Estimated positions: how many, where and how they count

Each of the 12 parties in the test has a position on each of the 63 questions: the 42 statements that make up the test and the 21 tie-break questions, which are only offered when a result is close. That makes 756 positions in all. For 516 of them we can point to a specific document. The remaining 240 carry the label “Estimated position”.

This article explains what lies behind that label, where such positions are found and how the calculation handles them. One caveat up front: the number of estimates reflects how many suitable documents we found for a given party and question, not what kind of party it is. It says nothing about the party, for better or worse.

What an estimated position is

We mark a position as estimated when we cannot find anything that settles the specific question: no passage in a manifesto, no vote, no verifiable statement. In that case we set it according to what the party stands for in general or on a related issue, and we say so openly. Sometimes a document touches on the question, such as an older or conflicting statement, but it is not enough to settle it. On the party's page such a position shows a label instead of a link to a source, and in Compare the parties it has an asterisk and a dashed border.

In one respect the estimates are more cautious than the other positions. The scale has five points, and sourced positions use all of them; of the 516, 98 sit at 1 and 110 at 5. Among the 240 estimated positions there is not a single 1 and not a single 5: 95 are at 2, 89 at 3 and 56 at 4. We only give an extreme value when a document supports it clearly and without qualification, and for an estimate we have no such document.

How many there are

Two groups of questions need to be kept apart. The figures on the methodology page and on the party pages refer only to the 42 test statements, because only these go into every result. The tie-break questions are counted separately.

In the test itself and in the quick test, then, more than one position in four is estimated. On 5 of the 42 statements, all twelve parties have a sourced position. The tie-break questions are a different story: every one of the 21 has at least one estimated position. They are there to separate parties that sit close together in the main test, and for many of them we did not find a suitable document for at least one party.

By party

The table lists the parties in the order used on the site. The first figure is the number of estimates among the 42 test statements, the second among the 21 tie-break questions.

The differences have mundane causes: above all, they depend on which documents exist. A good example is the vote of a parliamentary group, which often answers a question directly but is only available for parties that won seats in 2023. Republika, the Hungarian Alliance, Demokrati and Sme rodina did not enter parliament then, and Právo na pravdu, registered in November 2025, has not yet stood in an election. None of these five parties therefore has a vote as the source for any test statement, whereas the seven parliamentary parties have between 3 and 10 such positions each. Nor does an estimate mean that a party has no view on the issue; it means we did not find that view written down in a way that answers our exact question.

By issue

Across the test statements, estimates are unevenly spread:

On healthcare half of the positions are estimated, and on minorities more than two in five: there were several parties for which we found no document that addresses the particular wording of our statement. Migration (16.7%) and the rule of law (18.3%) have the fewest estimates. Among the tie-break questions, values and society has the most estimates (26 of 48 positions).

The statements with the most estimates

The statement with the most estimates is the one on official status for Hungarian in municipalities where Hungarians make up a large part of the population: 11 out of twelve, with only the Hungarian Alliance's position sourced. The statement on a single public health insurer has eight, and those on cutting the standard VAT rate and on a referendum on staying in the EU have seven each. Five more have six: on pulling out of the defence cooperation deal with the United States, on hiring doctors and nurses from other countries, on using frozen Russian assets for Ukraine's benefit, on cutting back support for refugees from Ukraine, and on keeping social and family benefits even if taxes have to rise. You will find the exact wording of these statements, with every party's position, on the pages of the relevant issues.

Among the tie-break questions, the one with the most estimated positions, 10 out of twelve, is on asking a doctor for help to end one's life, followed by an age limit of 16 for opening one's own social media account, with nine. Three have eight: a ban on charging for an earlier appointment, support for education and culture in Hungarian, and writing EU and NATO membership into the Slovak Constitution.

How they enter the calculation

We checked this directly in the code that computes the result, not only in the methodology text.

In your match, estimates carry the same weight as sourced positions. Every position enters the calculation with a confidence weight of 3, whether or not it has a source. The same applies to the issue-by-issue match and to where parties are placed on the political compass. If estimates were given less weight, the effect on parties would be uneven: those with more estimates would mechanically count for less in the calculation. So we opt for equal weight and a visible label. How positions become a percentage is set out in detail in the analysis How your match with the parties is calculated.

Estimates are not used when tie-break questions are chosen. When two or three parties at the top of your result are close together, the test picks questions that tell them apart. It leaves out any question on which even one of the tied parties has an estimated position. A close result should not be decided by our estimate.

Estimates are not quoted in the texts. The texts on party positions and on the differences within each issue include estimated positions in their figures, but never present them as documented. The text on a comparison of two parties is shown only if every position it relies on has a source; if one of them turns into an estimate, the text disappears until we rewrite it.

What the other 516 positions rest on

For every sourced position we give the name of the document, a link and a date. From the names it is possible to tell, with reasonable confidence, what kind of source each one is. We classified every position by the first document its source names (64 positions cite more than one document at once):

Of the 62 votes, 61 link directly to the record on nrsr.sk and one to the overview of votes compiled by KohoVolit.eu. Looking at the 42 test statements alone, the proportions are similar: of the 360 sourced positions, 217 come from party documents, 51 from parties' answers in voting advice applications, 50 from media reports, 41 from votes and 1 from the government programme statement.

We have not attempted a finer breakdown, such as separating manifestos from press releases. The source name does not allow that to be done reliably enough, and we would rather publish a coarser breakdown that can be checked.

How you can help

The quickest way for an estimate to become a documented position is for someone to know of a document we missed. If you know of one, send it to us through the Contact page. Tell us the party and the statement, the point from 1 to 5 that you think the party should have, the source itself (a page of a manifesto, a National Council vote with its date, or a statement, ideally with a link) and a short explanation. We will read the source. If the source confirms the position or points to a different one, we update it, add the document and the review date to the party's page, and reply to you either way, even if we make no change. This also applies to positions that already have a source: a better document can challenge that too.

Every change to positions is recorded in the revision log in the methodology.

What not to read into these figures

They are not a measure of the quality of a party or its manifesto. A party outside parliament has no National Council votes that can be cited, and a new party has had less time to publish documents. A party with a short manifesto need not hold a less clear view on an issue; it may simply not have written it down.

The figures are as of 5 November 2026. Some positions rest on manifestos written for the 2023 election; when parties publish their manifestos for the 2027 election, we will go through the positions again and the number of estimates may change. The current figures for each party are on its page in the Parties section.