No-calendar control, 2,190 training rows / 32 prediction rows

match ← All flat growth benchmarks

The same history and prediction rows as the country scaling cases with no events, isolating each implementation's total calendar cost.

Input

benchmark/inputs/v6/country-scaling-2190-batch-32.json: 2190 history rows and 32 forecast rows.

Output

Largest difference: 4.1e-12 (< 0.0001% of the input's range).

Every forecast row from both libraries.

Date	Effect Prophet	Python Prophet	Difference
2025-12-30	20.0121	20.0121	4.1e-12
2025-12-31	20.0121	20.0121	4.1e-12
2026-01-01	20.0121	20.0121	4.1e-12
2026-01-02	20.0121	20.0121	4.1e-12
2026-01-03	20.0121	20.0121	4.1e-12
2026-01-04	20.0121	20.0121	4.1e-12
2026-01-05	20.0121	20.0121	4.1e-12
2026-01-06	20.0121	20.0121	4.1e-12
2026-01-07	20.0121	20.0121	4.1e-12
2026-01-08	20.0121	20.0121	4.1e-12
2026-01-09	20.0121	20.0121	4.1e-12
2026-01-10	20.0121	20.0121	4.1e-12
2026-01-11	20.0121	20.0121	4.1e-12
2026-01-12	20.0121	20.0121	4.1e-12
2026-01-13	20.0121	20.0121	4.1e-12
2026-01-14	20.0121	20.0121	4.1e-12
2026-01-15	20.0121	20.0121	4.1e-12
2026-01-16	20.0121	20.0121	4.1e-12
2026-01-17	20.0121	20.0121	4.1e-12
2026-01-18	20.0121	20.0121	4.1e-12
2026-01-19	20.0121	20.0121	4.1e-12
2026-01-20	20.0121	20.0121	4.1e-12
2026-01-21	20.0121	20.0121	4.1e-12
2026-01-22	20.0121	20.0121	4.1e-12
2026-01-23	20.0121	20.0121	4.1e-12
2026-01-24	20.0121	20.0121	4.1e-12
2026-01-25	20.0121	20.0121	4.1e-12
2026-01-26	20.0121	20.0121	4.1e-12
2026-01-27	20.0121	20.0121	4.1e-12
2026-01-28	20.0121	20.0121	4.1e-12
2026-01-29	20.0121	20.0121	4.1e-12
2026-01-30	20.0121	20.0121	4.1e-12

Timing

Median of all measured repetitions.

Step	Effect Prophet	Python Prophet	Effect Prophet is
Fit	7.72 ms	18.4 ms	2.4× faster
Predict	200 µs	2.81 ms	14× faster
Save model to JSON	15 µs	2.42 ms	1.6e+2× faster
Load model from JSON	99 µs	5.02 ms	50× faster
Load and predict in a new process	166 ms	1.50 s	9.1× faster

Settings

The case declaration both libraries ran.

{
  "kind": "map",
  "comparison": {
    "kind": "equivalent-objective",
    "evidenceId": "country-calendar-public-v1"
  },
  "configuration": {
    "growth": "flat",
    "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
  }
}