# Carbon Cashmere Signal Accuracy – Per-Coin ML Model Backtesting

> Carbon Cashmere Signal Accuracy – Per-Coin ML Model Backtesting is a paid API for AI agents from api.carbon-cashmere.de, paid per call via x402, $0.1/call, status unknown (last checked 2026-09-14).

Returns historical directional accuracy metrics for ML signal models on a specific cryptocurrency, including per-model hit rates, edge over coin-flip in basis points, calibration checks, and confidence-bucket breakdowns across 7d/30d/90d windows.

## Facts

- Endpoint: GET https://api.carbon-cashmere.de/v1/signal-accuracy/:coin
- Price: $0.1/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-14
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/api-carbon-cashmere-de-1f0445ec
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_gCKSZgqexGUemUso3CPPi

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability api-carbon-cashmere-de-1f0445ec
```

Example prompt: Before I act on those ML trading signals, can you pull up the historical directional accuracy stats for Bitcoin — I want to see the per-model hit rates, edge over coin-flip in basis points, and confidence-bucket breakdown across the 7d, 30d, and 90d windows?

## When to prefer this

Use this endpoint when you need to audit or verify the trustworthiness of ML trading signals from Carbon Cashmere before acting on them, or when comparing model quality across coins. It is the verifiability layer specifically tied to Carbon Cashmere's own signal ensemble — prefer it over generic backtesting tools when working within that ecosystem.

## Known failure modes

- Unsupported coin symbol returns 404 or empty result
- Insufficient historical data for short windows on newly listed coins
- Stale model data if retraining pipeline is delayed
- Rate limiting or payment failure returns 402 or 429
- Model name not recognized if referencing deprecated versions

## How this service works

Historical directional accuracy of our ML signal models for a specific coin. Per-model hit rate, edge-vs-coinflip in bps, calibration check, confidence-bucket breakdown over 7d/30d/90d windows. Verifiability layer for signal endpoints. Informational only.

## Output

Returns a structured breakdown of historical ML signal accuracy for the requested coin: per-model hit rates, directional edge over random chance in basis points, a calibration quality check, and confidence-bucket performance breakdowns — all sliced across 7-day, 30-day, and 90-day lookback windows. Useful for verifying signal reliability before acting on recommendations.

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/api-carbon-cashmere-de-1f0445ec/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from api.carbon-cashmere.de](https://www.zero.xyz/host/api.carbon-cashmere.de/llms.txt)