Perfectly constant

match ← All logistic growth benchmarks

Every value is exactly 50, halfway to the ceiling.

Input

benchmark/inputs/v2/logistic-constant-96.json: 96 history rows and 120 forecast rows.

Output

Largest difference: 7.9e-12.

Every forecast row from both libraries.

Date	Effect Prophet	Python Prophet	Difference
2020-01-01	49.9952	49.9952	1.2e-12
2020-01-02	49.9961	49.9961	1.2e-12
2020-01-03	49.9970	49.9970	1.3e-12
2020-01-04	49.9978	49.9978	1.3e-12
2020-01-05	49.9986	49.9986	1.3e-12
2020-01-06	49.9993	49.9993	1.4e-12
2020-01-07	49.9999	49.9999	1.4e-12
2020-01-08	50.0005	50.0005	1.5e-12
2020-01-09	50.0010	50.0010	1.5e-12
2020-01-10	50.0015	50.0015	1.5e-12
2020-01-11	50.0019	50.0019	1.6e-12
2020-01-12	50.0022	50.0022	1.6e-12
2020-01-13	50.0025	50.0025	1.7e-12
2020-01-14	50.0027	50.0027	1.7e-12
2020-01-15	50.0028	50.0028	1.7e-12
2020-01-16	50.0029	50.0029	1.8e-12
2020-01-17	50.0029	50.0029	1.8e-12
2020-01-18	50.0029	50.0029	1.8e-12
2020-01-19	50.0027	50.0027	1.9e-12
2020-01-20	50.0026	50.0026	1.9e-12
2020-01-21	50.0024	50.0024	2.0e-12
2020-01-22	50.0021	50.0021	2.0e-12
2020-01-23	50.0017	50.0017	2.1e-12
2020-01-24	50.0013	50.0013	2.1e-12
2020-01-25	50.0008	50.0008	2.1e-12
2020-01-26	50.0003	50.0003	2.2e-12
2020-01-27	49.9997	49.9997	2.2e-12
2020-01-28	49.9990	49.9990	2.3e-12
2020-01-29	49.9983	49.9983	2.3e-12
2020-01-30	49.9975	49.9975	2.3e-12
2020-01-31	49.9967	49.9967	2.4e-12
2020-02-01	49.9958	49.9958	2.4e-12
2020-02-02	49.9948	49.9948	2.5e-12
2020-02-03	49.9938	49.9938	2.5e-12
2020-02-04	49.9927	49.9927	2.5e-12
2020-02-05	49.9916	49.9916	2.6e-12
2020-02-06	49.9904	49.9904	2.6e-12
2020-02-07	49.9891	49.9891	2.6e-12
2020-02-08	49.9878	49.9878	2.7e-12
2020-02-09	49.9893	49.9893	2.8e-12
2020-02-10	49.9907	49.9907	2.8e-12
2020-02-11	49.9921	49.9921	2.9e-12
2020-02-12	49.9934	49.9934	2.9e-12
2020-02-13	49.9947	49.9947	3.0e-12
2020-02-14	49.9959	49.9959	3.1e-12
2020-02-15	49.9970	49.9970	3.1e-12
2020-02-16	49.9981	49.9981	3.2e-12
2020-02-17	49.9992	49.9992	3.2e-12
2020-02-18	50.0002	50.0002	3.3e-12
2020-02-19	50.0011	50.0011	3.4e-12
2020-02-20	50.0020	50.0020	3.4e-12
2020-02-21	50.0028	50.0028	3.5e-12
2020-02-22	50.0035	50.0035	3.5e-12
2020-02-23	50.0043	50.0043	3.6e-12
2020-02-24	50.0049	50.0049	3.7e-12
2020-02-25	50.0055	50.0055	3.7e-12
2020-02-26	50.0060	50.0060	3.8e-12
2020-02-27	50.0065	50.0065	3.9e-12
2020-02-28	50.0070	50.0070	3.9e-12
2020-02-29	50.0073	50.0073	4.0e-12
2020-03-01	50.0077	50.0077	4.1e-12
2020-03-02	50.0079	50.0079	4.1e-12
2020-03-03	50.0081	50.0081	4.2e-12
2020-03-04	50.0083	50.0083	4.2e-12
2020-03-05	50.0084	50.0084	4.3e-12
2020-03-06	50.0084	50.0084	4.4e-12
2020-03-07	50.0084	50.0084	4.4e-12
2020-03-08	50.0083	50.0083	4.5e-12
2020-03-09	50.0082	50.0082	4.6e-12
