A record shown after the games can be cut to the good ones.

Five checks, and what each one misses

1. 1.Was each pick locked before its own game?

   Look for a time on each pick, and check it is earlier than that pick's own game, not just earlier than the first game on the card.

   Catches

   The late add: a pick slipped onto a card after the early games are final, borrowing the card's early time.

   Does not prove

   That the pick was any good. And a cropped screenshot proves nothing: only a time the platform itself records counts.

2. 2.Is the price the one on offer at that moment?

   Each pick should carry the odds that were available when it locked, and say where they came from.

   Illustration, made-up numbers, no real tout: 10 wins and 8 losses at 1 unit each is +1.1 units at -110, and -1.3 units at -150.

   Catches

   A record graded at -110 when the pick was really available at -150. Every win then counts for more than it paid.

   Does not prove

   That you can still get that price when you read the pick, or that a bet was placed at all.

3. 3.Can you see the whole record, losses included?

   Ask for every pick since the first one, in one list, with the losing ones still in it.

   Catches

   The filtered run: "9-2 this week", one sport only, a fresh start after a bad month, or losing posts deleted.

   Does not prove

   That nothing is missing. Where posts can be deleted, a deleted post cannot be detected; only a record nobody can edit afterwards settles that.

4. 4.Are units counted at the stated odds?

   A win rate means nothing without the price. At -110 a bettor needs about 52.4% just to break even; at -200 it is 66.7%.

   Catches

   A high win rate built on heavy favourites that still loses money, and unit sizes raised on the winners after the result.

   Does not prove

   That a unit means the same to two touts, or that someone following along would have matched it.

5. 5.How did the prices compare with the closing line?

   Compare the price at each pick with the last price before the game started. Over many picks, prices better than the close are the earliest useful sign of skill.

   Catches

   A hot streak sold as skill. Win-loss alone is slow: after 100 picks, a coin-flip bettor lands between about 40% and 60% wins, 19 times in 20.

   Does not prove

   Profit, or future results. A good price can still lose, a short run of good prices can still be luck, and the check only means something if the price was captured when the pick locked (check 2).

How SharpUnit runs the same checks

The five checks, done by a system instead of by eye. A creator sends picks before the games; anyone can read the result.

1. 1.Each pick is locked before its own game, or for fights before its own fight. A pick sent after the card started but before its own fight is labelled exactly that way.
2. 2.The price is captured at the lock, with the source it came from.
3. 3.The closing price is recorded where one is available, so check 5 can be run.
4. 4.Each pick is graded in public, and the losses stay on the ledger.
5. 5.A correction is added to a public log. Nothing is edited or deleted.

Where it stands, October 30, 2026: no outside creator audit is published yet. Our own picks run on the same system as a first-party record, not an independent one, and the record shows whether it is up or down.

No rankings, no leaderboards, and no tout pays to be checked.