Methodology

Twitter is full of cappers posting picks. Most of them claim records you cannot verify. That is the problem TailSlips solves.

We track every public pick from a curated list of sports-betting accounts. Picks are graded against the final outcome of the game, at the odds and units the capper actually posted. The leaderboard is the receipts, every account judged by the same rules.

Why this exists

A capper who tweets ten picks and only references the wins is selling a different account than a capper who tweets ten picks and grades every one of them. Most public-facing capper records are the first kind. TailSlips is built around the second. Every public pick a tracked account posts is captured at the moment it goes live and graded later, win or lose, at the odds and stake the capper actually committed to.

That means the leaderboard reflects what really happened, not what got highlighted. A capper sitting at the top of TailSlips is at the top because their full body of work, including the picks they would rather forget, nets out positive.

What counts as a pick

A tweet becomes a graded pick when it contains a clear wager: a side or selection, a line where applicable, the odds taken, and the unit stake. The four pieces of information that make a wager actionable are the four pieces TailSlips needs to grade it. If any are missing, the pick is logged for review rather than auto-graded.

• Public picks posted on the open timeline
• Captured at the moment the pick is posted, not retroactively
• Locked at the odds and units the capper posted, not adjusted later
• Graded against the actual final outcome of the game
• The same grading rules applied to every tracked account

What doesn't

Not every tweet that mentions a team or a line is a pick. The leaderboard would be useless if it were padded with stat sheets, hot takes, and giveaway promos. The rules below define what stays out.

• Picks behind paywalls, in subscriber DMs, or in private Discord rooms
• Picks that lock the wager terms behind paid-content language
• Tweets that name a subject but hide the side, line, odds, or stake
• Picks edited or deleted after the game starts
• Anything ambiguous enough that a clear wager cannot be identified
• Picks tied to a postponed game, which are voided rather than graded
• Stat sheets and probability tables, even when they list players and numbers
• Pure giveaway and engagement-bait posts
• Futures, season-long, and award bets that don't resolve on a single game

How grading works

Once a game finalizes, every pick tied to it is graded against the box score. A win pays out at the posted odds, a loss is the unit stake, a push returns the stake, and a void (postponed, suspended early, or an outcome that the wager cannot resolve against) takes the pick out of the record without counting as a result either way. The math is the same math any sportsbook uses to settle a bet. There is no ranking adjustment, no handicapping, no hindsight grading.

When a capper doesn't post the odds

Some cappers post a moneyline pick without specifying the American odds they took. Rather than fabricating a default, we grade those picks against the Pinnacle closing line for the same game and side. Pinnacle is widely accepted as the sharpest book, and the close is the consensus probability at first pitch, both factors that make it the fairest available proxy when the capper has not given us a number.

Picks graded this way are marked with a small "(close)" indicator next to the odds in the history table, so you can see at a glance which lines were graded against literal posted odds and which were graded against the Pinnacle close. This is moneyline only. Spreads and totals cluster near -110 in practice, so when those markets are posted without odds we use a -110 default and disclose it the same way.

Player props are different. Prop odds vary widely between players (a star hitter's 1+ hits is priced very differently from a bench bat's), and Pinnacle's prop coverage is sparse. So when a capper posts a player prop without committing to specific odds, we grade it on outcome only: the pick still counts toward Win %, but it is excluded from the unit profit and ROI calculation. Those rows show "no odds" in the history table where the odds and profit would normally be. The capper gets credit for being right or wrong, without a fabricated payout shaping the bottom line.

Parlays

Parlays are graded as a single wager against the combined odds the capper posted. Every leg has to win for the parlay to win. If any leg pushes or voids, that leg drops out and the parlay is recomputed against the remaining legs. The parlay's line in the record is one entry, not one per leg.

Reposts and edits

Cappers sometimes delete and repost the same pick to fix a typo, swap a header, or clean up the formatting. TailSlips treats those as the same wager and only counts them once. What does not count as a credibility-preserving repost: deleting a pick because the line moved against you and not putting it back up.

Doubleheaders and game binding

When two teams play twice in one day, picks are bound to the specific game the capper named (Game 1, Game 2, or by pitching matchup). When a player prop names a player whose team plays a doubleheader, the prop binds to the game the named player started. Wrong binding is a real failure mode and gets caught in the audit pass before grades go public.

Same rules, every account

The model entry on the leaderboard, the FADE AI account, is graded by the exact same rules as every human capper. No special treatment, no hand-waving on its losses. If something disqualifies a pick from a human capper's record, it disqualifies the same pick from the model's.

What you can verify

Every pick on every profile links back to the original tweet. Click through and you see what the capper posted, when they posted it, and what odds and units they committed to. The grade you see is computed from the final game outcome a public sportsbook also has on file. There is no hidden math.

The full pick history for every tracked account, wins and losses, is on their profile.

If you're a capper

The number on your TailSlips profile will almost never match the spreadsheet you keep yourself. That is intentional, and the reason it differs is the whole point of this site. A few specifics worth knowing if you arrive here ready to dispute something.

Your record is computed from publicly posted, fully specified picks. A tweet has to declare the side or selection, the line where applicable, the odds taken, and the unit stake before we can grade it. Picks that hide any of those behind paid content, lock emojis with no number attached, or vague language are kept out of the record. Most missing picks fall into this bucket, and they are missing on purpose.

We do not retroactively delete losses. Once a pick is captured and the game finalizes, the grade is permanent. Deleting the original tweet does not pull the row off your record. If anything, deleting a losing pick is detected and surfaced as a credibility signal of its own.

Moneylines without posted odds are graded at the Pinnacle close. Spreads and totals default to -110. Player props posted without odds count toward your win rate but stay out of units profit. All three are disclosed on the row itself. See the "When a capper doesn't post the odds" section above for the full rationale.

If you find a genuinely misattributed pick (wrong game binding, wrong player, parser misread of your odds), email corrections@tailslips.com with the tweet link and what should be different. Real errors get fixed, usually within a day. We do not honor requests to remove losing picks, fabricate odds the tweet did not declare, or recategorize stakeless "locks" as graded wagers.