The EurovisionOdds SF1 Forecast Result: 10 of 10 Qualifiers Called Correctly — The 5-Signal Model Validated For Eurovision 2026 Final Week

On May 12 at 09:00 CEST, twelve hours before Semi-Final 1 broadcast, we published the EurovisionOdds SF1 Forecast — a proprietary 5-signal model that called the top 10 qualifying countries plus three contrarian betting positions against the bookmaker market. The model was specific: a single 0-100 confidence percentage per country, with a fixed methodology disclosed in advance.

Last night at 22:54 CEST, the EBU announced the 10 qualifying countries from SF1. We can now grade the forecast against the actual result.

Headline grade: 10 of 10 qualifiers called correctly. 100% top-10 hit rate. One contrarian call hit (Georgia to NOT qualify, at 1.45-1.55 lay). One contrarian call missed (San Marino each-way at 17.00). Belgium top-5 SF1 — pending final SF1 placement publication.

This article publishes the full grade, the per-signal contribution analysis, and the methodology adjustments we are making for SF2 (broadcasting May 14 evening) and the Grand Final (May 16).

Betfred — Bet £10 Get £50 in Free Bets on Eurovision 2026 Final

The Headline Result

Source: Official EBU SF1 results published on eurovision.tv May 12 evening. Forecast published May 12 morning at /blog/eurovisionodds-sf1-forecast-top-10-qualifiers-may-12-2026-proprietary-5-signal-model.

What This Means For Public Models

Going into SF1, the betting consensus across major bookmakers had 7-8 of these 10 qualifiers as the top-7 favourites. Our model agreed on those 7 but added Belgium (75% vs market 50%), Serbia (72% vs market 52%), and Poland (68% vs market 50%) as contrarian qualifiers. All three qualified. The market caught up overnight — Belgium top-5 SF1 odds have shortened from 6.50 to 4.20 within hours, Serbia outright Final win from 40.00 to 28.00, Poland from 100.00 to 60.00.

Our 100% top-10 hit rate compares against the Audience Poll Predictive Power Audit's 87% top-10 hit rate across 3 years and the best public bookmaker model's 78%. We acknowledge a single contest is not a backtest — we expect to revert toward the mean in SF2 and the Final. But the SF1 result establishes a baseline of credibility we can apply to the Final-week bets.

Per-Signal Contribution Analysis

Of the 10 qualifiers, the model's confidence was driven by which of the 5 signals?

• Bookmaker probability (35% weight) — the dominant signal. Finland, Greece, Israel, Sweden, Croatia all had high market probability and the model simply weighted that correctly. 7 of 10 qualifiers were market-favoured.
• Audience poll vote share (25% weight) — the key differentiator on Moldova. Moldova's 28.2% poll lead lifted the model's confidence from 60% (market) to 90% (model). Result: model right.
• Jury show signal (20% weight) — the key differentiator on Belgium and Poland. EurovisionFun's coverage of Essyla's snowstorm-cinematic staging and Alicja's gospel performance both rated highly. Both qualified.
• Running order (10% weight) — Serbia's closing position 15 was the model's clearest application. The recency-bias effect held. Serbia qualified.
• Auto-qualifier benchmark (10% weight) — Italy and Germany's reference performances anchored the rubric. No qualifier scored against this signal in isolation.

Three Contrarian Calls — Grade Card

Call 1: Georgia NOT to qualify at 1.45-1.55 (lay market)

HIT. Georgia finished 15th of 15 and missed qualification by a wide margin. The vocal issues reported in the jury show coverage materialised on the live broadcast. Bzikebi's energy didn't translate. Anyone who took the lay market at 1.45-1.55 cleared positive expected value.

Call 2: San Marino each-way at 17.00 outright qualification

MISSED on the outright; partial on the each-way. San Marino finished 13th in SF1 — out of qualification but not last. The forecast's 48% probability vs market's 22% was directionally correct (we said the market under-priced San Marino) but the qualification still went the other way. Each-way bets paid on a top-3 SF1 finish, which San Marino did not achieve. Net result: the bet lost.

The model's overreliance on the jury show coverage of Senhit + Boy George's staging was the methodology error. Production quality matters less than vocal precision in a tight semi-final. We adjust Signal 3 (Jury show) weight downward for SF2 from 20% to 17%, redistributing 3pp to Signal 1 (Bookmaker probability).

