Feature matrix: RSI, MACD, returns

easyMLacceptance —~20 min

Implement features(closes, window) — the feature row for the LAST bar: [ret1, momentum, vol] where ret1 = closes[-1] / closes[-2] - 1, momentum = closes[-1] / closes[-1 - window] - 1 (the return over exactly window bars), and vol is the population standard deviation of the last window one-bar simple returns.

Every model in the ai quants folder starts from a table like this — aligned, leak-free, one row per bar. Return the three values rounded to 4 decimal places.

example 1

input

closes=[100.0, 100.0, 100.0, 103.0, 102.0, 104.0], window=4

output

[0.0196, 0.04, 0.0157]

example 2

input

closes=[100.0, 98.0, 101.0, 103.0, 102.0, 104.0], window=4

output

[0.0196, 0.0612, 0.015]

constraints

window + 1 ≤ len(closes) ≤ 10_000
every close > 0

teacher · 2 hints · the first is free

↗ read the theory

the chapter this problem is cut from

▶ open in the workbench

run the same math on a real tape

●solution.py· utf-8 · spaces:4

from statistics import pstdev


def features(closes: list, window: int) -> list:
    """[ret1, momentum, vol] for the last bar, rounded to 4 dp."""
    ret1 = closes[-1] / closes[-2] - 1
    # TODO: momentum looks back one bar short — the reference bar is closes[-1 - window]
    momentum = closes[-1] / closes[-window] - 1
    rets = [closes[i] / closes[i - 1] - 1 for i in range(len(closes) - window, len(closes))]
    return [round(ret1, 4), round(momentum, 4), round(pstdev(rets), 4)]

ln 1, col 1tab indent · esc leave editor

python 3.13 · pyodide

+ 6 hidden on submit

input · keyword arguments to features()

closes=[100.0, 100.0, 100.0, 103.0, 102.0, 104.0], window=4

expected

[0.0196, 0.04, 0.0157]

the judge calls features() with exactly these inputs · outputs are compared to 6 decimal places

● standing bypython 3.13 in your browser · pyodide 0.28.3 · judged on the server to 6 dp⌘↵ run · ⌘⇧↵ submit

paper only · educational · not advice