Pair ratio scan

How do you find which pairs of coins have drifted furthest from their usual ratio? Run a z-score on every pair's ratio with request.pairs, then hand the matrix and a per-coin summary to the Screener with report.*.

indicator("Pair ratio scan")
array<string> m = request.universe(10, exclude_tags = "stablecoin")
int len = input.int(30, "Z-score window", minval = 5)

// Inside request.pairs, close is the row coin's close over the column's
matrix<float> z = request.pairs(m,
  (close - ta.sma(close, len)) / ta.stdev(close, len))
report.matrix("ratio z", m, z)
report.column("sum of z", m, cs.row_sum(z))
plot(array.size(m), "Members")

Run it from the Screener: choose it in the script picker and press Run. The report shows a row per coin with its summed z-score, and the Matrix section shows every pair. request.pairs and report.* are AmenoScript additions, so the script is AmenoScript and runs leak-safe.

How it works

• exclude_tags = "stablecoin" leaves out coins pegged to the dollar. The ratio of two of them barely moves, so its z-score would be noise.
• request.pairs(m, expr) runs expr on every ordered pair of the ten coins, 90 of them. Row i, column j is coin i over coin j.
• The z-score is how many standard deviations the ratio sits from its 30-day mean. A large positive cell says the row coin is rich against the column coin.
• Each pair keeps its own ta.sma and ta.stdev, from the first day both coins have a price.
• report.matrix hands on the whole matrix, and its history, for the Screener's Matrix section.
• cs.row_sum(z) sums each row: how rich a coin is against all the others at once.
• plot(array.size(m)) gives the chart something to draw. A script that only writes a report says so on the chart.

Traps

• The matrix is antisymmetric only roughly. The ratio j over i is 1 over i over j, so its z-score is close to, not equal to, minus the other cell.
• A pair is only a close. high, low, open and volume are errors inside request.pairs, and so is ta.atr.

See also

• Each and pairs
• Reports
• Correlation matrix