CAPTRACKER
HOME› BLOG› This Week's ROI Math: Verified Leaderboard
DATA REPORT September 06, 2026 Rachel KimBY RACHEL KIM

This Week's ROI Math: Verified Leaderboard

This Week's ROI Math: Verified Leaderboard

I check the CapTracker leaderboard the way some people check the weather before they've even gotten out of bed. Not because I trust it blindly, I just like knowing what changed overnight. This week's board has a run of ROI figures that look almost fake: 930.6% and 618.0%. Those are real numbers off real settled picks, but real doesn't mean repeatable, and that gap is the whole subject of this column.

This week's top verified handicappers leaderboard, ranked with a gut check

The current window shakes out like this, best ROI first, with my honest read on each line. Small samples get called out because they matter more than the percentage sign sitting next to them.

How a losing record still tops the ROI column

Start with the vocabulary, because it matters here. ROI on CapTracker measures net units returned against units actually risked, and it reflects the real odds on each pick rather than a flat bet size. At standard -110 pricing, break-even sits at 52.38%. Anything posted far above that on a handful of picks is usually one bet doing most of the talking.

Smart_Competition_10 went 1-4 this week. One win, four losses. By record alone, that's a losing week, full stop. Yet the board shows +30.9 units and 618.0% ROI, which only happens if that single winning pick was priced as a real underdog, big enough to cover four losses and still leave a profit behind.

That's not a flaw in the math. It's exactly how plus-money underdogs behave: rare wins, oversized payouts when they land.

Flashback2500 tells a cleaner version of the same story with a winning record attached: 4-1 over five picks, +46.5 units, 930.6% ROI. Four wins and a loss shouldn't usually clear 900% at typical pricing on its own, so at least one of those wins came at a serious price. Good picking, sure. But five picks is a snapshot, not a sample.

Compare that to X_Underscore_X, the football name logging 51 picks this week. A 51.6% ROI on that kind of volume looks almost tame next to 618% or 930%, and that contrast is the whole tell. Big samples rarely produce explosive percentages, because one outlier pick gets diluted by fifty others sitting around it.

I've stared at boards like this for years, and the pull to round a five-pick heater up to "this person has it figured out" never fully goes away. It just gets a little easier to catch yourself doing it.

What CapTracker's season pool says that one week can't

Zoom out to the full pool and the picture gets a lot less dramatic, which is the point of zooming out. Across CapTracker's tracked pool, 988 handicappers have combined for 5,796 settled picks this season, and 489 of them, 49%, are net profitable. That sits close to a coin flip, and it should. Beating the closing line over and over is hard, and most people who try land near even money once enough picks pile up.

Narrow that to cappers with at least 15 settled picks, a reasonable floor before taking a record seriously, and 35 of 65, or 54%, are profitable. That's better than the full pool's rate, but still not far from a coin flip. It tells you something useful anyway: reaching a profitable season happens often enough to be real. Staying there is the actual test.

The season's top names back this up. X_Underscore_X leads with 27-17-7 over 51 picks, +389.8 units, 51.6% ROI, the same name sitting near the top of this week's board too. A hot week from someone with a hot season already behind them reads differently than a hot week from a name you've never seen before.

Volume by itself isn't proof of skill either. SportsBettingPicks has logged 548 picks this season and sits at -8.3 units. GuyBostonSports has thrown 490 picks and is down 56.6 units. CallingOurShot has 140 picks logged and is down 14.5 units.

Big samples clear out the noise. They just don't automatically point up.

How to read next week's handicappers leaderboard without getting fooled

None of this makes the weekly leaderboard useless. It means you read the sample size sitting right next to the percentage before you read the percentage itself. CapTracker auto-settles every pick against ESPN results, and once a pick is timestamped, it can't be edited or deleted after the fact. That's what lets a number like 930.6% ROI stand as an honest record of what actually happened, even on a week you wouldn't bet your season on.

The site tracks more than 900 handicapper profiles, and the free leaderboard ranks every one of them by units won and ROI, so this week's movers and the season-long names sit on the same page. Check the daily feed if you want to watch picks land in real time instead of catching them after the fact. And if ROI or closing line value still feels fuzzy, the glossary walks through the terms without the jargon.

One more thing worth tracking: convergence. When independent handicappers land on the same side without coordinating, that agreement carries more weight than any single hot week does, small sample or not. It gets covered in more depth on the methodology page, along with how the site handles voided picks and ROI math generally.

Bookmark the leaderboard. It won't tell you who to trust off one week alone, but it's free, and it updates against real, ESPN-settled results you can check yourself. Come back often enough and you'll have enough history to tell a five-pick heater from an actual track record. That's the whole game.

FREE TOOL

Track every pick. Verify every record. See which cappers actually beat the line, automatically settled by ESPN data.

VIEW LEADERBOARD →