From a market question
to a working API call.

Market comparison

Did ZEC finish ahead of BTC?

The example calls compare_relative_return using ZECUSDT and BTCUSDT on Binance spot. It compares each market’s own open-to-close return over non-overlapping 12-hour windows. It does not use the ZECBTC cross pair.

Relative return = ZEC return − BTC return, in percentage points. A positive value means ZEC finished ahead. A historical count is not a prediction of the next winner.

01 / Run it

Download the script, then run it in a terminal. There is nothing to install with pip.

curl -fSLO https://aipricepatterns.com/examples/beggy_zec_vs_btc.py
python3 beggy_zec_vs_btc.py

Default grid: 00:00–12:00 and 12:00–00:00 UTC. Prints a readable summary and the latest finished windows. Run python3 beggy_zec_vs_btc.py --help for all options.

{
  "jsonrpc": "2.0",
  "id": 2,
  "method": "tools/call",
  "params": {
    "name": "compare_relative_return",
    "arguments": {
      "base": "ZECUSDT",
      "versus": "BTCUSDT",
      "interval": "1h",
      "horizonBars": 12,
      "offsetBars": 0,
      "recentWindows": 6
    }
  }
}

POST https://aipricepatterns.com/api/mcp

The JSON above is the tool call after MCP initialization. The downloadable script performs the complete exchange. Switch to Full Python to copy everything.

03 / Know what comes back

Use --json for the full response. These are the useful blocks to start with:

brief

A readable summary of historical outcomes. Start here before interpreting individual statistics.

recentWindows

Finished windows with timestamped opens and closes, each market’s return, relative return and the observed winner.

requestedWindow

The specific round requested with --window-open, or an explicit unavailable result if it cannot be measured.

currentWindow

The round still in progress, bars elapsed and the side ahead so far. Its outcome is not final.

regime

Historical windows grouped by pre-open market conditions. Compares how wide relative returns were in calm, mid and wild regimes.

coverage / data_coverage

Measured history, source freshness, missing candles and limits. Inspect these before using observations in your application.