Probability, fair odds and value side by side

1. Probability

The model works out how likely each outcome is. The total is always 100 %, because there is no other option.

A probability doesn't say what will happen. It says how often that outcome would occur if the match were played many times under the same conditions. Fifty-five per cent is still close to a coin toss.

2. Fair odds

Fair odds are the reciprocal of the probability — the odds at which a bet would neither win nor lose over the long run:

probability 50 %  →  1 / 0.50  =  2.00
probability 25 %  →  1 / 0.25  =  4.00
probability 58 %  →  1 / 0.58  ≈  1.72

A bookmaker calculates its odds the same way, then builds in a margin. That's why market odds tend to sit below the true chance — and why it pays to compare them with the fair odds rather than with a gut feeling.

3. Value

Value is the gap between the market odds and the fair odds, expressed as a percentage. The formula is value = probability × odds − 1.

How to read value

Positive value means the odds are higher than the model's estimate implies. Negative means the opposite — the odds are worse than fair, and such a tip never makes it into the best-tips list.

One thing to watch: +7.6 % is not the profit on a bet. It's an expected value. Repeat the same situation many times, and if the model is right, it returns an average of 7.6 % of the stake. In one particular match it either lands or it doesn't, and neither outcome says anything about the model.

What this all rests on

Value only means something if the model's probability is closer to the truth than the odds are. When the model is wrong, high value just means it was wrong by a lot. So two things matter: what data the model runs on, covered in Methodology, data sources and how often it changes, and how it holds up in hindsight, which the chapter How tips are settled shows.