HKUDS/Vibe-Trading · error · ValueError

bootstrap_statistic needs confidence in (0, 1), got {confide

Error message

bootstrap_statistic needs confidence in (0, 1), got {confidence}

What it means

bootstrap_statistic requires confidence strictly inside (0, 1). Values of 0, 1, or outside give degenerate percentiles (min/max of the sample), so they are rejected before resampling.

Source

Thrown at agent/src/quantlib/timeseries.py:806

        n_bootstrap: Number of bootstrap resamples.
        confidence: Confidence level in (0, 1), e.g. 0.95 for a 95% interval.
        seed: Seed for the random generator; pass an int for reproducible output.

    Returns:
        Dict with keys ``point_estimate``, ``bootstrap_mean``, ``bootstrap_std``,
        ``ci_lower``, ``ci_upper`` (all float) and ``confidence`` (float, echoed).

    Raises:
        ValueError: If ``data`` is empty, ``n_bootstrap`` is below 1, or
            ``confidence`` is not strictly inside (0, 1).
    """
    sample = np.asarray(data, dtype=float).ravel()
    if sample.size == 0:
        raise ValueError("bootstrap_statistic needs a non-empty sample")
    if n_bootstrap < 1:
        raise ValueError(f"bootstrap_statistic needs n_bootstrap >= 1, got {n_bootstrap}")
    if not 0.0 < confidence < 1.0:
        raise ValueError(f"bootstrap_statistic needs confidence in (0, 1), got {confidence}")

    rng = np.random.default_rng(seed)
    n = sample.size
    # Resample one draw at a time. Materialising the whole (n_bootstrap, n)
    # index matrix would be ~160MB at the default 10000 draws over 2000 bars.
    bootstrap_stats = np.empty(n_bootstrap, dtype=float)
    for i in range(n_bootstrap):
        bootstrap_stats[i] = float(statistic_func(sample[rng.integers(0, n, size=n)]))

    alpha = 1 - confidence
    return {
        "point_estimate": float(statistic_func(sample)),
        "bootstrap_mean": float(np.mean(bootstrap_stats)),
        "bootstrap_std": float(np.std(bootstrap_stats)),
        "ci_lower": float(np.percentile(bootstrap_stats, alpha / 2 * 100)),
        "ci_upper": float(np.percentile(bootstrap_stats, (1 - alpha / 2) * 100)),
        "confidence": confidence,
    }

View on GitHub (pinned to 80ffdda44c)

Solutions

1. Pass a fraction strictly between 0 and 1, e.g. 0.95.
2. If your config stores percentages, divide by 100 at the call site (and assert 0 < value < 1 for percentages in (0,100)).
3. Reject 0.0/1.0 explicitly in config validation.

Example fix

// before
bootstrap_statistic(data, confidence=95)
// after
bootstrap_statistic(data, confidence=0.95)

if not 0.0 < confidence < 1.0:
    if 0.0 < confidence <= 100.0:  # percentage form
        confidence = confidence / 100.0
    else:
        raise ValueError(f"confidence must be in (0,1), got {confidence}")
bootstrap_statistic(data, confidence=confidence)

When it happens

Trigger: bootstrap_statistic(data, confidence=0.0 or 1.0 or 95); the classic mistake is passing a percentage (95) instead of a fraction (0.95).

Common situations: Config files storing confidence as 95 or 0.95 inconsistently; UI dropdowns returning percentages; refactoring from APIs that take alpha.

Related errors

• bootstrap_statistic needs a non-empty sample
• bootstrap_statistic needs n_bootstrap >= 1, got {n_bootstrap
• confidence must be in (0.0, 1.0), got {confidence}
• autocorrelation_test needs lags >= 1, got {lags}
• lags must be < the number of observations; got lags={lags} f

AI-assisted analysis of HKUDS/Vibe-Trading@80ffdda44c (2026-08-28). Data as JSON: /api/errors/375679e770942a01. Report an issue: GitHub.