Social feeds are full of "professional tipsters" posting screenshots of green profit lines. Some of them are genuinely skilled. Most are not. The problem is that from the outside, a lucky tipster and a skilled tipster look exactly the same for a surprisingly long time.
This guide gives you the maths to tell them apart. No guru opinions, just calculations you can repeat yourself in a spreadsheet.
TL;DR
- 5 wins in a row at 50% odds happens 1 time in 32. Among 1,000 tipsters, around 31 will post a 5-win streak in any 5-tip window, by pure chance.
- 100 tips is not a track record. At odds around 2.0, the uncertainty on a 100-tip ROI is roughly ±19 percentage points.
- To show a real +5% ROI with reasonable confidence at odds around 2.0, you need about 1,600 tips. At odds of 4.0, about three times as many.
- Closing line value (CLV) shows skill much faster than profit, because it doesn't depend on whether the ball hit the post.
- A paid tipster has to beat the bookmaker and the subscription fee. That second hurdle is bigger than most people think.
Skip to: the 6-point tipster checklist →
1. Streaks: why "10 wins in a row" means less than you think
Assume a tipster with zero skill picks bets with a 50% chance of winning. The probability of a winning streak of a given length, starting from any given tip:
| Streak length | Probability | Out of 1,000 zero-skill tipsters |
|---|
| 3 wins in a row | 1 in 8 (12.5%) | ~125 |
| 5 wins in a row | 1 in 32 (3.1%) | ~31 |
| 7 wins in a row | 1 in 128 (0.8%) | ~8 |
| 10 wins in a row | 1 in 1,024 (0.1%) | ~1 |
And that's for a single window. Over a season, each tipster gets hundreds of chances to start a streak, so long runs become close to guaranteed for somebody. Thousands of accounts post tips every day. Someone will always be on a hot streak, and that account is the one you'll see shared.
This is survivorship bias. The unlucky tipsters quietly delete their accounts. The lucky ones screenshot their run and sell a VIP group. How your brain gets fooled by streaks →
2. Sample size: how many tips before a record means anything?
Every record has noise. The question is how much.
For bets at odds of about 2.0, the profit on each one-unit bet is either +1 or −1, so the standard deviation per bet is about 1 unit. The uncertainty on ROI shrinks with the square root of the number of tips:
| Tips in record | Standard error of ROI | 95% range around the true ROI |
|---|
| 50 | ±14.1% | ±27.7 pp |
| 100 | ±10.0% | ±19.6 pp |
| 250 | ±6.3% | ±12.4 pp |
| 500 | ±4.5% | ±8.8 pp |
| 1,000 | ±3.2% | ±6.2 pp |
| 2,000 | ±2.2% | ±4.4 pp |
What this means in practice: a tipster showing +15% ROI over 100 tips could have a true ROI anywhere from about −5% to +35%. That record tells you almost nothing.
To be reasonably confident (two standard errors) that a +5% ROI is real, you need:
2 × (1 / √n) ≤ 0.05 → n ≥ 1,600 tips
Higher odds make it worse. At odds of 4.0 (a 25% win chance), the per-bet standard deviation is about 1.73 units, so you need roughly three times as many tips (around 4,800) for the same confidence. That's why accumulator and "long-shot" tipsters can hide a total lack of skill for years.
3. The simulation: 1,000 tipsters, zero skill, real bookmaker margin
Let's make it concrete. Imagine 1,000 tipsters with no skill at all. Every pick has a true 50% chance, and they bet at odds of 1.90, so the bookmaker margin gives them an expected ROI of −5%.
How many of them would still show an ROI of +10% or better, purely from luck?
| Tips posted | Zero-skill tipsters showing +10% ROI or better |
|---|
| 50 | ~130 out of 1,000 |
| 100 | ~57 out of 1,000 |
| 200 | ~13 out of 1,000 |
| 500 | fewer than 1 out of 1,000 |
(Normal approximation. Per-bet standard deviation at odds 1.90 is about 0.95 units.)
Read that first row again. After 50 tips, about 13% of completely skill-less tipsters look like +10% ROI winners. Those are the accounts that go viral and start charging subscriptions. By 500 tips, luck alone almost never produces that result, which is exactly why short-term screenshots are the main marketing tool.
4. Win rate is meaningless without odds
"70% win rate!" sounds impressive. It isn't, on its own. The break-even win rate is simply 1 ÷ decimal odds:
| Average odds | Break-even win rate | A 70% win rate gives… |
|---|
| 1.20 | 83.3% | −16% ROI |
| 1.50 | 66.7% | +5% ROI |
| 1.90 | 52.6% | +33% ROI (unrealistic long-term) |
| 4.00 | 25.0% | Not plausible at these odds |
A tipster with a 70% win rate on short favourites at 1.20 is losing money. A tipster with a 30% win rate at 4.0 odds is making +20%. Always ask for average odds and ROI, never win rate alone.
