As the rumor mills fly, optimism abounds for the upcoming 2010 season. Just how good are the Twins going to be in 2010? What are the biggest holes that need to be filled during the off season? And just as importantly, how can we assess the effect that potential moves could have to improve the ball club? Many of you have probably heard of using "Wins Above Replacement", or "WAR" as a method to evaluate a player's total value to a team, compared to a "replacement" player, one who is freely available at no cost from AAA, in season free agents, etc. For the last couple of years, the website Beyond the Boxscore, and more recently on a Twinkie Town Fan Post by Jon Kammerer, fans have compiled a list of player by player WAR projections to project an estimated number of wins for entire teams, divisions and leagues.
I've done similar projections for the Twins the last few years, with varying levels of success. This year, I'm taking my projections in a different direction. Starting with actual 2009 statistics as a baseline, can I track the Twins' WAR projections as moves are made throughout the off season? This article gives us a start for a move-by-move assessment leading up to opening day. First things first, where do I put the Twins as of today? Bottom line, I project the Twins at 84-85 wins as the team is currently constructed. I've posted the full WAR spreadsheet on Google Docs here. I'm sure many of you will immediately be up in arms at a projection of a 2 win drop-off compared to 2009, a full three wins below Jon's analysis. I'll get into much more detail after the jump, but before we get there, let's take a deeper look at the three 2010 projections I've done to date compared to the 2009 data.
Date | C | 1B | 2B | 3B | SS | LF | CF | RF | DH | SP | RP | Hit | Pitch | WAR | Wins | Notes |
2009 Season | 7.2 | 3.2 | (1.3) | 1.4 | 0.8 | 0.9 | 1.9 | 2.2 | 3.1 | 12.6 | 5.1 | 19.4 | 17.6 | 37.1 | 82.1 | Actual 2009 season data |
01 Nov 2009 | 6.1 | 3.1 | 0.6 | 0.4 | 1.4 | 1.1 | 2.4 | 2.1 | 2.2 | 12.5 | 5.3 | 19.5 | 17.8 | 37.3 | 82.3 | End of season. No free agents |
06 Nov 2009 | 6.1 | 2.9 | 1.2 | 0.6 | 2.9 | 0.3 | 2.7 | 1.7 | 2.3 | 12.5 | 5.3 | 20.7 | 17.8 | 38.5 | 83.5 | Hardy-Gomez, Buscher |
10 Dec 2009 | 6.1 | 2.9 | 1.2 | 0.6 | 2.9 | 0.3 | 2.7 | 1.7 | 2.3 | 13.5 | 5.4 | 20.7 | 18.8 | 39.5 | 84.5 | Pavano, Bonser |
How do I get to my projections for each player? I don't use any single projection system out there, but I don't use my own home-grown system either. Since some projections are more conservative than others (for example, Bill James' is notoriously optimistic), I put a number of systems, including Bill James and CHONE, side by side along with the previous three years data and eyeball wOBA and FIP performance for a given player. I put a lot more thought into how I assign plate appearances and innings pitched for each player on the team. You probably notice that most of my PA/IP projections appear a bit low. This allows me to account for risk of injuries, as well as assess a team's depth in the projections. Projecting every position player and member of the starting rotation to play a full season will over-project by quite a bit.
What have I found so far? First of all, I was very surprised that plugging in the final 2009 season statistics into the WAR projection spreadsheets came out at 82.1 wins, five fewer than the Twins actually won (87), and around four fewer than the Twins Pythagorean win total (86.6). Does this mean the WAR projection system is fundamentally flawed when applying to entire teams? I don't think so. I've written on a number of occasions about how the Twins do the "little things" that don't appear in the box score (or wOBA or FIP), but help a team score or prevent runs. I suspect that this explains most of the five run difference and a little luck explains the rest.
Step by step this off season, how has the team improved? Surprisingly, the moment the season ended, my projections actually jumped slightly from 82.1 to 82.3 wins. How could this be, considering that free agents Carl Pavano, Joe Crede, Orlando Cabrera and Ron Mahay were no longer on the roster? Looking at the table above, other than projected declines at catcher (-1.1) third base (-1.0) and DH (-0.9), other positions offset. Middle infield, in particular (+2.5) projected an improvement, primarily due to last season's second basemen (1.3 runs below replacement) being historically bad, and expecting some improvement from a Casilla-Punto-Tolbert-Tolleson combination.
Looking at the moves so far this off season in two chunks, by my projections the Twins have improved themselves by about two wins trading for Hardy and reinstating Pavano. As expected, the Hardy trade improved SS (+1.5) and 2B (+0.6, Punto being a better 2B option than Casilla). These gains were offset by a large loss in LF(-0.8, due to Delmon Young playing pretty much full time). In the end, just over a one win improvement due to the trade. Pavano accepting arbitration also improved the Twins by around one win by setting the #4 spot in the rotation. I only project 140 innings from Pavano, so while one marginal win may not be "worth" the presumed $6-7M he'll get in arbitration, those innings replace 140 innings from Swarzak, Manship and Perkins, which is an improvement over the above replacement value the others would have provided.
Looking forward, I'll continue to update this spreadsheet as the Twins make moves and roster spots are set leading up to opening day. According to the projections, the Twins biggest holes are in LF (Delmon Young slightly above replacement), 3B (Harris-Tolbert is not very good, and I don't project much from Valencia - although he projects better than Harris or Tolbert), and 2B (Punto). Same as the Twinkie Town consensus. In the meantime, I look forward to your comments. Where am I terribly wrong with the projections? What would you do to improve the ball club, and how much effect would it have?