PRESS & MEDIA KIT
data, facts and contact for journalists covering sports betting
THE SHORT VERSION
We run an automated collector that captures sports betting picks the moment
they are posted publicly, then grades them against official results. That produces an independent
performance record for handicappers who otherwise grade themselves. As of 2026-08-24 we hold
31,972 captured picks from 4,441 authors, of which 9,893 have reached a final
result. The data is free, and we will run custom cuts for reporters at no cost.
Data available to journalists
We will run bespoke queries against the dataset for any working journalist, free, with no conditions,
no embargo requirements and no approval rights over the story. Things reporters commonly ask for:
- A named handicapper's independently graded record — what they actually went,
against what they claimed, including picks they later deleted.
- Performance by sport, market or price band for any date range in the window.
- Platform comparisons — how picks posted to Reddit, YouTube or X performed.
- Claimed-record versus graded-record gaps for the YouTube channels we audit. We hold
473 audited videos and extract the record claims made in titles, thumbnails and narration.
- Custom aggregates for a specific angle you are working on.
Turnaround is usually same day. If we cannot support a claim with the data, we will tell you that
rather than stretch it — see the note on limitations below.
Current research
The State of Sports Handicapping 2026
An analysis of 6,906 graded picks from 1,384 handicappers, January to August 2026. Headline findings:
aggregate return of −0.58% on a 51.77% win rate; the share of profitable
handicappers falls as sample size grows (57.1% at 10+ picks down to 47.1% at 50+); and a
profitable first ten picks did not predict the next twenty.
The report publishes its own exclusion filters and flags every figure that rests on thin sample.
It is licensed CC BY 4.0
— free to reproduce, including commercially, with attribution.
READ THE REPORT
DOWNLOAD CSV
Fact sheet
| FACT | AS OF 2026-08-24 |
|---|
| Picks captured | 31,972 |
| Distinct authors captured | 4,441 |
| Picks graded to a final result | 9,893 |
| Handicappers with a graded pick | 2,192 |
| YouTube videos audited | 473 |
| Picks extracted from those videos | 4,057 |
| Collection start | January 2026 |
| Sports covered | MLB, NBA, NFL, NHL, NCAAB, soccer,
tennis, cricket, MMA, esports, WNBA |
How the data is produced
Three stages, all automated:
- Capture. Collectors monitor public sources and record a pick when it is posted,
with its timestamp, price and stated stake. This happens before the event starts, which is what makes
the record independent — nothing is entered after an outcome is known.
- Retention. Captured picks are retained even if the original post is later edited
or deleted. Deletion of losing picks is a real behaviour in this industry and a record that quietly
loses those is not a record.
- Settlement. A separate job grades each pick against official results and computes
profit at the price as posted.
The full pipeline description is at /methodology.
WHAT WE WILL TELL YOU ABOUT OUR LIMITATIONS
We would rather be quoted accurately than quoted a lot. Our coverage is not a
random sample of the industry — we see what is posted publicly, so paywalled-only sellers are
underrepresented. Grading is to the posted price, not the closing line. Automated extraction of picks
from text and video is imperfect and we exclude rows that fail to parse rather than guess at them. Every
published figure carries its sample size, and where n is small we say so in the same sentence.
Boilerplate
SHORT (25 WORDS)
CAPTRACKER independently grades sports betting picks captured from public posts before events start,
producing verified performance records for handicappers who normally grade themselves.
LONG (60 WORDS)
CAPTRACKER is a sports betting analytics service that captures handicappers’ picks at the moment
they are published, then grades them against official results. Because collection happens before the
event and retains posts that are later deleted, it produces an independent record that cannot be edited
after the fact. As of August 2026 it holds 31,972 captured picks from 4,441 authors.
HOW TO REFER TO US
CAPTRACKER (all caps), at captracker.markets. Not “CapTracker”, not “Cap Tracker”.
Usage and licensing
- Figures from The State of Sports Handicapping 2026 and
the accompanying CSV are CC BY 4.0. Reproduce them freely, including commercially, with
attribution to CAPTRACKER and a link to the source page.
- Screenshots of the site may be used in editorial coverage without asking.
- Custom data cuts we prepare for you are yours to publish. We ask only that you attribute the source
and link to
captracker.markets.
- We do not require sign-off on stories and will not ask to see copy before publication.
What we are not
Worth stating plainly, since it affects how you would characterise us:
- We are not a sportsbook and take no bets.
- We are not a tout. We do sell a paid subscription to the analytics product, and we
run a betting bot whose record is published separately from the handicapper dataset and is never mixed
into it.
- We are not accusing named individuals of fraud. The research is about a population.
Where a specific handicapper's claimed record differs from their graded record, we will show you the
data and let you do your own reporting on it.