Is this guy for real? 他是真货还是假货？ 原文连结：http://www.hardballtimes.com/main/article/is-this-guy-for-real/ by Sal Baxamusa July 14, 2008 The All-Star break is upon us, and if they haven't already, teams will be re-evaluating their position relative to the league. Are they buyers? Are they sellers? 全明星假期已经到了，每个球团将会（如果他们还没有开始的话）重新评估球队内各个位 置相对於联盟的强弱。他们应该开始兜售球员吗？还是他们应该藉由交易来补强实力呢？ That decision depends partially on where you are in the standings, but also on what you think your players will do in the coming months. Dumb teams will assume that their players will put up similar numbers in the second half as they did in the first half. But smart teams, of course, won't do that. 这个决定不只跟球队的战绩有一些关系，也和球队认为他们的球员在之後几个月内会表现 的怎样有关。白痴的球队会单纯得以为他们的球员会表现的跟上半季一样。而当然了，聪 明的球队不会这麽做。 Why not? Because what a player has done this year is not his true talent level. He could be significantly better or worse, or he may be playing up to his true talent level. But performance data are just a sampling from a player's true talent, and half a season really isn't enough information to say Barry Zito is cooked or that Xavier Nady has become a star. 为什麽他们不会这麽做呢？因为一个球员到目前为止的表现并不一定是他真正的实力。他 可能有超水准的惊人表现或是超水准的摆烂演出，也可能只是展现出他真正的实力。但是 这些表现出来的东西只不过是他真正实力的随机取样而已，所以这半个球季并不足以证明 Barry Zito 鸟掉了，或者 Xavier Nady 已经成为一个明星球员了。 As usual, a player's true talent level, or what we expect his performance to be going forward, can be determined by creating a projection utilizing all of his performance data—this year's, last year's, and even the year before that. And as I've focused on recently, the easiest way to do that is to run a simple Marcel projection. 跟其他东西一样，球员真正的实力（或者可以称为我们对他未来表现的预测）是可以透过 所有关於他表现的资料（今年的资料，去年的资料，甚至是前年的资料）来进行推算的。 而达成这个目的最简单的方法（也是我最近在研究的方向）就是使用 Marcel 这个预测系 统。 Today, I'm going to give you the tools to do that for yourself—with a major assist from Dave Studeman who helped me test and improve the initial versions. You might use it for your fantasy team or just to see if your favorite club should stick with a struggling player. 今天，我会把所有的工具都给你，然後你就可以自己来预测球员下半季的表现（我要特别 提一下，Dave Studeman 帮我测试以及修改最初的版本）。你可以用这些东西来该善你的 fantasy 球队，也可以用它们来看看你喜欢的球队是不是该继续相信那个成绩跌入谷底的 球员。 You can download a spreadsheet for pitchers here and for hitters here. Using them is simple. Download and open the spreadsheet and find the "Paste" tab. Enter your favorite player's name and his date of birth. Then, use our handy search feature to find his THT player page. Copy the first block of stats (this works best if you use Internet Explorer), all the way back to 2004 (which is as far as we go) and paste them in the appropriate place in the "Paste" tab. The next tab, called "Quick-n-dirty Marcel hitting" (or "Quick-n-dirty Marcel pitching") automatically computes that player's current projection for the rest of the year and his projected 2008 totals. 你可以在 ftp://ftp.hardballtimes.com/QDMarcel_pitching.xls 下载投手的试算表，而 在 ftp://ftp.hardballtimes.com/QDMarcel_hitting.xls 可以下载打者的。使用方式很 简单。下载并打开试算表，你可以看到一个叫做「Paste」的分页。输入你想预测球员的名 字和出生日期。