Value betting masterclass · Module 4: Football markets and the Value Predictor · Lesson 18 of 25 · About 7 min
Tipster sites, stats tools and trading communities love a track record: hit rates, winning streaks, members’ screenshots and “best month” banners. This lesson shows you how to judge any track record, ours included, in five checks, so you can tell a real edge from numbers that only look good.
The short answer is that a hit rate on its own tells you almost nothing. What matters is the price the bets were taken at, how many bets there were, and whether the rules were written before or after the results were known.
Read the transcript
Tipsters and stats tools love a track record. Here's how to tell a real edge from numbers that only look good.
First, a hit rate isn't a profit. Seventy-two per cent winners at one point three loses six point four per cent. Fifty-five per cent at one point nine makes four and a half. Always ask, at what odds?
To check, work out the break-even odds: one divided by the hit rate. A seventy-two per cent hit rate needs odds of one point three nine or more, just to break even.
Second, check the sample. A method that truly wins sixty per cent of the time could show anything from forty-six to seventy-four per cent over fifty bets. Look for hundreds.
Third, were the rules set before the results? Test twenty filters with no real edge, and there's a sixty-four per cent chance at least one looks great by luck. That's the one you'll be shown.
Fourth, could you really have got those prices? Allow for commission, which turns two point oh into one point nine five at five per cent, and for bookmaker stake limits.
And fifth, who's showing you the record, and why? Screenshots come from the winners, and subscriptions pay whether members win or lose.
A believable record shows the odds, the profit and the losing runs. Judge the price, not the pick, and please bet responsibly.
The short version
- A hit rate is not a profit. Always ask: at what odds?
- Break-even odds are 1 ÷ the hit rate. A 72% hit rate needs odds of 1.39 or more just to break even.
- Small samples swing wildly. 50 bets can be 14 percentage points away from the truth either way.
- Rules fitted to past results almost always look better than they will do on new matches.
- Check the prices were ones you could actually get, after commission and stake limits.
Check 1: a hit rate is not a profit
“Over 2.5 goals landed 72% of the time” sounds impressive. But a bet that wins often is usually priced short. The question that matters is whether the price paid more than the chance.
Worked example. Say a filter finds 100 matches, the over 2.5 goals bet lands in 72 of them, and the typical price was 1.30. With £10 on each:
| Step | Working | Result |
|---|---|---|
| Staked | 100 × £10 | £1,000 |
| Returned | 72 winners × £10 × 1.30 | £936 |
| Profit | £936 − £1,000 | −£64 (−6.4%) |
| Break-even hit rate at 1.30 | 1 ÷ 1.30 | 76.9% |
A 72% hit rate lost money, because at 1.30 you need to win 76.9% of the time. A 55% hit rate at 1.90 would have made 4.5%. So when you see a hit rate, work out the break-even odds (1 ÷ the hit rate) and compare them with the odds you can actually get. Our odds converter does this for you.
Check 2: how many bets?
Results bounce around a lot over a small number of bets. Even if a method truly wins 60% of the time, here is the range you would expect to see, 95% of the time, just from luck:
| Number of bets | Hit rate you might see |
|---|---|
| 50 | about 46% to 74% |
| 200 | about 53% to 67% |
| 1,000 | about 57% to 63% |
So a month of good results proves very little. Look for hundreds of bets, and be wary of anyone who shows only a short, hand-picked period. Our variance calculator shows how wide the swings can be.
Check 3: were the rules written before or after the results?
Take enough past data and slice it enough ways (by league, odds band, minute, form, day of the week) and some slices will always look brilliant. That is luck dressed up as a system.
Here’s why. If you test 20 filters that have no real edge, and each has a 5% chance of looking “significant” by luck, the chance that at least one looks great is 1 − 0.9520, which is about 64%. The person selling the filter shows you that one.
The fair test is to fix the rules first, then check them only on matches that came afterwards. Ask whether a record is a backtest (rules applied to the past) or live results recorded at the time.
Check 4: could you have got those prices?
- Which price? Results worked out at the best price across every bookmaker, or at the peak price before it shortened, are better than most people will get.
- Commission. On an exchange, commission comes off your winnings. At 5%, a back bet at 2.00 really pays like 1.95. At 2%, like 1.98.
- Stake limits. Bookmakers often limit accounts that keep winning or keep taking the best prices, so a strategy that relies on them may not last.
Check 5: who is showing you the record, and why?
- Screenshots and testimonials come from the people who won. The people who lost rarely post.
- Subscriptions earn the same whether members win or lose, so the incentive is to keep people subscribed and betting.
- Affiliate links pay when people sign up with a bookmaker. We earn this way too, which is why we tell you to judge every bet on its price, not on our say-so.
None of this means every track record is false. It means a believable one shows the odds, the profit as well as the hit rate, a large number of bets, rules fixed in advance, and the losing runs as well as the winning ones.
What this means for the Value Predictor
Our Value Predictor doesn’t sell tips. Recent form is already built into bookmakers’ prices, so a form shortlist on its own isn’t an edge. That’s why the Value Predictor shows the best UK price, the fair price from the betting exchanges, a Value tag when a price beats fair by 3% or more, and the break-even price for each market. The price is what decides whether a bet is worth taking.
Track record questions
What is a good hit rate?
There isn’t one. A 40% hit rate at 3.00 makes 20%, while an 80% hit rate at 1.20 loses 4%. A hit rate only means something next to the odds.
How many bets do I need before I trust a record?
Several hundred at the very least, and more for bigger odds, because results at long odds swing more. Even then, check whether the bets beat the closing price, which is a quicker and steadier sign of an edge than profit.
Is a backtest worthless?
No. A backtest is a useful first check, as long as the rules were set before looking at the results and the prices used were ones you could have got. Treat it as a reason to keep testing, not as proof.
How do I judge my own betting?
Keep a record of every bet, with the odds you took and the closing odds. Our free bet tracker works out your profit, return on investment, longest losing run and closing line value for you.
Related lessons: implied probability, betting variance, closing line value, keeping betting records and value betting myths.
Lesson 18 of 25 · Football markets and the Value Predictor
18+ only. Betting should be fun, and every bet can lose. Set a budget you can afford to lose and stick to it. For free, confidential support, visit BeGambleAware.org or call the National Gambling Helpline on 0808 8020 133.