We Audited Every "3-0 SWEEP" Thumbnail on YouTube. Zero Checked Out.

Open YouTube, search "MLB picks today," and count the thumbnails screaming a record at you. 3-0 SWEEP YESTERDAY. 4-0 RUN. Green check marks, broom emoji, the works.

We wanted to know what those numbers actually describe. So we built the boring version of the answer: since May 5, 2026 we have pulled 386 videos from 9 capper channels, extracted 3,358 individual picks from the audio and the on-screen bet slips, and settled 1,676 of them against MLB box scores.

Then we did the one thing nobody does. When a thumbnail says "3-0 yesterday," we went and graded every pick in yesterday's video.

We found 8 testable record claims. Zero of them matched.

The receipts

Here is every claim where we had a large enough sample of the previous day's card to say something real. The "actual" column is the full settled record of that channel's picks for the day the thumbnail is referring to.

Look at the first row again. The July 19 video contained 38 picks. Twelve won. Fourteen lost. The next morning the thumbnail said 3-0.

Both statements can be true at the same time. That is the entire trick.

How "3-0" and "12-14" are the same day

Nobody has to lie to put 3-0 on a thumbnail. They just have to decide, after the games are over, which three picks were the card.

A ninety-minute video with 38 picks in it gives you an enormous menu. Three of them went 3-0. Some other three went 0-3. If you name your winners the next morning, you will never have a losing day again, and you will never have said a false sentence on camera.

That is not a hot streak. That is a reporting choice, made after the result is known, in the one place a new viewer looks before they decide whether you are good: the thumbnail.

The tell is the ratio. If a capper posts 3 plays and goes 3-0, that is a real 3-0. If a capper posts 38 plays and the thumbnail says 3-0, the number is not describing the day's performance. It is describing the day's best three.

What those channels actually run at

Full settled records across everything we have graded from each channel since May:

Guy Boston Sports is the volume leader of the group with 541 settled picks, and it is running at a coin flip. On the 130 of those picks where a price was actually stated on camera, the number comes out to -15.8% ROI at an average price of -52.

That is the honest headline for that channel. It has never once appeared on a thumbnail.

What we are not saying

We are going to be precise here, because precision is the whole product.

We are not saying these channels invented a 3-0. We cannot see their private card, and it is entirely possible that three specific plays they consider "official" did go 3-0 on those dates. What we can prove is the thing that matters to you as a viewer: the video that number is attached to contained 38 picks, and the full set went 12-14. The headline is describing a slice, and nothing on the thumbnail tells you it is a slice.

We also pulled four other claims out of this piece rather than pad the list. Two came from days where we had only 4 settled picks on record, which is too thin to call anybody out. One came from a day our own settlement had not finished. One was a 3-1 that could plausibly be a real 3-0 with an extra pick of ours in the pile. If we are going to hold cappers to receipts, the receipts have to survive us being hard on our own data first.

The one channel that never did this

Worth saying out loud: StumpTheSpread has the largest settled sample in our index at 722 picks, is running 55.0%, and has made zero record claims on a thumbnail. It just posts the picks.

The channel with the best number is the one not advertising a number. That is usually how it goes.

How to check any of this yourself

Every pick, every settlement, and every claim in this article is on our YouTube capper audit pages, broken out per channel and per video, with the source video linked on every row. If we graded something wrong, it is visible and you can see exactly which pick.

Picks are extracted by a model watching the full video, so it catches plays spoken aloud and plays that only ever appear on an on-screen bet slip. Settlement runs against MLB Stats API box scores. No editorial, no adjusting, no quietly dropping a loser.

That last part is the point. A capper's record is only worth something if somebody other than the capper is keeping it.