再用 http://www.hardballtimes.com/thtstats/main/player/index.php 找出那个球员的资料。然後把第一个表格内从 2004 年开始的资料（这是这个程式的限制 ）贴到「Paste」分页内相对应的栏位（建议使用 Internet Explorer）。而同时，另一个 叫做「Quick-n-dirty Marcel hitting」的分页（或是「Quick-n-dirty Marcel pitching 」）会自动推算出球员下半球季的表现和 2008 球季的整体成绩。 Marcel has a lot of caveats. Basically, Marcel only knows three things: a player's stats, the stats of the rest of the league, and the player's age. Marcel doesn't know anything about injuries, parks, minor leagues or platoon splits. But by and large, this simple system isn't a whole lot worse than the heavy hitters like CHONE or PECOTA or ZiPS. And there's a reason why I call them "quick-n-dirty." The spreadsheet I made is a pretty simplistic tool. It doesn't forecast RBI or runs—sorry fantasy gurus—and the estimated playing time is based purely on extrapolation. But it does do a pretty good job at most of the other stuff. Marcel 预测系统有很多要注意的地方。基本上，Marcel 只考虑三件事：这个球员表现出 来的成绩，整个联盟的平均成绩，以及球员的年纪。Marcel 完全不考虑伤病，球场，小连 盟成绩，或者是左打右打等因素。不过整体而言，这个预测系统比起那些重量级的 CHONE ，PECOTA，和 ZiPS 却毫不逊色。而我把我的工具叫做「快速而且肮脏」（quick-n-dirty ）是有原因的。这个试算表非常的阳春。它并没有包含打点以及得分等数据（fantasy 的 玩家，对不起了）而且球员上场的时间完全是由上半季的资料直接推算。不过它在其他方 面应该表现的还算不错。 How good is this guy? 他到底有多强？ For example, Chipper Jones is currently hitting .373/.473/.613. That's pretty darn good, but we all know that, good as Chipper is, he can't keep that up. Pretty much nobody can. That's not a reflection on Jones; it's just the truth. His projection for the rest of the year is a still-awesome .316/.409/.553, which would put him at .346/.441/.585 at year's end. Impressive. 举例来说，Chipper Jones 现在的打击三围是 .373/.473/.613。这是非常吓人的成绩，不 过大家应该都知道，就算像 Chipper 这麽强的一个球员，他也没有办法一直维持这样的成 绩。这应该是没有人可以做到的。我可不是想扁低 Jones，这只是事实。预测系统对他下 半球季的推测仍然是十分棒的 .316/.409/.553，而整个球季的平均会是 .346/.441/.585 。我只有一个字可以形容，强。 What about Jose Vidro? There's a blogospheric debate over how bad he is right now. Yes, Vidro has hit terribly this year (.214/.261/.311—yeesh!), but we can't look only at this year's stats. That would be dishonest. His current projection for the rest of the year is .269/.333/.374. Yeah, that's bad, especially for a DH. No, he probably couldn't help any team in the league, unless he could still play second base acceptably. But he's not the stinky pile of poo that M's fans make him out to be. He's more like a, well, let's say an odorless pile of poo. Still offensive, though. 那 Jose Vidro 又如何呢？在 blog 界正在争论着他到底有多烂。没错，Vidro 今年的打 击成绩实在是惨不忍睹（.214/.261/.311），不过我们不能只看今年的成绩。短期的成绩 是会骗人的。预测系统推测他下半季的成绩是 .269/.333/.374。嗯，还是很烂，尤其是他 还是个 DH。他应该是没有办法帮助任何一支球队，除非他有还办法恢复成一个平均水准的 二垒手。但是，他并不是像水手球迷口中所说的一坨恶心的臭大便。该怎麽说呢，他比较 像是一陀没有什麽臭味的大便。这样好像还是很酸。 And pitchers? Bay Area whipping boy Zito is sporting a 4.77 FIP. That's not not what you expect from a $120 million ace. Is he really that bad? Kind of, yeah. His projected FIP for the balance of the year is a slightly-better-but-still-not-good 4.49. 那投手呢？湾区的众矢之的 Zito 有着 4.77 的 FIP。这不应该是一个领 1.2 亿的王牌投 手该有的成绩。他真的那麽烂吗？嗯，是真的。他这个球季的整体预测 FIP 是「有点进步 但是还是很差的」4.49。 