Skip to content

How to check a tipster: 7 steps and red flags

By BetBuddyRank editorial team

BetBuddyRank runs this check automatically for 434 tipsters on X: 50,197 picks graded against final scores, each linked to the original post. You can look a capper up by handle with the tipster check; below is how to do the same by hand and what to look for in a verified record.

Tipster checklist

StepWhat to checkRed flag
1Every pick in order for a monthOnly screenshots of winners
2Post time before kick-offResults posts after the fact
3Outcome of every pick from the scoreGenerous grading of half wins
4Hit rate against the probability implied by the oddsHigh percentage on short prices
5ROI with 1-unit stakesNo ROI, or ROI on “bankroll”
6Sample sizeFewer than 50 picks
7Parlays, deleted posts, promisesGuarantees and “fixed matches”

Step 1. Collect every pick in order

The most common mistake is judging a tipster by what they choose to show. Screenshots of winning slips, pinned “+15 units this week” posts and recaps with check marks are almost always cherry-picked. Open the timeline and list every pick for the period, in order, including the ones the author never mentioned again.

A calendar month works well: it shows both the streaks and the slumps. If a capper posts picks as images, you will have to grade them by hand, because automatic records do not count image-only posts.

Step 2. Check when each pick was posted

A pick counts only if the post went up before kick-off. Hover over the post date on X to see the exact time. Edited posts are labelled and deleted ones disappear, so save the links right away.

A separate red flag: results posts such as “5 of 6 hit yesterday” with no picks posted before the games. The capper leaderboard ignores them.

Step 3. Grade every pick against the score

Record the outcome of each pick: win, loss or push. Mind the market rules: football (soccer) totals and handicaps use regular time, NFL, NBA and MLB scores include overtime, and Asian quarter lines produce half wins and half losses. All grading rules are in the tipster rating methodology.

Step 4. Compare the hit rate with the odds

A hit rate on its own tells you nothing. For each pick, work out the probability implied by the odds: 1 divided by decimal odds. The sum of these probabilities is the expected number of wins. A capper who wins 70 of 100 picks when 72 were expected is losing to the market, however good 70% looks.

Odds (decimal / American)Break-even hit rateTypical market
1.50 / −20066.7%football favourites
1.91 / −11052.4%NFL and NBA spreads and totals
2.00 / +10050.0%even money
2.50 / +15040.0%underdogs
3.00 / +20033.3%draws, long underdogs

To turn American or fractional odds into a probability, use the odds converter.

On X, 31% of picks come with odds, and 52 of 295 tipsters with 10 or more picks never post them. Without odds, the expectation has to come from team strength, which is how the tipster index works.

Step 5. Compute ROI on picks with odds

ROI is total profit divided by the number of bets, with the same stake on every pick. It captures both the hit rate and the prices, so it is the key money metric. The formula and realistic numbers are in the guide to ROI in sports betting.

In our data, the 97 tipsters with 30 or more priced picks have a median ROI of −7.8%. 30 are in profit, and only 5 show a profit that luck cannot explain at 95% confidence.

Step 6. Judge the sample size

Over a short run, luck matters more than skill. We tested this on 144 tipsters with 60 or more picks by splitting each record into two halves in time. The top quarter by the first half had an edge of +10.4 pp; in the second half it kept only +1.2 pp. The bottom quarter, at −10.0 pp, moved back to almost zero: −0.3 pp.

A past record does say something about the future, but about half of a hot start turns out to be luck. The tipster index accounts for this by shrinking small samples.

PicksWhat you can conclude
under 50almost nothing: luck decides
50-100early signs, but one streak changes the picture
100-300a real signal: look at the edge over expectation
300 and morea confident verdict

Step 7. Look for red flags

  • Only screenshots of winning slips, no timeline of picks posted before the games.
  • Deleted losing posts and edited picks.
  • A claimed 70-80% hit rate at odds of 1.80 (−125) or longer: it does not hold over time.
  • Promises of “fixed matches” and guarantees.
  • Paid access with no public record you can verify.
  • Picks with no market or odds: “Real Madrid today” without saying what exactly.
  • Long parlays: parlays of 6 or more legs from tipsters in our database hit 0.7% of the time against 3.7% expected.

FAQ

How many picks do you need to judge a tipster?

Under 50 picks tell you almost nothing, from 100 there is a signal, and from 300 you can judge with confidence. In our data the best tipsters lose about half of their edge in the second half of their record.

Can you trust screenshots of winning bets?

No. A screenshot shows one bet and nothing about the losses. Check the full timeline of picks posted before kick-off.

What if a capper deletes losing picks?

Save the links to posts before kick-off and check them afterwards. Deleting losses is a red flag in itself: an honest capper has nothing to hide.

Why does a tipster have a high win rate but lose money?

Because they back short prices. At odds of 1.40 (−250) you need a 71.4% hit rate to break even, so 70% over a long run means a loss.

Are paid capper picks worth it?

Price is not a quality signal. Judge only a long public record with a positive ROI over hundreds of bets: in our data such tipsters are rare, and past profit does not guarantee future profit. 18+.

Getting started

  1. Look the capper up with the tipster lookup by handle.
  2. If they are not in the database, grade their picks by hand using the steps above.
  3. Compare them with the best football tipsters or the tipster leaderboards by league.
  4. Follow new picks through the RSS feed on the tipster page.

Sources

Public tipster posts on X; final match results from LiveScore and ESPN; BetBuddyRank calculations over the site’s open data. Verified September 2026.