---
url: >-
  https://jer-k.github.io/effect-prophet/benchmarks/flat/country-se-flat-2190-batch-32.md
---

# Sweden with a smaller prediction batch

 [← All flat growth benchmarks](/benchmarks/flat)

The identical 2,190-row history and frozen fitted columns as the 512-row case, with only 32 future rows; input bytes are frozen separately.

## Input

[`benchmark/inputs/v6/country-scaling-2190-batch-32.json`](https://github.com/jer-k/effect-prophet/blob/main/benchmark/inputs/v6/country-scaling-2190-batch-32.json): 2190 history rows and 32 forecast rows.

## Output

Largest difference: **4.8e-10** (< 0.0001% of the input's range).

Every forecast row from both libraries.

| Date | Effect Prophet | Python Prophet | Difference |
| --- | ---: | ---: | ---: |
| 2025-12-30 | 20.0033 | 20.0033 | 9.6e-11 |
| 2025-12-31 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-01 | 23.0066 | 23.0066 | 3.7e-10 |
| 2026-01-02 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-03 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-04 | 20.0091 | 20.0091 | 7.5e-11 |
| 2026-01-05 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-06 | 19.9937 | 19.9937 | 4.8e-10 |
| 2026-01-07 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-08 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-09 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-10 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-11 | 20.0091 | 20.0091 | 7.5e-11 |
| 2026-01-12 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-13 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-14 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-15 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-16 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-17 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-18 | 20.0091 | 20.0091 | 7.5e-11 |
| 2026-01-19 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-20 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-21 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-22 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-23 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-24 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-25 | 20.0091 | 20.0091 | 7.5e-11 |
| 2026-01-26 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-27 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-28 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-29 | 20.0033 | 20.0033 | 9.6e-11 |
| 2026-01-30 | 20.0033 | 20.0033 | 9.6e-11 |

## Timing

Median of all measured repetitions.

| Step | Effect Prophet | Python Prophet | Effect Prophet is |
| --- | ---: | ---: | --- |
| Fit | 13.3 ms | 43.0 ms | 3.2× faster |
| Predict | 518 µs | 7.44 ms | 14× faster |
| Save model to JSON | 299 µs | 2.54 ms | 8.5× faster |
| Load model from JSON | 4.72 ms | 5.91 ms | 1.3× faster |
| Load and predict in a new process | 183 ms | 1.60 s | 8.8× faster |

## Settings

The case declaration both libraries ran.

```json
{
  "kind": "map",
  "comparison": {
    "kind": "equivalent-objective",
    "evidenceId": "country-calendar-public-v1"
  },
  "configuration": {
    "growth": "flat",
    "countryHolidays": {
      "country": "SE",
      "coverage": "reject-known-gaps",
      "priorScale": 10
    },
    "scaling": "absmax",
    "seasonalityMode": "additive",
    "holidaysMode": "additive",
    "changepoints": {
      "mode": "auto",
      "count": 25,
      "range": 0.8
    },
    "changepointPriorScale": 0.05,
    "seasonalities": [],
    "events": [],
    "regressors": []
  },
  "pythonOptimizer": {
    "algorithm": "Auto",
    "maxIterations": 10000,
    "newtonFallback": true,
    "sigFigs": 12
  }
}
```