2020-03-10	50.0080	50.0080	4.6e-12
2020-03-11	50.0078	50.0078	4.7e-12
2020-03-12	50.0075	50.0075	4.7e-12
2020-03-13	50.0072	50.0072	4.8e-12
2020-03-14	50.0068	50.0068	4.9e-12
2020-03-15	50.0063	50.0063	4.9e-12
2020-03-16	50.0058	50.0058	5.0e-12
2020-03-17	50.0053	50.0053	5.1e-12
2020-03-18	50.0046	50.0046	5.1e-12
2020-03-19	50.0040	50.0040	5.2e-12
2020-03-20	50.0033	50.0033	5.3e-12
2020-03-21	50.0025	50.0025	5.3e-12
2020-03-22	50.0016	50.0016	5.4e-12
2020-03-23	50.0008	50.0008	5.5e-12
2020-03-24	49.9998	49.9998	5.5e-12
2020-03-25	49.9988	49.9988	5.6e-12
2020-03-26	49.9978	49.9978	5.7e-12
2020-03-27	49.9967	49.9967	5.7e-12
2020-03-28	49.9955	49.9955	5.8e-12
2020-03-29	49.9943	49.9943	5.8e-12
2020-03-30	49.9930	49.9930	5.9e-12
2020-03-31	49.9917	49.9917	6.0e-12
2020-04-01	49.9903	49.9903	6.0e-12
2020-04-02	49.9889	49.9889	6.1e-12
2020-04-03	49.9874	49.9874	6.2e-12
2020-04-04	49.9859	49.9859	6.2e-12
2020-04-05	49.9843	49.9843	6.3e-12
2020-04-06	49.9826	49.9826	6.4e-12
2020-04-07	49.9810	49.9810	6.4e-12
2020-04-08	49.9792	49.9792	6.5e-12
2020-04-09	49.9774	49.9774	6.6e-12
2020-04-10	49.9755	49.9755	6.6e-12
2020-04-11	49.9736	49.9736	6.7e-12
2020-04-12	49.9717	49.9717	6.8e-12
2020-04-13	49.9696	49.9696	6.8e-12
2020-04-14	49.9676	49.9676	6.9e-12
2020-04-15	49.9654	49.9654	7.0e-12
2020-04-16	49.9633	49.9633	7.0e-12
2020-04-17	49.9610	49.9610	7.1e-12
2020-04-18	49.9588	49.9588	7.2e-12
2020-04-19	49.9564	49.9564	7.2e-12
2020-04-20	49.9540	49.9540	7.3e-12
2020-04-21	49.9516	49.9516	7.4e-12
2020-04-22	49.9491	49.9491	7.4e-12
2020-04-23	49.9465	49.9465	7.5e-12
2020-04-24	49.9439	49.9439	7.6e-12
2020-04-25	49.9413	49.9413	7.6e-12
2020-04-26	49.9386	49.9386	7.7e-12
2020-04-27	49.9358	49.9358	7.8e-12
2020-04-28	49.9330	49.9330	7.8e-12
2020-04-29	49.9301	49.9301	7.9e-12

Timing

Median of all measured repetitions.

Step	Effect Prophet	Python Prophet	Effect Prophet is
Fit	17.8 ms	26.7 ms	1.5× faster
Predict	733 µs	2.95 ms	4.0× faster
Save model to JSON	24 µs	788 µs	32× faster
Load model from JSON	148 µs	2.38 ms	16× faster
Load and predict in a new process	156 ms	1.52 s	9.8× faster

Settings

The case declaration both libraries ran.

{
  "kind": "map",
  "comparison": {
    "kind": "equivalent-objective",
    "evidenceId": "logistic-python-reconciliation-v1"
  },
  "configuration": {
    "growth": "logistic",
    "scaling": "absmax",
    "seasonalityMode": "additive",
    "holidaysMode": "additive",
    "changepoints": {
      "mode": "explicit",
      "timestamps": [
        "2020-02-08T00:00:00.000Z"
      ]
    },
    "changepointPriorScale": 0.05,
    "seasonalities": [],
    "events": [],
    "regressors": []
  },
  "pythonOptimizer": {
    "algorithm": "Auto",
    "maxIterations": 10000,
    "newtonFallback": true,
    "sigFigs": 12
  }
}