Call 3: Belgium top-5 SF1 at 6.50

PENDING. The EBU publishes the SF1 detailed scoring (jury rank, televote rank, total points) typically 24-48 hours after the broadcast. Belgium's top-5 SF1 status will be known by May 14. Press centre signals from last night suggested Belgium finished in the upper half of SF1 — provisionally positive but unconfirmed.

What Changes For SF2 + Grand Final

The model is fundamentally validated. Three minor adjustments:

1. Signal 3 (Jury show) weight: 20% → 17%. Production quality is a lagging indicator. Vocal precision is leading. We re-weight.
2. Signal 1 (Bookmaker probability) weight: 35% → 38%. The market is more accurate than we initially modelled. Trust the price more.
3. Signal 5 (Auto-qualifier benchmark) extends to Big 4 + Austria in the Final. Previously calibrated to Italy + Germany only. For the Final, France, UK, Italy, Germany, and Austria all become reference benchmarks.

The 5-signal architecture stays. The weights shift incrementally based on what the data taught us last night.

SF2 Forecast — Coming Within 2 Hours

We publish the SF2 Forecast for the 15 SF2 entries (Bulgaria, Azerbaijan, Romania, Luxembourg, Czechia, Armenia, Switzerland, Cyprus, Latvia, Denmark, Australia, Ukraine, Albania, Malta, Norway) within the next 2 hours, using the SF2 audience poll data once it releases this afternoon, plus the recalibrated weights above. The pre-SF2 audience poll typically lands between 14:00-17:00 CEST today.

Methodology Limitations Disclosed

1. One contest is not a backtest. 100% top-10 hit rate over a single 15-country semi-final does not establish a 100% expected hit rate going forward. The Audience Poll Predictive Power Audit's 3-year sample is the more conservative reference; we expect the model to revert to 87-93% across the SF2 + Final cycle.
2. The contrarian-call grade is asymmetric. The Georgia lay-market bet had a defined positive expected value (1.45-1.55 implied 65-69% vs forecast 70% non-qualification); the San Marino each-way bet had higher variance (1/4 fractional odds at 17.00 requires top-3 SF1 to break even on each-way). The fact that one hit and one missed is consistent with the forecast probabilities.
3. The audience poll signal cannot be replicated for SF2 until it releases. The May 11 SF1 audience poll was the strongest single fresh signal in our SF1 model. SF2's equivalent poll typically drops Wednesday afternoon. Our SF2 forecast will integrate it once available.
4. Geopolitical disruption is not modeled. The Stadthalle protest incident during Israel's performance, the pro-Palestinian demonstration planned for May 16 in Vienna, and the 5-country boycott (Spain, Slovenia, Netherlands, Ireland, Iceland withdrawn) all introduce variance the 5-signal model does not capture. Separate analysis articles cover these factors.

How To Cite This Validation

Ferretti, M. (2026). "The EurovisionOdds SF1 Forecast Result Analysis: 10 of 10 Qualifiers Called Correctly." EurovisionOdds.org, May 13, 2026.

The Bottom Line

The 5-signal model called Eurovision 2026 SF1 with a 100% top-10 hit rate. One contrarian betting call hit (Georgia lay market), one missed (San Marino each-way), one pending (Belgium top-5 SF1). The model's overweighting of jury show production quality was the only methodology adjustment required. Weights shift modestly for SF2 + Final. We will be measured against the SF2 forecast tomorrow.

Stake — Crypto Betting with Instant Payouts on Eurovision 2026

Related Articles

• The EurovisionOdds SF1 Forecast (the original, published 12 hours before SF1)
• The EurovisionOdds Audience Poll Predictive Power Audit
• The EurovisionOdds Bloc Voting Quotient: 4,200+ Twelve-Point Exchanges
• The EurovisionOdds Last-Week Mover Index: Final-Week Odds Movement
• The 10 SF1 Qualifier Deep Dives — One Article Per Country

SF1 forecast methodology and results verified May 13, 2026 from official Eurovision Song Contest 2026 results published by the European Broadcasting Union on eurovision.tv. 18+. Please gamble responsibly. BeGambleAware.org. When the fun stops, stop.