69% accuracy — is that good? The break-even rate explained
A number that means nothing on its own
"90% accuracy" is the most common promise in tipping channels. It sounds convincing, and on its own it is meaningless.
A hit rate tells you what share of bets landed. It says nothing about the prices. You can hit 90% and lose money — just back favourites at 1.05. You can hit 40% and profit, if your average price is above 2.50.
The only question that matters is whether the money staked came back, and how much on top.
The break-even rate
Every average price has a hit rate at which you break even:
Break-even rate = 100 / average odds
Some examples:
- average odds 1.20 → you need 83.3%
- average odds 1.50 → you need 66.7%
- average odds 1.65 → you need 60.6%
- average odds 2.00 → 50% is enough
- average odds 3.00 → 33.3% is enough
Back to "90% accuracy". At odds of 1.05 the break-even rate is 95.2%. A channel hitting 90% at those prices is losing money, even though it is almost never wrong.
Our numbers
We publish everything, so this is easy to check. At the time of writing:
- 271 settled bets
- 69.4% hit rate
- average odds 1.65
- break-even rate 60.6%
- ROI +13.5%
The margin above break-even is 8.8 percentage points, and that gap is the profit: every $100 staked returned about $114.
Notice how modest that looks next to "90% accuracy" — and how much more honest. Current figures are always on the statistics page; they are recalculated after every match, losses included.
Why ROI beats hit rate
ROI answers the only question that matters — what came back on what went in:
ROI = (total profit / total staked) × 100%
A winning bet returns (odds − 1) of the stake; a losing one takes the whole stake. Add it up across every bet and divide by their number.
The professional benchmark in football is ROI between 5% and 15% over a few hundred bets. Anything above 30% across a long run usually means one of two things: the sample is too small, or the losses are not in it.
How to check any tipping service
Three questions settle it:
- What is the average price? Without it a hit rate cannot be interpreted.
- How many bets are in the sample? At 20 bets any percentage is noise. Meaningful conclusions start around 200–300.
- Are losses published? If the feed is all winners, you are looking at a shop window, not a record.
If any of the three has no answer, the headline percentage means nothing.
How we calculate probabilities and find the gap against the bookmaker's line: see what a value bet is and the Poisson model.