VATOMETRS

Latvia’s political “vatnik” vote statistics
Satirical project. Not affiliated with the Central Election Commission or any government agency.

C26 · central specification

Methodology

Vatometrs is a deliberately simple weighted election model. It sums votes for party lists fully or partly included in the target political niche, applies the declared weights, and divides by the number of valid ballot envelopes.

C26 = SV/AJ + ST! + 0,25 × LPV + 0,20 × (Saskaņas Centrs + Gobzems)
Essential interpretation rule: coefficients are analytical weights, not a claim about what percentage of each party’s voters personally fits an ideological label.

2026 party-list weights

Party listWeightRole in modelNote
SUVERĒNĀ VARA / APVIENĪBA JAUNLATVIEŠI1,00coreFully included
Stabilitātei!1,00coreFully included
LATVIJA PIRMAJĀ VIETĀ0,25partialCentral estimate; sensitivity range 0.20–0.30
Saskaņas Centrs0,20partialMixed / residual overlap
Gobzema saraksts0,20partialPartial overlap with protest voters
Other 2026 party lists0,00excluded from modelNot counted in the central specification

The core is also shown separately: SV/AJ + ST!. In 2026 the national core is 13.71%, while the full C26 is 17.34%.

Why LPV is a sensitivity parameter

LPV is the largest source of modelling uncertainty. The central weight of 0.25 is therefore not treated as sacred: results should also be considered across the plausible 0.20–0.30 range.

16,68%LPV weight 0.20lower sensitivity bound
17,34%LPV weight 0.25canonical C26
17,99%LPV weight 0.30upper sensitivity bound
±0,65 ppapproximate deviation from central estimatenationally

Denominator: valid ballot envelopes

Vatometrs uses the number of valid ballot envelopes, rather than just the total of valid ballot papers. The frozen 2026 dataset contains 804,501 valid ballot envelopes and 793,846 valid ballots. The difference is deliberate and consistently preserved throughout the calculation.

The national total includes overseas polling stations. The municipality and state-city map covers only domestic areas, so its weighted aggregate differs from the national figure.

Comparison with 2018 and 2022

Party systems change, so comparing elections is not simply a matter of adding up parties with the same names. Each election has a predetermined classification intended to measure the same political niche.

ElectionSpecificationNational index
2018LKS + 0,35 × Saskaņa10,13%
2022ST! + LKS + Suverēnā Vara + 0,20 × (Saskaņa + LPV + KATRAM UN KATRAI)16,62%
2026SV/AJ + ST! + 0,25 × LPV + 0,20 × (Saskaņas Centrs + Gobzems)17,34%
This involves judgement. Election specifications are not laws of nature. Classifications and sensitivity ranges are part of the result, not a hidden technical detail.

Geographic aggregation

Latvija

On the municipality and state-city map, polling-station results are first summed within each administrative area, then the weighted index is calculated. One colour scale is used everywhere.

Rīga

The Rīga neighbourhood map is a special case. Ordinary polling stations are assigned to neighbourhoods using their official addresses and neighbourhood polygons. It reflects the geography of polling locations, not voters’ declared addresses. Rīga voters could choose any polling station in the city in 2026, so this layer must not be interpreted as a precise map of residents’ voting preferences.

Neighbourhoods without an ordinary polling station are shown as “no data” instead of receiving an invented score. Prison and special-location commissions are not mixed into ordinary neighbourhood polling-station aggregates.

What the index cannot tell you

Reproducibility

Calculations are performed during site generation using frozen integer vote counts and official denominators. The browser only displays precomputed values. SHA-256 checksums are kept for source snapshots, intermediate data and build artefacts.