[Editor's note: By request, I'm featuring this at the top for easy access until I do a companion Yankees piece. I'll try to tackle that this weekend, though I warn you it will be painful, and I probably will return from the experience a bitter, shaken man. Oh, and from now on, please refer to me as "this dude." Thanks.]
Perhaps the only thing more fun for baseball fans (and by “baseball fans,” I mean “dorks”) than actually seeing baseball start in April is predicting how baseball will go after it starts in April. Here’s a fairly comprehensive look at the Red Sox, how they did in 2006, and how that compares to what the well-respected Baseball Prospectus/PECOTA and Bill James/Baseball Info Solutions projections say they’ll do in 2007.
First, the Red Sox' starting nine in 2006 were pretty disappointing, compared to the monster offenses put out from 2003-05:
Youkilis -- .279/.381/.429 13 HR, 159 H, 100 R
Loretta -- .285/.345/.361 5 HR, 181 H, 75 R
Ortiz -- .287/.413/.636 54 HR, 137 RBI, 29 2B, 119 BB
Ramirez -- .321/.439/.619 35 HR, 102 RBI, 27 2B, 100 BB
Varitek -- .238/.325/.400 12 HR, 55 RBI
Nixon -- .268/.373/.394 8 HR, 52 RBI, 24 2B
Lowell -- .284/.339/.475 20 HR, 80 RBI, 47 2B
Crisp -- .264/.317/.385 8 HR, 36 RBI, 22 2B, 22 SB
Gonzalez -- .255/.299/.397 9 HR, 50 RBI, 24 2B
Figuring out a typical lineup is pretty difficult, thanks to significant injuries to Varitek, Nixon and Crisp, as well as the lineup switching between Youkilis and Crisp, plus the significant playing time given to Cora and Pena. Anyway, that’s somewhat typical – and not very good. It’s a wonder Ramirez got any pitches to hit, considering the gap between him and Lowell.
Based on Cyril Morning;s and Ken Arneson’s Lineup Analysis tool, as posted on Baseball Musings, that lineup should have scored 5.58 runs per game. The best lineup involving those nine players (Youk-Ramirez-Nixon-Ortiz-Lowell-Gonzo-Tek-Crisp-Loretta) would have scored 5.80. The Sox actually scored 5.06 RPG, a testament to their injuries and the amazingly horrific lineup Terry Francona was forced to field at times. That amounts to 83 lost runs – or roughly eight wins. The underperformance of Varitek, Crisp and Nixon – when they were actually “healthy” – didn’t help either. Was this lineup good enough to win the AL East? No. But clearly the Red Sox would have been a playoff contender well into September without their bad luck.
So what about the 2007 Red Sox lineup?
The Bill James Handbook projects the Sox this way (caveats being that the new players were not projected for Fenway Park, which suppressed home runs and enhances doubles, both to severe extents, last season; adjust accordingly. I also added some counting stats because those are more interesting than percentages):
Lugo -- .277/.343/.399 11 HR, 156 H, 85 R, 26 SB
Youkilis -- .283/.395/.433 14 HR, 165 H, 101 R
Ortiz -- .285/.391/.592 47 HR, 138 RBI, 42 2B, 103 BB
Ramirez -- .305/.414/.590 37 HR, 118 RBI, 33 2B, 90 BB
Drew -- .283/.398/.493 24 HR, 82 RBI, 27 2B
Varitek -- .259/.343/.434 17 HR, 69 RBI
Lowell -- .273/.341/.452 18 HR, 77 RBI, 36 2B
Crisp -- .284/.337/.419 11 HR, 54 RBI, 30 2B, 23 SB
Pedroia -- .284/.355/.418 10 HR, 72 RBI, 47 2B, 67 BB, 43 K
This strikes me as very conservative – slight rebounds from Crisp and Varitek, a slight decline from Lowell, and a solid (47-double) debut by Pedroia. If Drew puts up those numbers over his projected 138 games, I’ll be quite pleases.
As you’d expect, these numbers produce a much better result – 5.93 runs per game, or 961 total runs – 60 more than they should have scored with the 2006 lineup and about 150 more than they did score in 2006. So, based on the BIS projections, the Sox added between six and 15 wins with their lineup changes – adding Lugo and Drew in place of Gonzalez and Nixon, and allowing Pedroia to replace Loretta. The Varitek and Crisp improvements seem to be a wash with the slight regressions from Ortiz and Lowell.
