Understanding Winning Probability
Winning Probability is a predictive framework that estimates a candidate's chances of winning based on historical election data, party organizational strength, and current political trends.
Important Notice
This is purely a data-driven prediction model. It relies strictly on publicly available historical election results and current measurable indicators to estimate winning probability. It may be inaccurate for newly emerging candidates or parties lacking historical voting records.
What is Winning Probability?
It calculates a probability score (0–100%) for each candidate in a given constituency by evaluating various qualitative and quantitative signals.
Previous vote shares, direct election results, and historical win/loss records (e.g., 2074 and 2079 elections).
Analysis of party structures, political alliances, voter loyalty, and grassroots network presence.
Candidate visibility, current activity levels, media presence, and constituency-specific political dynamics.
Data Sources & Analysis
The prediction model synthesizes data from publicly available official records and structured datasets. These fundamental pillars contribute specific weights to generate the final winning probability score.
Historical Voting Patterns
Evaluation of official results from the 2074 and 2079 elections, factoring in PR (Proportional Representation) votes, direct votes, and consistent patterns of victory or defeat.
Party Structure & Alliances
Impact assessment of party mergers, splits, electoral alliances, candidate defections, and historical vote transferability.
Current Political Indicators
Current public visibility, engagement levels, recent political developments, and unique local political contexts.
How the Probability is Calculated
Each candidate is assigned an initial base score derived from multiple weighted signals. These scores are then adjusted competitively and normalized into a final percentage value.
Calculation Steps
- 1
Base Foundation
Establishing a base score largely determined by the party's PR vote strength and the candidate's past win/loss context.
base = PR_vote × (won ? 1.6 : 1.3)
- 2
Personal Adjustments
Applying positive weights for previous incumbency, strong constituency ties, and historically proven electoral performance.
- 3
Competitive Dynamics
Factoring in current candidate visibility, activity metrics, and adjustments for highly marginal, close-contest scenarios.
- 4
Normalization
Finally, converting all candidates' aggregated scores proportionally to formulate a relative winning probability out of 100%.
Winning Probability (0–100%)
Why Do New Candidates/Parties Score Differently?
Because our predictive model heavily relies on proven historical voting data to establish confidence. Completely new entrants or recently formed parties inevitably lack historical PR data or verified candidate track records. Consequently, the model assigns lower confidence and may underestimate their actual ground support.
How to Interpret the Output
Score
Probability Score (0–100%)
A higher percentage indicates a mathematically stronger position. However, it is an estimate, NOT a guarantee of victory.
Confidence Interval
The confidence of our prediction increases proportionally with the availability of rich historical data.
- High Confidence Data Rich
- Low Confidence Data Poor