To verify a tipster or handicapper you need three things: every pick they made, posted before the game with the odds they took; the return on investment (ROI) calculated at those odds, losses included; and enough picks to separate skill from luck, which means several hundred. If the odds or the full record are missing, the tipster cannot be verified, whatever win rate they advertise.

What does a tipster need to publish to be verifiable?

• Date and time of every pick, before kickoff or tip-off, somewhere it cannot be edited afterwards (a channel with timestamped posts, a read-only shared sheet).
• Game, market and selection, with no ambiguity: "over 2.5 goals, regulation time", not "goals game".
• The odds taken when the pick was posted. It is the data point most often missing, and the one that decides whether a pick makes money.
• The stake in units, always on the same scale (for example 1 unit = 1% of the bankroll).
• The result of every bet, losses and voids included, not just screenshots of the winners.

Most tipsters never publish the odds they took or a complete record. That is not necessarily bad faith, but without those two things their return cannot be checked.

How do you calculate a tipster's ROI?

1. Add up the units staked across every settled bet.
2. Work out the profit of each bet: a win pays units × (decimal odds − 1); a loss costs the units; a void is zero.
3. Divide the total profit by the units staked: that is the ROI, the return on turnover.
4. Compare it with the average odds: at -133 (1.75) you need to win 57.1% of bets (1 ÷ 1.75) just to break even.

An example: 200 picks of 1 unit, 120 of them winners at an average of -133 (1.75). Profit: 120 × 0.75 − 80 = +10 units on 200 staked, an ROI of +5%. The same 120 winners at an average of -167 (1.60) would give 120 × 0.60 − 80 = −8 units: the same 60% win rate, and a losing record.

How many picks do you need before you trust a record?

More than you think. Take a tipster with no edge at all who always bets outcomes at fair odds of +100 (2.00): in the long run their ROI is zero. Over a few picks, though, luck can make them look good:

Picks posted	Chance that a tipster with no edge shows an ROI of +10% or better
50	24.0%
100	18.4%
250	5.7%
500	1.4%
1,000	0.1%

After 100 picks, almost one tipster in five with no edge whatsoever is showing +10%. Below 300-500 bets the ROI says little; what matters more is whether the odds taken beat the fair odds, which is the subject of What is CLV?

What does a verifiable record look like?

For comparison, here is the record of the ValueBets NinjaBet members log in the Bet Tracker, published with every loss on the ValueBet results page (as of October 16, 2026, settled straight bets since August 1, 2024, all NinjaBet sites):

Figure	Value on October 16, 2026
Settled bets	21,793
Return on turnover	+9.59%
Profitable months	25 of 27
Worst month	January 2026, −2.73%
Best month	March 2026, +47.81%

These are not numbers to promise anyone: members enter them themselves and they swing a lot from month to month. But it is the format to ask a tipster for: every bet, the return on turnover, the good months and the bad ones.

Which warning signs should make you suspicious?

• A win rate with no odds, or with the best price that could be found the next day.
• Words like "lock", "can't lose", "100%": no sporting result is.
• A record that resets every month, or that is made only of screenshots of winning slips.
• Returns promised up front in exchange for a VIP group subscription.
• A simple question with no answer: if a tipster had a lasting edge, they would make more by betting it than by selling it for $20 a month. Why sell it?

How do you check a single pick you are sent?

Compare the price you are offered with the fair odds of the same outcome, the consensus price of the worldwide betting market. MarketOdds shows them for every selection and the ValueBet calculator gives you the edge: if the pick's price is below the fair odds, the pick has no edge, however the game ends. The full method is in What is a value pick?

Want to see it in practice? +EV picks shows members' verified picks and their public record, Check a pick prices any tip you are given, and the Telegram channels get them first.

Frequently asked questions

What ROI is believable for a tipster?

Over thousands of bets at average odds, a few percentage points is already an excellent result. A double-digit ROI over a few dozen picks says more about luck than about skill.

Isn't the win rate enough?

No. Without the average odds you cannot tell whether it makes money: 70% winners at -250 (1.40) lose 2% of turnover, while 40% winners at +200 (3.00) make 20%.

Is a record on a verified-tipster site reliable?

It is better than nothing if it records the odds at posting time and does not let picks be deleted. Still check the number of bets and the return on turnover.

Why should a tipster publish their losses?

Because a return is calculated over every bet. Leaving out the losers turns any record into a success.

How many units should you stake on a pick?

Few, and always the same relative to your bankroll: 1-2% is the starting point. The Bankroll and Kelly guide explains how to adjust them.

Sources

• NinjaBet, ValueBet results, figures as of October 16, 2026.
• NinjaBet calculation: binomial probability of winning at least 55% of independent picks at fair odds of +100 (ROI ≥ +10%), for 50-1,000 picks.

You must be 21 or older (18 in a few states) to bet in the United States, and bet only with operators licensed in your state. Gambling can be addictive: if you or someone you know has a gambling problem, call or text 1-800-GAMBLER, or the National Problem Gambling Helpline at 1-800-MY-RESET (ncpgambling.org), free and confidential, 24/7. Results vary and require time and capital; this article is information, not financial or legal advice. Last updated: October 16, 2026.