[情報] Sean Smith談WAR

看板MLB作者時間13年前 (2012/11/14 22:48), 編輯推噓15(15推 0噓 21→)
留言36則, 26人參與, 最新討論串1/1
http://tinyurl.com/bqy9de3 "The thing about it I wouldn't question is the format," Smith said. "You've got your player's batting, his fielding, his baserunning. How well he stays out of double plays, turns double plays. How he prevents runners from advancing. You can express all of those things in runs. And once you do that, you add them up, and it seems so common sense, I don't understand why people would question it." Sean Smith (BR版WAR開發人): 我從不懷疑公式 你計算球員的打擊 防守 跑壘 避免雙殺 製造雙殺 防止搶壘 把這些換算成分數 然後相加 這不是很合乎常識嗎? 我不懂為什麼人們要質疑它 That said, they don't invalidate the metric's more accurate pieces. As Smith said: "I think you can use WAR without crunching a single defensive number. You can take scouting reports and divide guys into above, average, below. Fudge the numbers. Use scouting-based defensive numbers. If they all match somewhat, use it. If they don't match, you've got to make a judgment call as to what's best. "不用數據也可以計算WAR 例如使用球探報告 把球員分成上中下 推測個數字 然後帶入 如果不合 就再做另一次猜測" "I just couldn't justify ignoring defense entirely." "我只是無法忽略防守不計" --- PS: 數據派前輩Nate Silver預測2008總統大選猜50中49 預測2012大選猜50中50 -- ※ 發信站: 批踢踢實業坊(ptt.cc) ◆ From: 140.112.25.133
推
11/14 22:51, , 1F
戰!
11/14 22:51, 1F
→
11/14 22:55, , 2F
連公式都沒有統一的東西是要怎麼不被質疑
11/14 22:55, 2F
→
11/14 22:56, , 3F
所以Nate Silver跟這篇文章的關係..?
11/14 22:56, 3F
推
11/14 22:56, , 4F
結論:Trout拿MVP
11/14 22:56, 4F
推
11/14 22:58, , 5F
誰知道他在哪隊工作？
11/14 22:58, 5F
→
11/14 22:58, , 6F
舉Silver是想說真正的數據頭強者便是兇狠強勁霸吧 XD
11/14 22:58, 6F
→
11/14 23:06, , 7F
計算化學的公式也沒統一阿 可以餵給電腦不同公式得到
11/14 23:06, 7F
→
11/14 23:06, , 8F
不同答案
11/14 23:06, 8F
→
11/14 23:08, , 9F
他只是想說WAR有其價值，只是每個人標準不一樣
11/14 23:08, 9F
推
11/14 23:09, , 10F
有看文章都知道為什麼Nate Silver會出現
11/14 23:09, 10F
推
11/14 23:09, , 11F
別人家算出來的有時候排名還會不太一樣...
11/14 23:09, 11F
推
11/14 23:14, , 12F
BA, BP, ESPN, Sickels 給的新秀排名有一樣過嗎？
11/14 23:14, 12F
→
11/14 23:15, , 13F
無法準確評價新秀就不用做球探報告了嗎？
11/14 23:15, 13F
推
11/14 23:24, , 14F
等等會不會有人要說棒球不是加減乘除就可以知道的了(誤
11/14 23:24, 14F
→
11/14 23:31, , 15F
兩家用不同的公式來表達,發展,完成同一個概念
11/14 23:31, 15F
→
11/14 23:32, , 16F
不代表這個概念沒價值阿
11/14 23:32, 16F
推
11/14 23:32, , 17F
其實美國總統大選，難預測的就只是那幾個「搖擺州」(5-6)，
11/14 23:32, 17F
→
11/14 23:33, , 18F
50州裡面40-45都是大勢底定，沒啥好猜測的。你看他們兩黨
11/14 23:33, 18F
→
11/14 23:34, , 19F
候選人只跑「搖擺州」造勢就知道了。
11/14 23:34, 19F
→
11/14 23:36, , 20F
所以不用Nate Silver，任何人預測準確率鐵定都40/50以上，
11/14 23:36, 20F
→
11/14 23:36, , 21F
不過這次的搖擺州情勢真的是超級接近，能全隊確實不簡單。
11/14 23:36, 21F
→
11/14 23:54, , 22F
因為公式會讓你聽不到純粹的聲音（誤
11/14 23:54, 22F
推
11/14 23:57, , 23F
公式統不統一不是重點,重點是這個數字的概念
11/14 23:57, 23F
→
11/14 23:57, , 24F
最起碼他是把跑壘,防守考慮進去了
11/14 23:57, 24F
推
11/15 00:09, , 25F
pure sound只有頂級球探聽的到阿XD
11/15 00:09, 25F
→
11/15 00:11, , 26F
數據是一種解釋方式，看你信不信阿，還是信我感覺比
11/15 00:11, 26F
→
11/15 00:11, , 27F
較準？XD
11/15 00:11, 27F
推
11/15 00:24, , 28F
就WAR作者自我的感覺跟我們觀眾自己自我的感覺，看相信誰
11/15 00:24, 28F
→
11/15 00:45, , 29F
棒球不只是算算數!! 球迷的心是熱的!!!! 靠感覺!!!
11/15 00:45, 29F
→
11/15 00:48, , 30F
我覺得跑預測跑出來50州都命中其實還好 重點就是那幾州而已
11/15 00:48, 30F
推
11/15 00:49, , 31F
問題就那幾州(舉個例8州搖擺，8中8，8中7其實都很準了
11/15 00:49, 31F
推
11/15 01:15, , 32F
還是很強吧
11/15 01:15, 32F
→
11/15 02:15, , 33F
不太認同2樓說的話
11/15 02:15, 33F
推
11/15 15:03, , 34F
“純粹的聲音”是人生決勝球的哏嗎？XD
11/15 15:03, 34F
→
11/15 15:10, , 35F
二樓絕對不適合做計算材料化學研究
11/15 15:10, 35F
推
11/17 00:25, , 36F
考慮進去是一回事 問題在計算方是合不合理是一回事
11/17 00:25, 36F
‣ 更多 abc12812 的文章
文章代碼(AID): #1GewynAR (MLB)