World population · World · 5 years ahead

2030: the model assigns an 80 percent probability to 8,536,707,706 to 8,601,609,314 people

Point estimate 8,601,609,314 people from Holt (exponential smoothing with additive trend), against 8,215,424,893 people observed in 2025. holt beats last_value on out-of-sample MAE by 91.2 percent over 12 cutoffs (minimum 8, margin 5 percent).

ModeledMetric page

Distribution over the history

Fan chart

Solid: observed (people). Dashed and shaded: model point estimates and 80 percent intervals from empirical residual quantiles. Filled dot: the record on this page. Green or red dot: the realised value, inside or outside its interval.

Quantiles: p5 8,532,709,855 · p10 8,536,707,706 · p25 8,542,596,714 · p50 8,564,168,738 · p75 8,575,671,826 · p90 8,597,147,284 · p95 8,599,459,817
Source: empirical quantiles of 12 in-window 5-step residuals, in log space; no normality assumed

The record

Immutable fields

Key: world-population:WLD:5
Id: 40a65470-2795-4e83-8396-d0c1422c8022
Created: 2026-09-07 22:16 UTC
Data cutoff: 2026-09-07 22:16 UTC
Origin: 2025 = 8,215,424,893 (obs 137382)
Target period: 2030 (1826 days)
Point: 8,601,609,314 people
Interval: 8,536,707,706 to 8,601,609,314 at 80 percent
Model: holt models_v1.targets_v1
Baseline: last_value
Snapshot: a5ac17fb1e3f40abc6f27c01c332ccc95fa4b77fb8a966fd6c75bd7b6967b76d
Data mode: ingested_at
Status: active

Drivers

Statistical association, not causal

Holt (exponential smoothing with additive trend) uses only the target's own history; there are no driver contributions.

Outcome

Not yet realised

The observation for 2030 has not been ingested. The weekly forecast.evaluate job attaches it when it arrives; the record above does not change.

Scenario assumptions

What the forecast takes for granted

No structural break in world-population between the data cutoff and the target period.
The source keeps publishing the series on the same definition; a rebasing or redefinition invalidates the comparison.
The smoothed level and trend at the cutoff persist over the horizon.

Falsifiers

What would show this forecast wrong

If the first published value of world-population (WLD) for 2030 is below 8,536,707,706 people or above 8,601,609,314 people, the 80 percent interval is falsified.
The model assigns a 10 percent probability to a value below 8,536,707,706 people and a 10 percent probability to a value above 8,601,609,314 people; interval misses should occur about one time in five over many forecasts, and a run of misses well above that rate falsifies the calibration.
The point forecast is above the last observed value (8,215,424,893 people); a realised value at or below that level falsifies the direction call.

Competing models

Scorecard for this target and horizon

Model	Status	Cutoffs	MAE	RMSE	Direction	80% coverage	Skill vs last value
Holt (exponential smoothing with additive trend) models_v1.targets_v1	champion	12	35,742,300	41,925,467	100%	50%	91.2%
Drift models_v1.targets_v1	challenger	12	225,271,426	229,591,978	100%	33%	44.6%
Last value models_v1.targets_v1	baseline	12	406,812,060	408,023,901	0%	100%	0.0%
Linear trend models_v1.targets_v1	challenger	12	67,857,265	78,527,034	100%	58%	83.3%
Ridge regression on lags and drivers models_v1.targets_v1	challenger	12	37,327,110	45,684,477	100%	42%	90.8%

Model card

Holt (exponential smoothing with additive trend)

Purpose. Point and 80 percent interval forecasts of world-population (WLD) 1, 2, 5 years ahead. Owner: Forecast Lab (HV 3.0).

Method

Simple exponential smoothing with an additive trend. Level and trend smoothing parameters (alpha 0.95, beta 0.95) are chosen by grid search on one-step-ahead squared error inside the training window. The series is modelled in natural logarithms and transformed back, so intervals are asymmetric in level terms.

Uncertainty

Intervals are empirical: the model is refitted at earlier origins inside the training window, the realised errors at the same horizon are collected, and the interval is the 10th to 90th percentile of those errors around the point forecast. No normality is assumed.

Training and evaluation

Training window: 66 periods, 1960-01-01 to 2025-01-01. Evaluation: rolling-origin backtest over 12 cutoffs using pseudo real-time data (the earliest ingested value per period; no vintage archive exists for this series, so revisions between the cutoff and today are not reproduced).

Known limitations

Structural breaks after the cutoff are not modelled.
Driver contributions are statistical associations, not causal effects.
Promotion to champion requires beating the last-value baseline on out-of-sample MAE by at least 5 percent over the minimum number of cutoffs; otherwise the baseline stays champion.
Retraining: every scheduled forecast.run refits on the data then available; backtests re-run from /admin/forecasts.

Version chain

Supersedes

This is the first record for its key.