gerry
Mare
A price is value only when the odds you take are higher than the true probability of the outcome.
That’s it.
Not “higher than Betfair.”
Not “higher than the bookies.”
Not “looks big.”
Not “drifted.”
It’s purely a comparison between:
- Your estimated probability, and
- The implied probability of the odds you’re offered.
Odds imply a probability.
For example:
- Odds of 4.0 imply a probability of 1/4=0.25 or 25%.
If you believe the horse actually has a 30% chance of winning, then:
- Your probability (30%) > market probability (25%)
- That’s value.
If you believe the horse has a 20% chance, then:
- Your probability (20%) < market probability (25%)
- That’s negative value, even though the odds look “big.”
This is where most bettors fall down.
The market is usually efficient because:
- Thousands of people are pricing the same race
- Information is shared instantly
- Betfair money corrects bad prices quickly
So to find value, you need an edge that lets you estimate probability better than the market. That could come from:
- Deep form analysis
- Sectional times
- Pace maps
- Trainer patterns
- Ground biases
- Inside information (legal or otherwise)
- Statistical modelling
- Identifying market overreactions
Without an edge, you’re just guessing.
Betfair is often the most accurate reflection of true probability.
If a bookmaker is offering bigger odds than Betfair, it’s usually because:
- They’re slow to update
- They’re boosting to attract bets
- They’re wrong (rare)
- You’re comparing at the wrong moment
- The horse is actually less likely than the bookmaker thinks
You only have value if your probability estimate beats Betfair’s.
You think a horse has a 40% chance of winning.
- Betfair price: 2.5 (implies 40%)
- Bookie price: 3.0 (implies 33.3%)
Is that value?
Only if your 40% estimate is correct.
If the true probability is actually 33%, then the bookie’s 3.0 is fair, not value.
To know whether a price is value, you need a method for estimating true probability. That’s the entire game.
If you want, we can go deeper into:
- How to build a probability model



