How the numbers are made

Every number on this site is grounded in real data. We cover one sport per template and turn it on when its season is live — right now that’s MLB. The specific weights and formulas behind our model are proprietary, but here’s an honest look at what goes into it.

Where the data comes from

• MLB stats: the official MLB Stats API — game-by-game batting and pitching lines.
• Market lines & odds: The Odds API — player-prop lines and prices from regulated US sportsbooks.

What goes into a projection

For each player and prop we build a per-game projection from a few ingredients:

• Established level — how the player has produced over the season.
• Recent form — whether they’re heating up or cooling off lately.
• The matchup — the opposing probable starting pitcher and the ballpark; for pitchers, the opposing lineup.

We cover the core MLB markets: hits, total bases, home runs, RBIs, runs, and pitcher strikeouts.

Comparing the projection to the market

A projection only means something next to the price. So instead of ranking by the raw gap between our number and the line (which over-rewards Overs on low lines), the edges board ranks by expected value: we convert each projection into a win probability, compare it to the best available priceacross the regulated US sportsbooks we track (DraftKings, FanDuel, BetMGM, BetRivers, and other licensed books), and show the side and book where the price and our number disagree most. We don’t surface offshore or unregulated operators, and we temper a single-season model toward the market rather than trusting it blindly.

We also only show bets that clear a minimum expected-value bar. When we backtested the model against a full season of real, pre-game prices, the thin edges turned out to be roughly break-even — only the larger disagreements held up. So we leave the marginal ones off the board rather than dress them up as value.

We focus where the model is provably good. We tested a team-level game-totals model the same way and it couldn’t beat the market, so we don’t publish it — we’d rather ship fewer numbers we trust than dress up noise. It’s all analysis, not advice.

What we leave out (and why)

We’d rather surface a short, high-quality list than dump every number on the board at you. So we deliberately filter out:

• Longshot & alternate lines— e.g. a player to hit 3+ home runs at +19900. Tiny “edges” on huge prices are almost always noise or a stale number, not value.
• One-sided or off-market prices — we only rank standard, two-sided markets at mainline prices, where the number is trustworthy.
• Stats our model doesn’t handle well— RBIs and runs, for instance, are lumpy and depend heavily on teammates, so we don’t rank them.
• Offshore / unregulated books — we only surface regulated US sportsbooks.

That’s why our boards lean toward hits, total bases, and strikeouts: those are the markets that are widely offered, fairly priced, and well-modeled. (On daily-fantasy pick’em it’s similar — and some props, like home runs, simply aren’t offered there at all.) It’s intentional, not a gap.

How often it updates

MLB refreshes daily — new box scores, fresh odds, and recomputed projections. Pages revalidate automatically, so what you see is current without us redeploying.

On variance

A projection is the center of a range of outcomes, not a prediction of what will happen. In a single game anything can occur; an edge only shows up over a large sample — see variance and sample size. When we publish hit rates, we’ll include the misses too. Everything here is informational, not betting advice.