail page allows 5–120 days.
- The array is normalized to 100%, then profit ratio, average cost, 90% and 70% ranges, peaks and patterns are derived from it.
The result is an estimate of the current cost structure. It is not a shareholder register and cannot see OTC trades, internal transfers or wallet movements.
Peaks and pattern classification
A peak is a local concentration of chips above a minimum mass threshold. The scanner classifies the structure into low, mid or high single peak, dual peak, dispersed or multi-peak based on peak count, concentration and position within the loaded range. The classification is mechanical; it does not “know” the news or the fundamentals.
Volume profile
Volume profile uses the same price bins but accumulates traded value without decay. The point of control is the highest-value bin; the value area expands outward from the POC until it contains 70% of traded value. The window can be all loaded candles, from the range low, from the range high, or the last 90 candles.
Open interest distribution
OI increases are spread across the price range of the candle in which they occurred; OI decreases are removed proportionally from the existing distribution. This produces a 30-day estimate of recent entries that remain open. Because OI is netted, the distribution cannot separate longs from shorts; funding is used as a directional clue.
Funding normalization
Funding intervals vary by contract. ChipEye sums the funding settled inside each candle and converts it to an 8-hour equivalent so different contracts can be compared. The 24-hour average shown in the header is also normalized.
Structure bias score
The bias score adds fixed points for observable conditions:
- Price above or below the main peak, and the pattern type.
- Profit ratio extremes and distance from average cost.
- Position relative to the value area.
- Distance and mass of the nearest trapped and support peaks.
- 7-day OI and price combinations, funding extremes, and the location of recent OI entries.
Total ≥ 1.5 is labeled bullish, ≤ −1.5 bearish, and the middle is a range. When several range conditions apply, a strong score is downgraded to “weak”. The score describes structure; it is not a forecast or investment advice.
Refresh and caching
The 200-coin scan is recomputed server-side about every 10 minutes. Coin detail (candles, OI, funding) is cached for 60 seconds. Market data can be delayed, unavailable or revised by the exchange; the page shows the last update time.
Known limitations
- Exchange volume is not the same as unique holder cost. Wash trading and market-maker activity can distort the input.
- Newly listed coins lack enough history for the turnover model to be representative.
- Airdrop, private-sale and unlock supply is not in the model and can create selling pressure that the chart does not show.
- Futures-based chips move faster than spot-based chips because leverage amplifies turnover.
- Patterns and scores are descriptive, not predictive. Use them to structure a question, not to skip risk management.
ChipEye provides public-data organization and visualization. It does not provide investment advice, execute trades or manage funds.