Powered by

#API

Currently, the API does not implement authentication. For production or public-facing deployments, we strongly recommend integrating a secure method (e.g., API keys, JWT tokens) to control access.

#Base URL & Environment Variables

http://164.90.156.209/api/testing

#Delta Calculation

Delta is a volatility measure used to model price uncertainty over time. It represents the average absolute percentage change in price over a given window.

#Step-by-step:

Step 1: For each pair of consecutive average prices P[t] and P[t+1], compute the absolute percentage change:

delta_t = abs((P[t+1] - P[t]) / P[t])

Mathematically:

deltaₜ = |P[t+1] - P[t]| / P[t]

Step 2: Take a rolling average of delta_t over a window (default: averaging_range = steps / 2). This smooths out short-term noise:

rolling_delta = delta_series.rolling(window=averaging_range).mean()

#Endpoints

#POST /calculate-single-probability

#Description

Calculates the probability that a given asset's price will remain within a specified range (± threshold) for a certain number of time steps, using manually provided inputs.

Request

  • Methods: POST
  • URL :/calculate-single-probability

#Request Body (JSON)

FieldTypeDescriptionExample
currentPricefloatCurrent price of the asset.2100.25
deltafloatVolatility measure (absolute delta; capped at 0.05).0.0045
stepsintegerPrediction horizon (in minutes).120
rangefloatTarget range threshold (e.g., 0.005 for ±0.5%).0.005

#Response

#Success (HTTP 200)

{
  "probability_within_range": 0.6781,
  "lower_bound": 2089.75,
  "upper_bound": 2150.75,
  "confidence_interval_80": {
    "lower": 2092.11,
    "upper": 2159.30
  }
}
  • Probability_within_range: Predicted chance (in fraction) that the future price will be within the target range.
  • Lower_bound / upper_bound: The target range based on the current price and the given threshold.
  • Confidence_interval_80: 80% confidence interval computed from the underlying normal distribution.

#POST /calculate-current-market-probability

#Description

Computes the probability using live market data. This endpoint retrieves the latest price data, calculates the volatility measure, and then invokes the prediction algorithm.

#Request

  • Method: POST
  • URL: /calculate-current-market-probability

#Request Body (JSON)

FieldTypeDescriptionExample
stepsintegerNumber of minutes ahead to run the prediction120
rangefloatRange threshold (e.g., 0.005 for ±0.5%)0.005

#Response

The response format is identical to/calculate-single-probability.

#POST /test-historical

#Description

This endpoint runs the prediction model across an entire historical dataset. In addition to returning overall metrics (e.g., average probability, accuracy, interval scores), it includes a debug section with detailed prediction data for selected timestamps.

#Request

  • Method: POST
  • URL: /test-historical

#Request Body (JSON)

FieldTypeDescriptionExample
datasetstringName of the dataset CSV fileuniswap_eth_usdc_minute_2023-08-01_2024-07-31.csv
stepsintPrediction steps ahead (in minutes).120
rangefloatThreshold for the target range (e.g., 0.005 for ±0.5%).0.005

#Response

{
  "overall": {
    "processed_rows": 498189,
    "average_probability_within_range": 0.6473,
    "naive": {
      "total": 498189,
      "in_wins": 323268,
      "naive_accuracy_percentage": 64.89
    },
    "high_confidence": {
      "range_in_total": 39547,
      "range_in_wins": 30818,
      "range_in_accuracy_percentage": 77.93,
      "range_out_total": 50,
      "range_out_wins": 46,
      "range_out_accuracy_percentage": 92.00
    }
  },
  "static": {
    "total": 323268,
    "average_probability": 66.99,
    "naive_accuracy_percentage": 100.0,
    "cases_prob_over_80": 30818,
    "cases_prob_under_1": 4
  },
  "non_static": {
    "total": 174921,
    "average_probability": 60.52,
    "naive_accuracy_percentage": 0.0,
    "cases_prob_over_80": 8729,
    "cases_prob_under_1": 46
  },
  "debug": {
    "2023-12-28T09:55:00": {
      "initial_price": 2392.14,
      "future_price": 2402.47,
      "delta": 0.00074,
      "predicted_probability": 0.4480,
      "predicted_range": [2380.18, 2404.10],
      "target_range": [2380.18, 2404.10],
      "actual_in_range": true},
    "2024-03-25T20:27:00": {
      "initial_price": 3635.95,
      "future_price": 3606.98,
      "delta": 0.00046,
      "predicted_probability": 0.2429,
      "predicted_range": [3642.69, 3691.88],
      "target_range": [3617.77, 3654.13],
      "actual_in_range": false}
    // ... similar entries for other timestamps
  }
}