5. Closing line value: the shortcut to spotting skill
Profit is noisy because results are noisy. Closing line value (CLV) removes most of that noise.
CLV compares the odds a tipster posted with the final odds just before kick-off at a sharp bookmaker such as Pinnacle. Closing odds reflect all the information and money in the market, so consistently getting a better price than the close is strong evidence of skill, even over a few hundred tips.
CLV = (odds taken ÷ closing odds) − 1
Tip posted at 2.10, closed at 1.95 → CLV = +7.7%
Tip posted at 1.80, closed at 1.90 → CLV = −5.3%
A tipster whose picks regularly close at shorter odds than posted is finding value before the market does. A tipster whose picks drift out is usually just following public sentiment. Full CLV strategy guide → · Free CLV calculator →
One practical signal: bettors who consistently beat the closing line tend to get their bookmaker accounts limited. Being limited isn't proof of skill on its own, but a tipster who claims years of big profits and has never been limited anywhere deserves extra scrutiny.
6. The subscription hurdle
A paid tipster has to beat two things: the bookmaker margin and the fee you pay.
Example: a €30/month service costs €360 a year. If you stake €20 per tip and the tipster genuinely delivers +3% ROI, each tip earns you €0.60 on average:
€360 ÷ €0.60 = 600 tips per year just to break even on the fee
That's roughly 12 tips a week, every week, at a real +3% edge, before you see a cent of profit. With smaller stakes or a smaller edge, the subscription eats everything.
The 6-point tipster checklist
Before you follow or pay anyone, check all six:
1. Is the full record public, with timestamps?
Every tip, posted before kick-off, with odds and stake. Screenshots of winning slips are not a record.
2. Is it independently verified?
Self-reported spreadsheets are easy to edit. Use records held on an independent verification platform where tips can't be deleted after the result.
3. Is the sample big enough?
Use the table above. Under 500 tips, treat any ROI as mostly noise. For high-odds tipsters, demand much more.
4. Do they report CLV, not only profit?
A tipster who beats the closing line on average is a far better bet than one with a lucky profit run and negative CLV.
5. Are the drawdowns visible?
Every genuine bettor has long losing runs. A record with no red months is either very short or edited.
6. Do they explain why?
"Home side's press against a back line that plays out from the keeper, and the referee lets contact go" is analysis you can check. "LOCK OF THE DAY 🔥 SLAM IT" is marketing.
The alternative: become your own analyst
Once you see how much noise sits inside a tipster's record, the logical conclusion is to apply the same scrutiny to your own bets: log every pick, track CLV, and judge yourself on hundreds of bets, not a good weekend.
StatPair gives you the data side of that process: player and team stats with home/away and half-by-half splits, referee stats, a Trends Finder and AI projections for shots, fouls, cards and corners. It doesn't sell picks.
Try StatPair free for 7 days → · Expected value calculator → · Build a simple betting model in Excel →
FAQ
Q: How many tips do I need to judge a tipster?
A: At odds around 2.0, 100 tips gives an ROI uncertainty of roughly ±20 percentage points, which is far too wide. Treat anything under 500 tips as mostly noise, and expect to need 1,000–2,000 tips to confirm a modest edge. At higher odds, you need several times more.
Q: Why is CLV better than win rate for judging a tipster?
A: Win rate depends entirely on the odds: 70% at 1.20 loses money, while 30% at 4.0 makes money. CLV compares the price taken with the closing price, which controls for odds and shows skill with far fewer bets than profit does.
Q: Is a long winning streak proof of tipster skill?
A: No. Five wins in a row at 50% happens 1 time in 32. With thousands of tipsters posting daily, long streaks happen every week by chance alone.
Q: Are paid football tipsters worth it?
A: Only if their verified edge is large enough to cover the subscription. At €30/month and €20 stakes, a genuine +3% edge needs about 600 tips a year just to cover the fee.
Q: What is the biggest red flag in a tipster's record?
A: A record built on screenshots, with no timestamped, independently verified history. If losing tips can be deleted, the record is meaningless.
Bottom line
A tipster's record is a small, noisy sample presented by someone with a financial interest in making it look good. Check the sample size, the odds, the CLV and the verification, and most "proven winners" disappear.
P.S. You can run every check in this guide on your own betting too. Log your bets, record the closing odds, and give it 500+ bets before you decide whether your method works.