Sportsbooks shade strikeout totals for Luis Castillo based on his last three starts. You’re looking at his full-season K% against right-handed batters in pitcher-friendly parks. That gap is money. The question is whether you know how to find it consistently.

MLB player props are mispriced more often than almost any other bet type, and the reason is structural. Books set lines fast, using recent performance and public betting patterns. They’re not running a custom Statcast model for every mid-rotation starter on a Tuesday afternoon. That’s your opening.

When a sportsbook sets a strikeout prop for, say, Logan Gilbert at 5.5, they’re leaning heavily on his last four or five starts, his season K/9, and where the public money is likely to land. That’s it. They’re not cross-referencing his xFIP against left-handed batters in domed stadiums, or checking whether today’s opposing lineup ranks in the bottom third of the league in contact rate against four-seamers above the zone.

You can check all of that. Most bettors don’t.

The numbers that actually matter vary by prop type, and mixing them up is where most losing bettors go wrong. For strikeout props, the relevant inputs are K% (not K/9, which inflates for pitchers who throw more innings), xFIP, and swinging-strike rate. These predict future strikeout performance better than raw totals. For total bases and hits props, you want barrel rate and hard-hit percentage on the hitter side, and BABIP regression on the pitcher side. A hitter with a .220 BABIP over 200 at-bats is likely running unlucky. A prop priced as if that .220 is his true talent level is a buy.

Here’s the thing: books see most of the same season-level stats you do. Leaning entirely on K% and barrel rate won’t give you a long-term edge because those numbers are available to everyone. The edge lives in the variables books are slowest to price.

Platoon splits are one. A book sets a hits prop for a left-handed hitter facing a right-handed starter. Standard. But if that pitcher has a .310 opponent batting average against lefties over his last 18 starts, and the book is using his overall ERA to set the line, you’re getting a number that doesn’t reflect reality. Baseball Reference and FanGraphs both publish L/R splits going back years. Checking them takes four minutes.

Ballpark factors are another. Globe Life Field in Arlington suppresses strikeouts relative to league average [VERIFY: current 2024 park factor ranking]. Coors Field inflates hits. Oracle Park in San Francisco kills power but not contact. A total bases prop priced the same whether a game is in Denver or San Francisco is a prop the book hasn’t fully adjusted.

Same-day lineup context is the fastest-moving variable of all. When the Dodgers rest Freddie Freeman on a Wednesday getaway day, any prop touching that lineup needs to be re-evaluated. Books update quickly, but not instantly. The window between lineup release (usually 3-4 hours before first pitch) and sharp money moving the line can be 20 to 30 minutes of stale numbers.

You’ve done the work. You think a pitcher’s strikeout prop is set a full strikeout too low. Now you need the best number available.

DraftKings might have the over at -115. FanDuel has it at -108. BetMGM has it at +100. Those three lines, on the same prop, on the same day, happen constantly. The difference between -115 and +100 on a bet you’re making 200 times a year is not a rounding error. It’s the difference between a losing record and a profitable one.

Bet365 and Caesars both tend to move lines later than DraftKings, which means they sometimes hold better numbers longer into the day. That’s not a rule. It’s a tendency worth tracking for your specific markets.

If you’re using one book, you’re accepting whatever price that book wants to charge. You’re removing the most controllable variable in the entire process.

Where you get a line matters. When you get it matters just as much.

Pitching props open 24 to 48 hours before first pitch on most major books. The early line is usually the softest because it’s set before sharp bettors have fully modeled the matchup. Overnight and early-morning lines on weekday games, in particular, tend to have wider margins for error. A sharp bet at 10am on a 7pm first pitch can move a strikeout total by half a unit before the public even knows the prop exists.

The other critical window is immediately after lineup release. If a team posts a lineup with three regulars resting, and a pitcher’s outs-recorded prop hasn’t moved yet, that’s a live opportunity. Books have humans monitoring lineup releases, but they’re monitoring every game simultaneously. There are gaps.

Injury news follows the same logic. When a team’s cleanup hitter is scratched 90 minutes before first pitch, total bases props for the opposing pitcher may not reflect that for several minutes. Fast-moving information creates fast-moving edges.

The process breaks down if you apply it to every prop on every game. There are 15 games on a given night in mid-July. You are not finding genuine edges in all of them. Trying to do so means diluting your sharpest spots with mediocre ones, and it means betting tired.

A simple game filter cuts the slate to games where props are most likely to be mispriced. Start with pitching matchup quality: games with a clear mismatch between a sharp pitcher and a contact-heavy lineup create more exploitable strikeout props than a pitcher-versus-pitcher wash. Check the over/under total. High-total games (9.5 and above) often have inflated run environments that affect hits and total bases props in ways books don’t fully price on both sides of the matchup. Then check pace of play and recent lineup stability. Teams that have shuffled their lineup three times in five days due to injury are producing unreliable baseline data.

Two or three well-researched props beat eight guesses every time.

No single input creates a repeatable edge. K% alone doesn’t do it. Line shopping alone doesn’t do it. Timing alone doesn’t do it.

The bettors who grind consistent profit from MLB player props are running a stack: the right statistical model for the specific prop type, the right platoon and park adjustments layered on top, the right book offering the best number, and the right timing to get that number before the market corrects. Remove any one piece and the math gets harder. Keep all four working together and you’re regularly finding yourself on the right side of lines the book should have moved an hour ago.