ODDS DNA ENGINE

How does Odds DNA work?

1. Odds are converted into probabilities. Each price is turned into an implied probability with 1/odds. Odds of 1.80 mean roughly 55.6%.
2. The bookmaker margin is removed. The implied probabilities in a market add up to more than 100% (the overround). Each family is normalised on its own, so matches priced with different margins can be compared on the same scale.
3. Similarity is measured. The 1X2, over/under and both-teams-to-score profiles are compared with past matches using a weighted distance, producing a DNA similarity from 0 to 100. Only matches above the threshold enter the sample.
4. Realised outcomes are counted. From the real scores of the sampled matches we derive the distributions of match result, over/under, both teams to score, half time and HT/FT.

Sample tiers

Every result shows how many similar matches it is based on. Four tiers are shown according to sample size:

• Low sample — rates are highly exposed to chance and should not be read on their own.
• Limited sample — a few matches can change the distribution noticeably.
• Moderate sample — a tendency can be read, but it is not certain.
• Strong sample — the distribution is relatively stable.

Statistical honesty

The percentages shown are the realised frequency in past matches with a similar odds structure; they are not this match's probability. For example, “71.4% of 640 similar matches ended Over 2.5” is a count; it does not mean “this match will end Over with a 71.4% chance”. Values that cannot be calculated are shown as “—” and are not counted as zero. No output is betting advice or a guarantee of winnings. To study odds movements, see Odds Radar; for the ten nearest neighbours of an opening-odds profile, see the DNA & Fingerprint screen.