Tennis, UFC, cricket, college football, and NFL predictions, graded in public

The edge: every call has a receipt

Before a ball is served, a cage door closes, or the coin goes up at the toss, we lock in a pick and a probability. After the result it gets marked right or wrong, and it stays on the record either way, wrong ones with a red badge.

That’s the whole pitch. Anyone can tweet a winner after the fact. What we keep is a public track record of every call we’ve ever made, win or lose, at every confidence level. That’s the difference between a receipt and a brag.

Who will win — and how

Five sports are live today — tennis, UFC, cricket, college football, and the NFL. Each gets a model built for how that sport actually plays out.

Tennis picks

We rate every player on how well they actually play — kept separate by surface, because a clay grinder and a grass-court server aren’t the same threat. The gap between two players becomes a win probability for their next match, tuned for the quirk that makes tennis tricky: you can win more points than your opponent and still lose. See today’s tennis picks.

UFC predictions

For every fight we rate, our model scores both fighters on what actually decides fights — age, reach, style, form, finishing ability — and makes its own call: who wins, and how it ends (KO/TKO, submission, or decision). When our pick goes against the betting favorite, we say so — that’s an upset call, on the record like everything else. See today’s UFC predictions.

Cricket predictions

For every T20 and ODI we cover, the model prices a side and projects a first-innings total — both locked in before the toss, no exceptions, so the number never leans on knowing who bats first. When the model can barely split two sides, we say “toss-up” instead of faking conviction. See today’s cricket predictions.

College football predictions

Every week of the season, we fit power ratings to more than a decade of game results and publish one number per game: the chance our pick wins, straight up. “Ohio State, 79% to win” is the whole call — nobody’s line to beat and nothing to shop. Our own projected spread and total ride along as context, written the way you read a spread anywhere else, but only the win probability is graded. See our college football predictions.

Here’s what makes that number useful when you’re weighing a decision. It’s calibrated: tested against more than six thousand past games, a stated 68% has come true about 68% of the time — an independent second opinion you can hold your own read against, not a vibe with a percent sign. It’s audited: every number lands on the public track record after the game, so you never have to take the calibration on faith. And it’s selective: games we can’t clear a 55% confidence floor on don’t get published, and neither do the few where our number sits far enough from the market consensus that our own history says we’re probably the ones wrong. Knowing when we have no call is part of the product — a coin flip sold as conviction helps nobody.

NFL predictions

Every week of the season, one number per game — the chance our pick wins, straight up — for every game on the slate. The NFL board is the one place we never withhold: there is always a pick side, so the closest games publish just past 50% instead of being left off, and our confidence is capped at 90%, the edge of the calibration table the model was validated on. Our own projected spread and total ride along as context, only the win probability is graded, and the one or two ties a season are void — they never count for or against the record. See this week’s NFL predictions.

And what we don’t claim: we don’t publish sportsbook numbers, we don’t promise an edge over closing lines, and nothing here is betting advice. What you get is our honest, checkable read — the rest of the decision is yours.

The percentages are graded, not promised

A number only means something if it’s honest. So every percentage we publish gets checked against what actually happened, and the whole thing lives on the public track record— the confidence levels where we nail it, and the ones where we’ve run hot. Tennis, UFC, cricket, college football, and the NFL are all graded there. That’s the difference between a number you can build a read on and one you just have to take on faith.

We publish predictions, not bets. No units, no “lock of the day,” no telling you how much to put down. If you play, play within your limits — the picks are here to sharpen your read, not to chase.

Frequently asked questions

How do you make your picks?

Every pick comes from our model, not a hunch. In tennis we rate players on how they actually perform, surface by surface. In UFC we rate every fighter on the things that actually decide fights — age, reach, style, form, finishing ability — and call the winner ourselves, plus how it ends: KO, submission, or decision. In cricket we price a side and project a first-innings total before the toss, no exceptions. In college football we fit power ratings to over a decade of game results and publish a calibrated win probability for every game we have a call on. In the NFL we publish a calibrated win probability for every game on the weekly slate — the full board, capped at 90%. Every number gets graded against the real result afterward.

Do you actually track your record?

Yes. Every pick is frozen before the event and graded against the real result on our public track record, wins and losses alike. Misses stay up with a red mark, and a published pick is never edited after the fact.

What sports do you cover?

Tennis, UFC, cricket, college football, and the NFL — a pick for the tennis matches our model rates, every UFC fight we can rate, every cricket match we cover (T20 and ODI), a weekly college football board with a win probability on every game we have a call on, and an NFL board that calls every game on the weekly slate.

How often are picks posted?

Daily for tennis and cricket — fresh picks go up for each day's matches, and cricket picks always land before the toss. For UFC, predictions land ahead of every event for the fights we rate, main card and prelims alike. College football and the NFL publish weekly: each board goes up early in the week and every game is graded after it's played.

What does a college football win probability actually mean?

When we publish "68% to win," that number has been tested against more than six thousand past games, where numbers like it came true at their stated rate. It's a calibrated estimate to hold your own read against, and our public track record shows whether the calibration is still holding.

Why are some college football games missing from the board?

Because we only publish games where we actually have a call. Anything our model can't clear a 55% confidence floor on is left off entirely — a coin flip published as a pick would just be noise. We also withhold the handful of games where our number lands far from the market consensus, because our own history says those are the calls we're most likely getting wrong. A shorter board is the honest board.

Do you cover every NFL game?

Yes — unlike college football, the NFL board carries every game on the slate. There is always a pick side, so the closest games publish just past 50% rather than being withheld, and our confidence is capped at 90%, the edge of the calibration table the model was validated on. The one or two ties a season are void: they count neither for nor against the record.

Do you tell me who to bet?

No. Maiden publishes predictions, not betting advice. We don't tell you what to stake and we don't promise profit. We give you our read on who wins and how — you make your own call.

Do you win every pick?

No, and we'd never claim to: nobody wins them all. Every percentage we publish is graded against the real result on our public track record, which shows how our picks at each confidence level have actually held up. Tennis, UFC, cricket, college football, and the NFL are all graded there.