Impact Scores, Prediction Markets and Election Results

The official results of the 19 April vote allow us to conclude the Election Observatory series with a summary and assessment of the various approaches we applied over the past few weeks. In the run-up to the early parliamentary elections, we launched the project with the intention of exploring innovative approaches to measuring the presence of key political actors in the information space.

We identified the most-discussed candidates on social media and the extent to which they relied on controversial language in their electoral communications. We tracked prediction markets to determine how effectively they could be used to forecast election results, and tested what information the Bulgarian voter would receive from large language models.

Influence Score: Who Are Social Media Talking About?

The Influence Score (Impact Score) - a proprietary metric formulated by Identrics - ranks politicians based on their presence in social media, taking into account not only posts from their official profiles but also posts from other profiles that mention them by name. A high Influence Score indicates that a given politician is widely present in online conversations, regardless of whether they are an active participant or the subject of discussion.

Only registered candidates for the parliamentary elections of 19 April are included, which means that President Iliyana Yotova, the Prime Minister and ministers of the caretaker government, as well as individuals with a political background and positions who are not part of the candidate lists, are excluded from the rankings. The study covers the period from 2 March to 19 April 2026 - seven weeks prior to the vote.

It is notable that the Influence Score of a large proportion of the leaders (measured in thousands of points) very precisely reflects the election result of their parties and coalitions (measured in percentages). For example, Rumen Radev tops the ranking with a score of just over 45,000, whilst the Progressive Bulgaria coalition he leads received 44.6% of the vote. The minimal difference between Boyko Borisov (GERB-SDS) and Asen Vasilev (PP-DB) perfectly mirrors the extremely close results of the second and third political forces.

The positions of the leaders in the influence ranking correspond to the final ordering of the parties, with the only exception being Krum Zarkov, who ranks higher than the BSP-OL coalition he leads. For DPS chairman Delyan Peevski, the Influence Score is formed almost entirely from mentions of his name in posts by other politicians and media, whilst he himself is not active on social media - his personal score of just over 9,000 points has no direct correlation with his party's result (7.1%).

A trend is also observed whereby the leaders of some political forces - Vazrazhdane (Kostadin Kostadinov), MECh (Radostin Vasilev), Velichie (Ivelin Mihaylov), and Siyanie (Nikolay Popov) - enjoy personal online popularity that does not fully translate into votes.

Radev's total dominance comes as no surprise given the impressive election victory of his coalition. During the monitoring period, PP chair Asen Vasilev gradually closed the gap on Boyko Borisov of GERB-SDS and even overtook him in the final two weeks before the vote. This online trend was borne out in the final results: GERB-SDS finished ahead of PP-DB by less than 0.8%, whilst opinion polls over the previous month had suggested a lead of between 7% and 12%.

Progressive Bulgaria

Most of the remaining individuals in the top ten can be divided into two categories: politicians and experts associated with the caretaker governments appointed by Rumen Radev (Ivan Demerdzhiev, Galab Donev, Ivan Shishkov, Anton Kutev), and well-known athletes standing on the coalition's lists (volleyball player Vladimir Nikolov, swimmer Petar Stoychev, and karateka Ivet Goranova).

GERB-SDS

In her role as Speaker of the National Assembly in the period before the vote, Nazaryan emerged as the most recognisable face of GERB after Borisov. The greatest popularity during the monitoring period was generated by posts commenting on her accidentally caught expletive remark from the parliamentary podium. The sharp and confrontational style of Delyan Dobrev earns him fourth place.

PP-DB

Of note is the presence in the top ten of two of the young faces of the anti-government protests of late last year: Anna Bodakova (DB) and Petar Tanev (PP). Another recognisable "Generation Z" figure, Malena Malcheva (PP), remains just outside the top 10 in 11th place. The fact that Anna Bodakova received over 6,800 preferential votes and re-ordered the PP-DB list in the 24th MIR suggests that this approach likely brought tangible benefits to the coalition.

DPS

The scores of Stoyanov and those following him - Yordan Tsonev and Stanislav Anastasov - are shaped largely by media coverage of and user reactions to their sharp accusations against the caretaker government. Those ranked from sixth place downwards have significantly lower scores.

Vazrazhdane

The popular journalist Martin Karbovski exerts considerable influence on the top ten, having published interviews, interview excerpts, and individual statements from eight of the politicians in the ranking during the monitoring period. The final two - Dimitar Panayotov and Stefan Petrov - are on Vazrazhdane's lists as representatives of the Marxist-Leninist organisation Movement "23 September", which promoted them in posts from its official profiles.

Controversy: What Are the Hot Topics of Conversation?

Another indicator by which we measure politicians' social media presence is controversy. Over the course of the official election campaign (20 March - 17 April), we tracked the most controversial posts on a daily basis - including not only registered candidates but also a broad range of politically connected individuals, media outlets, and influencers.

The ranking demonstrates Facebook's complete dominance among social media platforms - only two of the leading daily posts came from TikTok. The clear leader among authors is Progressive Bulgaria leader Rumen Radev: eight of his posts top the daily controversy ranking, driven by the large number of followers his profiles attract and high engagement.

The Radev posts that top the daily rankings most frequently contain accusations against previous governments and the parties within them, as well as declarations that Progressive Bulgaria is unwilling to enter into coalition with compromised partners. Political opponents are described as "the oligarchy", "corrupt figures", and "coalition-mongers". PP-DB was accused of having "destroyed the Constitution and attempting to corrupt the elections" (17 April), whilst Delyan Peevski was described as "an ordinary oligarch who will be swept from the stage" (9 April).

The other two party leaders who frequently adopt a confrontational tone are Asen Vasilev (PP) and Kostadin Kostadinov (Vazrazhdane). Vasilev attacked GERB with accusations of corruption (11 April) but also doubted that Radev was unwilling to work with "the oligarchs" (4 April). Kostadinov declared that the euro had been introduced on the basis of "fabricated data" and promised to "restore" the lev (25 March), and called for the arrest of those who signed the agreement with Ukraine (2 April).

Prediction Markets: How Did They Forecast the Bulgarian Elections?

During the official election campaign, we tracked the markets on the Polymarket prediction platform. The total volume of the eight markets linked to the Bulgarian elections exceeded $1.66 million on the eve of the vote - an increase of over 740% compared to the start of the campaign on 19 March, with the total volume of bets doubling in the final week alone.

These discrepancies show that, whilst the "collective wisdom" of prediction-market participants offers an interesting perspective on pre-election sentiment, it is not a perfect forecasting instrument. The probabilities reflect the objective views of market participants and are largely free of emotion and bias - but, at this stage in Bulgaria, Polymarket users are few in number and belong to a comparatively limited demographic group, and frequently place bets on the basis of media coverage and opinion polls that may be incomplete or misleading.