Methodology

This page answers one question: can you trust the number on screen, and within what limits. We describe sources and assumptions rather than hiding them behind the word “algorithm”.

The order book

We keep persistent connections to exchanges' public streams and write book snapshots to our own storage: every 5 seconds for recent history and a 5-minute aggregate for the long one. The band around price is configurable (±0.5%, ±2%, ±5%) — a wider band shows distant walls but coarsens the detail near price.

A limit that cannot be worked around

Exchanges do not publish order book history. Data accumulates forward only, from the moment a coin is added to collection. For any earlier period the book does not exist — not for us and not for anyone else. If a service shows you years of “order book history”, it is showing a reconstruction, not a recording.

Spot, not futures

The calculations are based on spot books from several exchanges. A futures book exists only for Binance. Some venues expose limited depth: Kraken returns at most 500 levels, about ±1.9% from price, so the ±2% and ±5% bands cannot be filled for it.

The heatmap

Each cell is the total volume of limit orders at that price level during that time interval. Brightness is normalised over the visible window rather than in absolute terms: otherwise one anomalous wall would black out the rest of the picture across a long history.

There is a stability filter: a level is shown not “as of the bar close” but at the size it held for at least a chosen share of the time. That suppresses the flicker near price, where a wall is eaten and refilled, and keeps the real slabs.

The liquidation map

This is a model, not a fact

Exchanges do not publish where participants' positions sit. We estimate liquidation levels from a leverage range (2–125x by default) and accumulated volume. The higher the leverage, the closer the computed level sits to price. A computed level coinciding with a real liquidation cascade is probabilistic, not guaranteed.

Buyer and seller pressure

The metric compares the current imbalance of limit buyers over sellers with that same coin's usual behaviour, not with an absolute figure — coins differ in scale. “Usual” is the median over the chosen normalisation window. The value is expressed in sigmas: 0 is business as usual, +2 means buyers are noticeably denser than normal.

For robustness against outliers we use robust statistics: the median and the median absolute deviation in log space. A plain mean and standard deviation will not do here — a single spike skews the scale for weeks.

There are two sources: candles (executed volume, full history available, impossible to fake) and the order book (intent — finer and leading, but accumulating forward only and vulnerable to spoofing). The source is switchable in the indicator settings, and the chart states honestly which one is in use.

Prices and candles

Quotes and candles come from exchanges' public interfaces. The daily candle closes on exchange time (UTC), not on the visitor's local time.

The macro calendar

Dates and values for US and EU macro indicators come from open sources (official statistics, central bank schedules, public calendars). “BTC reaction” is computed from daily Binance candles around the release date. For some indicators the release date is an estimate — this is marked on the event page itself.

If you find a discrepancy

Write to support@in-crypto.app with the coin, timeframe and time. We check it, and if the error is ours we fix it and note here what changed.