Projecting win-loss record for all 30 teams now that Bryce has signed

With the Phillies’ signing of Bryce Harper last week, we can start to turn our full focus on spring training, as the time-honored baseball calendar dictates. Most of what Scott Boras would call the “steak” of the free-agent market has been served. And while there are some quality players still looking for landing spots — a list that includes lefty starters Dallas Keuchel and Gio Gonzalez, longtime relief ace Craig Kimbrel and veteran outfielders Carlos Gonzalez and Adam Jones — we now have a pretty good idea of how each team’s roster will look when the 2019 season begins later this month.

It’s been awhile since we have taken stock of the majors through the prism of my ever-evolving projection system (MLBPET, for those who care), so here are a few technical notes before we swing around the big league landscape.

1. The biggest change since last season is that, as part of my procedure for constructing the recent series on positional tiers, I have separated defensive projections from offensive projections. That is, for position players, I used to simply project a version of WAR based on player track records and career aging curves. In analyzing my hits and misses from the past couple of seasons, it became clear that the defensive component of WAR was problematic from a forecasting standpoint. By separating hitting and fielding, I am hopeful that not only will the accuracy of the team forecasts improve but I’ll be better able to judge the strengths and weaknesses of each team.

2. The projected records you will see are baseline forecasts. They have not been cycled through a schedule simulator, a process that adjusts the records for strength of opponent and generates the probabilities for each team qualifying for the postseason. We’ll save that for the Opening Day version of Stock Watch.

3. The runs scored and runs allowed listed are context-neutral. They have been adjusted for league and ballpark, so you can look at comparisons between teams as an apples-to-apples relationship. However, bear this in mind. When you look at, for example, the Rockies’ projection, bear in mind that their real-life numbers will skew higher because of the high run-scoring environment of Coors Field.

4. The unit ranks are based on each player’s subcomponent of my version of WAR, which isolates hitting from fielding, and starting pitching from relief pitching. Each player’s score in each of these areas is weighted for anticipated playing time. The team score is a simple sum of these measurements.

5. The change from 2018 measures the team’s current win projection as compared to its final power ranking at the end of last season. The measurement reflects personnel changes, expected statistical regression and the aging patterns of the players on the roster. Teams with extreme 2018 records — such as 100-plus wins or 100-plus losses — will tend to have a heavy degree of statistical regression.

Leave a Reply

Your email address will not be published. Required fields are marked *