(For what it’s worth, the most productive lineup theoretically would be: Youk-Manny-Pedroia-Ortiz-Drew-Lowell-Tek-Crisp-Lugo, which would produce 6.04 runs per game.)
Anything could happen, of course. Assuming that last year was a once-in-a-generation fluke and the Sox don’t underperform their statistics by 83 runs, a significant offensive boost seems likely. Does PECOTA agree?
Lugo -- .284/.347/.406 8 HR, 134 H, 74 R, 19 SB
Youkilis -- .271/.376/.456 18 HR, 142 H, 90 R
Ortiz -- .289/.406/.577 41 HR, 130 RBI, 37 2B, 109 BB
Ramirez -- .297/.400/.567 33 HR, 105 RBI, 32 2B, 84 BB
Drew -- .285/.392/.476 15 HR, 61 RBI, 27 2B
Varitek -- .274/.357/.453 14 HR, 55 RBI
Lowell -- .273/.333/.441 15 HR, 74 RBI, 33 2B
Crisp -- .310/.361/.452 13 HR, 63 RBI, 27 2B, 21 SB
Pedroia -- .294/.360/.431 9 HR, 60 RBI, 36 2B, 47 BB, 39 K
Striking similarities: Both give Mike Lowell and essentially the same projection. The same with David Ortiz and, to a lesser extent, Julio Lugo and Kevin Youkilis. Both like Pedroia a lot – BIS likes his power; PECOTA likes his patience. Both project him to be a high-doubles, low-K, high-contact player, and both say he’ll be better than Loretta.
Interesting differences: PECOTA sees a significant dropoff from Ramirez, as well as a disappointing season for Drew. That’s been discussed here before [LINK]. It sees Crisp rebounding to a major extent, and Varitek coming back nicely. BIS seems more willing to project bigger home run numbers for the middle of the lineup. PECOTA seems to be giving a lot of weight to Fenway’s home run suppression.
The PECOTA projections would give the Sox an even better lineup than the BIS projections – 5.95 runs per game (as opposed to 5.93), or 964 runs – a whopping thee runs better. So, despite the differences in some of the players, both systems essentially have the Sox scoring the same next season and, by extension, improving their offensive output significantly.
That’s only half of the equation, of course. Pitching counts, too.
Let’s use the same system for pitchers. Those who made at least 10 starts for the Sox in 2006:
Schilling -- .276/.303/.459 15-7, 3.97 ERA, 8.07 K/BB
Beckett -- .245/.317/.450 16-11, 5.01 ERA, 2.14 K/BB
Wakefield -- .248/.322/.413 7-11, 4.63 ERA, 5.79 K/BB
Lester -- .294/.381/.434 7-2, 4.76 ERA, 1.40 K/BB
Clement -- .291/.392/.438 5-5, 6.61 ERA, 1.13 K/BB
Snyder -- .314/.366/.518 4-5, 6.56 ERA, 2.85 K/BB
Ugly. The top four arms in the bullpen:
Papelbon -- .167/.211/.254 4-2, 0.92 ERA, 5.77 K/BB
Foulke -- .271/.298/.490 3-1, 4.35 ERA, 5.14 K/BB
Timlin -- .305/.349/.449 6-6, 4.36 ERA, 1.88 K/BB
Tavarez -- .293/.374/.451 5-4, 4.47 ERA, 1.27 K/BB
The Red Sox gave up 773 ER in 1,441 innings last year for an ERA of 4.83.
How do the same positions stack up under our two projection systems?
BIS (with a qualifier that James doesn't belive in projecting pitching performances and a qualifier from me that it picked only two pitchers to win as many as 18 games next season -- Santana, 20, and Carpenter, 19):
Schilling -- 12-8, 3.50 ERA, 5.9 K/BB
Beckett -- 13-10, 3.68 ERA, 2.55 K/BB
Matsuzaka* -- 15-10, 3.55 ERA
Wakefield -- 8-8, 4.14 ERA, 1.96 K/BB
Papelbon -- 14-6, 2.98 ERA, 3.77 K/BB
Lester -- 4-5, 4.38 ERA, 1.80 K/BB
* Matsuzaka was not projcted by BIS, so I used an average of various projections compiled here.
