# Delx Reliability Curve

> Delx Reliability Curve is a paid API for AI agents from api.delx.ai, paid per call via x402, $0.001/call, status unknown (last checked 2026-09-14).

Computes a calibration reliability curve by binning forecast probabilities against observed binary outcomes to assess forecast accuracy.

## Facts

- Endpoint: POST https://api.delx.ai/api/v1/x402/reliability-curve
- Price: $0.001/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-14
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/delx-reliability-curve-cfcb134f
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_2MiPVInDWHF4j8qsHL34p

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 delx-reliability-curve-cfcb134f -d '<json body>'
```

Example prompt: Check how well calibrated these forecasts are — I have 8 binary outcomes [0,1,1,0,1,0,1,1] and matching forecasts [0.1,0.9,0.8,0.2,0.7,0.3,0.85,0.75], split them into 5 calibration bins and show me the reliability curve.

## When to prefer this

Choose this endpoint when you need a lightweight, pay-per-call calibration curve computation without setting up a full ML evaluation library. Ideal for agents that periodically audit forecast quality, validate model outputs in production pipelines, or need to present calibration diagnostics without local compute dependencies.

## Known failure modes

- Mismatched array lengths between forecasts and outcomes — likely returns a validation error
- Forecasts outside [0,1] range — may produce undefined bin assignments
- Zero samples in a bin — bin may be omitted or show undefined hit rate
- Invalid bins count (zero or negative) — likely returns a parameter error
- Empty forecasts or outcomes arrays — returns empty points array or error

## How this service works

Reliability diagram points (mean forecast vs hit rate per bin). Call when plotting calibration without a notebook stack. Returns bins with mean_forecast and hit_rate as deterministic JSON for $0.001 USDC via x402 on Base. First-party local math only—no RPC, live market feed, API keys, storage, or mediagen. Advisory only; the caller owns capital, risk, and production controls.

## Output

Returns an array of reliability curve points, each containing the bin index, number of samples in that bin (n), the mean forecast probability for that bin, and the actual hit rate (proportion of positive outcomes). Follows the schema delx/util-reliability-curve/v1.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "bins": {
   "type": "integer",
   "description": "Number of calibration bins"
  },
  "outcomes": {
   "type": "array",
   "description": "Binary outcomes 0/1 aligned with forecasts"
  },
  "forecasts": {
   "type": "array",
   "description": "Forecast probabilities in [0,1]"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "points": [
   {
    "n": 2,
    "bin": 0,
    "hit_rate": 0,
    "mean_forecast": 0.15
   },
   {
    "n": 2,
    "bin": 3,
    "hit_rate": 1,
    "mean_forecast": 0.8500000000000001
   }
  ],
  "schema": "delx/util-reliability-curve/v1"
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/delx-reliability-curve-cfcb134f/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from api.delx.ai](https://www.zero.xyz/host/api.delx.ai/llms.txt)