About

Houston to Williamsburg

Ben Guez grew up in Houston, where he played center field at Lamar High School and made first-team All-District twice. He went east to the College of William & Mary, a school known more for its classrooms than its ballpark, and grew into one of the best hitters in the Colonial Athletic Association.

His junior season in 2008 was the breakout. In 57 games he hit .390 with a .470 on-base percentage and a .662 slugging percentage, with 13 home runs, 51 runs batted in and 15 stolen bases. His 75 runs scored ranked ninth in the country and second in school history. He was named All-CAA, Second-Team ABCA East All-Region and All-ECAC First Team, and that June the Detroit Tigers took him in the 19th round of the draft.

Climbing the ladder

He started at short-season Oneonta, then broke out at West Michigan in 2009: Midwest League Player of the Week in July, the Tigers' Minor League Player of the Month for August, and a .483 slugging percentage that tied for fourth in the league. By 2010, his third pro season, he was in Triple-A with the Toledo Mud Hens.

Toledo was home for five seasons. His best came in 2012, when he hit .289 with a .403 on-base percentage across Toledo and Double-A Erie and was named an MiLB.com Organization All-Star and the Mud Hens' Player of the Year by Scout.com. In 2013 he played a career-high 128 games and hit a career-high 18 home runs. After moving to the Milwaukee Brewers' organization, he got on base at a .410 clip for Triple-A Colorado Springs in 2015.

Winters were for more baseball. The first took him to the Dominican Republic, where he joined the Leones del Escogido for the 2010–11 season. He then spent three winters in the Mexican Pacific League, two with Culiacán and one with Mexicali, where he reached base at a .385 clip with seven home runs in 2013–14, and one in the Venezuelan league with Aragua, hitting .290 with a .484 slugging percentage.

Team Israel

In September 2012 he played the outfield for Team Israel in the qualifier for the 2013 World Baseball Classic, at Roger Dean Stadium in Jupiter, Florida. Brad Ausmus managed, in a dugout that also included former big leaguers Gabe Kapler and Shawn Green. The roster featured a future All-Star in Joc Pederson, along with Nate Freiman, Josh Satin and Adam Greenberg. Israel beat South Africa and Spain, then lost the winner-take-all final to Spain, 9–7 in ten innings, one game short of the main tournament.

What the numbers said, even then

His career line tells a story analysts would recognize. Across nine minor league seasons he hit .258, but his on-base percentage was .346, 88 points higher, built on 304 walks. He played all three outfield spots and hit 84 home runs with 94 stolen bases in the minors. Counting winter ball and the Arizona Fall League, he finished his pro career with an even 100 home runs. Much of his value lived in the parts of the game a batting average misses.

Seven seasons in Triple-A also put him close to the decisions this site is about: who gets called up and who gets sent down, which veteran gets the 40-man spot, how a club fills the hole when a starter goes down. In Triple-A, those calls aren't abstract. They decide careers, and they quietly decide pennant races.

Why Roster Memo

Ben loves this game, and he loves the analytics that have reshaped it. Roster Memo is where the two meet. It takes the most rigorous tools available outside a front office (projections, simulation, contract data and a public record of forecasts) and pairs them with something models can't supply: what a roster decision looks like from the clubhouse, the bus and the batter's box.

The numbers do the heavy lifting where they're strong. Where they go quiet, often with the players at the back of the 40-man, that experience has to take over. Every memo here is a recommendation to a real club, and every forecast is graded in public. The goal is simple: find value in rosters that others miss, and be honest about the misses along the way.

How to read a memo

Each memo is addressed to a real club and opens with the recommendation. The evidence fits on about a page. Methods, charts and code links go in the appendix. Every numeric forecast in a memo is logged with a date on the prediction log and graded when it resolves.