A much better rotation, for sure, with all five starters outperforming their previous output or the previous output of their slot, and four of the five (including Schilling himself) posting ERAs lower than Schilling's 2006 staff best. BIS is extremely conservative with its victory totals. If Papelbon does post an amazing sub-3 ERA, which is improbable, is there any way he wins only 14 games? Highly doubtful.
The BIS bullpen:
Timlin -- 72 G, 3.86 ERA, 2.56 K/BB
Donnelly -- 62 G, 3.41 ERA, 2.67 K/BB
Tavarez -- 62 G, 4.56 ERA, 1.39 K/BB
Delcarmen -- 53 G, 3.88 ERA, 2.19 K/BB
Romero -- 62 G, 4.40 ERA, 1.5 K/BB
Because it goes to press so soon after the end of the season, the Bill James Handnook failed to catch any of the significant bullpen changes, except Papelbon's move to starter. As a result, the Sox have no projected closer. It does project Joel Pineiro as a reliever/spot starter, but if he starts 14 games for the Red Sox as is projected, 2007 will be a horrible season indeed. Based on these numbers, I'd give the closer's job to Brendan Donnelly.
Projecting the runs and innings pitched from these 11 pitchers to last year's 1,441 IP gives us 587 runs allowed -- just a ridiculous number fewer than last year. It amounts to roughly a 3.70 ERA, which would have blown away Detroit's MLB-best 4.17 ERA this year. While I think the Red Sox have created one of the best -- if not the best -- starting staff in the AL, that would be a best-case scenario, obviously unlikely, considering the combination of the 19 pitching wins with the minimum six offense wins combines to 25 additional wins (111-51? I could live with that).
Perhaps PECOTA will be more instructive:
Schilling -- 13-9, 4.02 ERA, 4.24 K/BB
Beckett -- 11-10, 4.47 ERA, 2.27 K/BB
Matsuzaka -- 12-9, 4.01 ERA, 3.18 K/BB
Wakefield -- 9-9, 4.77 ERA, 1.69 K/BB
Papelbon -- 9-7, 4.28 ERA, 2.82 K/BB
Lester -- 6-7, 5.20 ERA, 1.64 K/BB
Talk about your two extremes. In PECOTA's scenario, every pitcher, including Matsuzaka, underperforms Schilling's team-best 2006 season. PECOTA and BIS share one similarity -- they both trend very conservatively when predicting wins and strikeouts/walks. I consider Schilling and Matsuzaka putting up K/BB ratios that low just as improbable as Papelbon posting a sub-3 ERA. I also view a 4.77 ERA from Wakefield as unlikely, though the rest seems realistic if a bit conservative.
Timlin -- 36 G, 4.24 ERA, 2.30 K/BB
Donnelly -- 46 G, 4.36 ERA, 2.15 K/BB
Tavarez -- 43 G, 5.05 ERA, 1.22 K/BB
Delcarmen -- 47 G, 4.29 ERA, 2.23 K/BB
Romero -- 51 G, 4.65 ERA, 1.44 K/BB
This is very ugly indeed. PECOTA also conservatively projects games played, Romero leading the team and Timlin apparently suffering an injury. None of these sets of numbers is impressive, and PECOTA doesn't project saves at all. It projects Pineiro also as a part-time starter. Lord, I hope not.
Under these projections, and adjusting for IP, the Red Sox would allow 710 ER, or a 4.43 ERA -- still an improvement of 73 runs, or seven wins. Based on PECOTA's projections, the Sox have improved a total of at least 13 wins over last season, projecting to a 99-63 record -- ignoring for a moment that the Sox have, um, absolutely no one in the closer's role at the moment.
Using James' Pythagorean theorum and PECOTA's seemingly more realistic projections, 964 runs scored and 710 runs allowed would project roughly (I used the power of two instead of the power of 1.83) to a 105-57 record. Um, wow.
This is all in fun, of course. They play the games for a reason. My math may be off, so those of you more numbers-inclined, feel free to drop me an email and let me know where I'm wrong, and I'll dutifully correct it. Due credit to Baseball-Reference and ESPN.com for stats, as well as Baseball Musings for the lineup calculator and the idea, taken from this post.