Your daily sports hub.

A model that shows its work

MatchWiz is an independent sports analytics site: one statistical model that scores every matchup on the day's slate, publishes its projections next to the market's numbers, and then grades itself against the real result — in public, every day. No hot takes, no hand-picked winners, no highlight reel. The same math rates every player and every game, and the full archive rolls back date by date so you can read exactly what we said and what happened.

For baseball that means daily matchup boards for hitters and pitchers, DFS projections and a free lineup optimizer, live scores and standings with simulated playoff odds, and a page for every player. Golf gets a live win-and-finish model every tournament week. Where the WizLine disagrees with the market enough to matter, it surfaces as a Play of the Day — graded at the posted line, hit or miss.

Everything here is a research signal, not betting advice — the methodology, the glossary, and the story behind the site are all public, because the record is the product.

How to read a matchup score

Every player on every board carries a 0–100 matchup score, and the most common mistake is reading it as a rating of the player. It isn't. It rates the spot in front of him, relative to the rest of that day's slate — so a star facing an ace in a cold pitcher's park can score below an average hitter facing a struggling arm in a bandbox. That's the model working, not the model broken.

Next to it sits confidence, which is coverage rather than certainty. It says how much real data backs the score: plate appearances in the relevant split, how much of the opposing arm's profile we can see, whether the lineup is confirmed. A rookie with eleven career at-bats against lefties gets low confidence no matter how strong his score looks. Read the two together — a high score with low confidence is a thin read, and we'd rather show you that than hide it.

Neither number is a win probability. Where we publish one of those — on moneyline and run totals, or the UFC fight board — it's labelled as such and shown beside the book's implied odds, so you can see exactly where the model and the market disagree.

What the model is bad at

Most sites in this space don't publish this part, which is exactly why it's worth reading. Efficient markets beat us. The moneyline is the sharpest, most liquid market in baseball, and our game-level model grades close to break-even against it. Game totals are much the same. We publish those numbers as context and as transparency, not dressed up as plays, because the honest read on a well-priced market is that there's nothing there.

Several prop markets look efficient too. We've run full reviews on home runs, total bases and hits+runs+RBIs and found no edge worth acting on — the books price them about as well as we can. Those boards stay up because the projections are genuinely useful for lineup and DFS decisions, and because quietly burying a board the moment it stops producing plays would leave you with a misleadingly rosy picture of what the model does. The UFC model is calibrated but hasn't proven it can out-price the books, so it posts no picks at all and says so on the page.

Where the model has found real room is the softer corners — team totals more than game totals, and the strikeout market, where projected workload and the specific lineup a starter draws matter more than a season-long rate. Those are the boards that surface plays. The rest earn their place by being accurate rather than by being profitable. The full version of that argument, with the season record attached, is in the methodology.

Every call is graded, including the bad ones

A projection is scored against the real box-score result, at the line the market actually offered when we called it. The score for a game is locked at first pitch, so once play starts nothing can be quietly rewritten to look smarter afterwards, and the price is frozen with it — if a line moves our way after we post, we don't take credit for the better number. Once a play surfaces it stays shown and stays graded even if the edge drifts before kickoff, because dropping the ones that stopped looking good is the oldest trick in this business.

The whole archive rolls back date by date. Pick any board, walk it backwards a few weeks, and read what we published that morning against what happened that night. That's the part worth checking — everything else on this page is us describing ourselves. Data sources and their licences are listed on the attributions page; nothing here is betting advice, and no outcome is guaranteed.