Cliff Lee has been awesome this year, posting a 2.31 ERA—and that's a legit mark, given that his FIP is 2.37. But this is coming after years of mediocre-to-poor performance, including a brutal 5.59 FIP last year. It's not fair to brush off this year's hot half, but it's not fair to ignore all of last year either. Marcel sees him posting a 3.64 FIP for the balance of the year. That's remarkable, considering that his Marcel projection entering the year had him tabbed for a below-average 4.37 FIP going into the year. Cliff Lee 今年的表现太棒了，他有着 2.31 的 ERA（而由他 2.37 的 FIP 可以看出这可 不是运气）。但他前一年的表现不过是在平均水准或以下，而且还有着 5.59 的 FIP。这 虽然不足以完全否认他今年惊人的上半季，但是我们也不能直接忽略他去年的成绩。 Marcel 预测他今年整体上会有 3.64 的 FIP。和季初 Marcel 推测他会有低於平均的 4.37 FIP 相比，这可是很棒的表现。 Before I leave you, I'll share with you what Marcel thinks the end-of-the-year leaderboards will look like (on pace for a minimum 500 PA): 在结束之前，我想告诉你 Marcel 所推测今年球季结束的打击排行榜会是怎样（所有列出 的球员都被预测会有超过 500 次 PA）： Player AVG OBP SLG OPS Pujols, Albert .337 .451 .596 1.047 Berkman, Lance .322 .424 .604 1.028 Jones, Chipper .346 .443 .584 1.027 Holliday, Matt T .331 .408 .558 .966 Rodriguez, Alex .303 .394 .567 .962 Bradley, Milton .300 .414 .546 .961 Burrell, Pat .270 .399 .538 .938 Ramirez, Hanley .311 .387 .550 .938 Utley, Chase .293 .376 .558 .934 Uggla, Dan C .274 .359 .553 .913 Those stats are a combination of performance already in the bank plus the Marcel projection for the balance of the year. Here's who Marcel thinks are, at this very moment, the best hitters in the league. 这个成绩包含了上半季的成绩以及 Marcel 对下半季的预测。而下面是 Marcel 认为目前 全联盟最好的打者。 Player AVG OBP SLG OPS Pujols, Albert .324 .427 .586 1.014 Jones, Chipper .316 .409 .554 .964 Ortiz, David .289 .397 .565 .963 Rodriguez, Alex .297 .396 .560 .956 Holliday, Matt T .322 .391 .556 .948 Berkman, Lance .295 .400 .546 .946 Cabrera, Miguel .317 .397 .543 .941 Howard, Ryan J .272 .376 .560 .936 Braun, Ryan J .304 .355 .571 .927 Wright, David A .307 .396 .523 .919 ...and the worst (among guys who'll see significant playing time). ......以及最差的打者（这是从有很多上场机会的球员中选出来的）。 Player AVG OBP SLG OPS Izturis, Cesar .254 .311 .337 .649 Everett, Adam .236 .288 .353 .642 Berroa, Angel .244 .286 .350 .637 Molina, Jose .236 .281 .349 .631 Bako, Paul .224 .296 .328 .625 Vizquel, Omar .239 .304 .319 .624 Cairo, Miguel .238 .295 .329 .624 Ausmus, Brad .229 .307 .312 .619 McDonald, John .237 .286 .329 .615 Pena, Tony F .239 .273 .336 .610 So go ahead, download, and enjoy! 赶快去下载并好好玩玩我们的工具吧。 References and Resources 参考资料以及资源 PS. Xavier Nady is projected at .282/.342/.467 for the rest of the year. 附注：Xavier Nady 下半季的预测表现是 .282/.342/.467。 The spreadsheets are located at the following URLs: 你可以从下面这些位置下载到文中所提得试算表： ftp://ftp.hardballtimes.com/QDMarcel_pitching.xls ftp://ftp.hardballtimes.com/QDMarcel_hitting.xls --

※ 发信站: 批踢踢实业坊(ptt.cc) ◆ From: 98.222.63.117 ※ 编辑: leecher 来自: 98.222.63.117 (07/15 01:49) ※ 编辑: leecher 来自: 98.222.63.117 (07/15 01:54) ※ 编辑: leecher 来自: 98.222.63.117 (07/15 01:57) ※ 编辑: leecher 来自: 98.222.63.117 (07/15 03:24)