#Explanation

  • Overall: Aggregated metrics for the entire dataset (naive predictions, high‑confidence predictions, etc.).
  • Static / non_static: Breakdowns based on the actual behavior:
    • Static: Timeframes where the future price remained within the target range.
    • Non-static: Timeframes where the future price fell outside the target range.
  • Debug: Optional. Contains detailed prediction data (e.g., for timestamps “2023-12-28T09:55:00”, “2024-03-25T20:27:00”, etc.) for advanced debugging and analysis. Debug cells setup manually in the code by request.

#Datasets

The following historical datasets are available for testing and evaluating the prediction model. Each file contains minute-level price data for the ETH/USDC pair on Uniswap.

#1.uniswap_eth_usdc_minute_2023-08-01_2024-07-31.csv

Period: August 1, 2023 → July 31, 2024
Chain: Ethereum
Download: 📥 Google Drive Link

#2. uniswap_eth_usdc_minute_2024-12-31_2025-04-14.csv

Period: December 31, 2024 → April 14, 2025
Chain: Ethereum
Download: 📥 Google Drive Link

#Error Handling

If an error occurs, the API returns an HTTP 500 status with an error message in JSON format:

{
  "error": "Description of error."
}

Check the server logs for more details if you encounter issues.

#Examples:

Test 1: Request

{
    "dataset": "uniswap_eth_usdc_minute_2024-12-31_2025-04-14.csv",
    "steps": 240,
    "range": 0.01
}

Response

{
    "debug": {},
    "non_static": {
        "average_probability": 42.08178131360263,
        "cases_prob_over_80": 8789,
        "cases_prob_under_1": 9900,
        "total": 61679
    },
    "overall": {
        "average_probability_within_range": 0.5067226948098416,
        "high_confidence": {
            "range_in_accuracy_percentage": 74.21143745782108,
            "range_in_total": 34081,
            "range_in_wins": 25292,
            "range_out_accuracy_percentage": 68.0225367596537,
            "range_out_total": 14554,
            "range_out_wins": 9900
        },
        "naive": {
            "in_wins": 81552,
            "naive_accuracy_in_range_percentage": 56.93739483770971,
            "naive_accuracy_out_range_percentage": 43.06260516229029,
            "total": 143231
        },
        "processed_rows": 143231
    },
    "static": {
        "average_probability": 57.16938444660028,
        "cases_prob_over_80": 25292,
        "cases_prob_under_1": 4654,
        "total": 81552
    }
}

Test 2:

Request

{
    "dataset": "uniswap_eth_usdc_minute_2024-12-31_2025-04-14.csv",
    "steps": 120,
    "range": 0.005
}

Response

{
    "debug": {},
    "non_static": {
        "average_probability": 42.14056895283416,
        "cases_prob_over_80": 304,
        "cases_prob_under_1": 4645,
        "total": 77410
    },
    "overall": {
        "average_probability_within_range": 0.46652023281710875,
        "high_confidence": {
            "range_in_accuracy_percentage": 65.33637400228051,
            "range_in_total": 877,
            "range_in_wins": 573,
            "range_out_accuracy_percentage": 80.40505452657088,
            "range_out_total": 5777,
            "range_out_wins": 4645
        },
        "naive": {
            "in_wins": 66061,
            "naive_accuracy_in_range_percentage": 46.04484529974699,
            "naive_accuracy_out_range_percentage": 53.95515470025301,
            "total": 143471
        },
        "processed_rows": 143471
    },
    "static": {
        "average_probability": 51.938526356117066,
        "cases_prob_over_80": 573,
        "cases_prob_under_1": 1132,
        "total": 66061
    }
}