Are the Core Players Delivering?

Projection System

Every projection starts with a Marcel-style baseline. Marcel, named after Tom Tango's simple projection system, uses each player's last three seasons of FanGraphs data, weighted 5/4/3 in favor of the most recent year. Seasons are also weighted by playing time — plate appearances for batters, innings pitched for pitchers — so a 600 PA season counts far more than a 77 PA cameo.

This weighted average is then regressed toward the league mean based on sample size. A player with 1,500 career PA over three seasons keeps nearly all of his projection; a player with limited history gets pulled closer to average. The amount of regression varies by stat — strikeout and walk rates stabilize faster (around 400 PA) than batting average or BABIP (around 500 PA).

Next, an age curve adjusts the projection based on where the player falls in his career arc. Players aged 25-27 are treated as being at their peak. Younger players get a boost — a 21-year-old like Jackson Holliday receives roughly a 6-point wRC+ bump for expected development. Older players see a modest decline, steepening after age 33.

For batters, the Marcel baseline is enhanced with Statcast data from Baseball Savant. Exit velocity, barrel rate, and hard-hit rate are compared against league averages (88 mph, 7.5%, and 38% respectively). Differences from baseline adjust the projection's wOBA, which in turn shifts SLG, OPS, and wRC+. Barrel rate also directly adjusts the HR projection — roughly 52% of barrels become home runs, so a player who barrels the ball more often than average gets additional projected homers.

For starting pitchers, projected innings come from a weighted historical workload estimate rather than a flat number. The system looks at the pitcher's last three years of innings, applies the same 5/4/3 weighting, and regresses toward a full starter workload of 170 IP. If a pitcher has a healthy season (150+ IP) in the window, the projection won't drop below 75% of that workload — this prevents two injury-shortened seasons from projecting an ace for 70 innings when he's expected to be healthy. Injured players also get pro-rated based on their expected return date.

Setting Benchmarks

Each player has exactly three benchmarks, hand-picked to answer the most important questions about their 2026 season. We chose benchmarks based on each player's role and development path — Gunnar Henderson's benchmarks test whether he can return to MVP-caliber power production, while Colton Cowser's focus on the contact-versus-power balance that will define his career.

Targets are fixed, round-number thresholds calibrated against both our projections and FanGraphs Depth Charts (which blends ZiPS and Steamer). They're designed to be meaningful stretch goals — not so easy that they're met by default, and not so hard that they're unreachable. For pitchers, innings targets account for injury history: Bradish and Rogers are both set at 130 IP given their recent workload limitations, while Shane Baz is set at 140 IP after a full healthy season in 2025.

During the Season

Once games start, benchmarks update daily using current stats from FanGraphs. Counting stats like home runs and innings are shown as 162-game pace numbers so they're comparable to full-season targets throughout the year.

The "ROS Proj" column shows a rest-of-season projection that blends actual performance with the preseason projection. Early in the season, the blend leans heavily on the preseason number; as plate appearances and innings accumulate, actual performance takes over. This prevents a hot April from wildly inflating projections before the sample is meaningful.

Data Sources

Historical stats and current in-season data come from FanGraphs. Statcast metrics (exit velocity, barrel rate, hard-hit rate) come from Baseball Savant. All data is fetched and processed automatically by an AWS pipeline that runs daily.