FanPost

春天在什么培训“物”为响尾蛇?

亚利桑那共和报 - 今日美国网络

After a set of 7 losses in a row (as of the writing of this article), Diamondbacks fans are likely wondering this very point as a strong correlation between Spring Training performance and that of the regular season could mean the D-backs are in for a long and cold 2020 season.But!As realists point out, no team’s record in Spring Training counts for anything so sit back, enjoy the nice spring weather Arizona, and watch non-roster invitees miss routine catches in late innings.After all, the majority of innings in Spring Training will be played by prospects, has-beens, and the never-will-be.But could the cumulative tangibles and intangibles of a team be an indicator of regular season success nevertheless – hitting strategies, fielding strategies, coaching abilities, as well as overall offensive?Let’s take a closer look…

Why does Spring Training exist in the first place?I’d venture three key historical reasons – getting pitchers stretched out for the regular season, honing position players at the plate and in the field, and ensuring team sobriety at the end of the offseason (looking at you, ’86 Mets).In the Age of Sabermetrics, the first two reasons have diminished as pitchers and position players are utilizing locations like Driveline or the Velocity School in the offseason to become better, stronger, faster, and准备for the regular season.Nowadays, my belief is that Spring Training is a way for teams to judge just how well the players have prepared themselves for the regular season and to gauge performance metrics (via an Edgertronic camera, for instance, for pitchers) relative to last season or the last measurement period.

To that effect, I began looking at Spring Training and regular season data to come to a conclusion on whether Spring Training really does correlate well to regular season performance.

First, let’s take a look at W-L records compared between Spring Training and the regular season across baseball from 2017 to 2019 (the Age of Sabermetrics):

w ^-L记录比较

春训常规赛W¯¯-L记录

  • 2017 R2 = 0.073
  • 2018 R2 = 0.0936
  • 2019 R2 = 0.0555

When completing a simple linear regression between the data, what we find is a比较一致R2 (coefficient of determination) value between the last three seasons.For the statistically uninitiated, what does this mean?The value isn’t ‘0’ (no relationship) but it isn’t anywhere close to ‘1’ (direct/sole relationship) – an R2 value in this neighborhood means there is a很轻微的相关性between Spring Training results and the results of the regular season.

Some teams out-perform their Spring Training stats in the regular season – for instance, the Braves have regularly outperformed their Spring Training stats by an average W-L delta of 0.130 (the equivalent of 21 games in the regular season) while the Royals have underperformed their Spring Training stats by an average W-L delta of -0.156 (25 game equivalent).The Diamondbacks?– their average W-L delta is 0.068 (11 game equivalent).

球队

2017 STW¯¯-L

2018 STW¯¯-L

2019 STW¯¯-L

2017和RSW¯¯-L

2018和RSW¯¯-L

2019和RSW¯¯-L

2017年无功。

2018瓦尔

2019瓦尔

Avge瓦尔

Dbacks

0.5

0.5

0.4

0.574

0.506

0.525

0.074

0.006

0.125

0.068

皇家

0.531

0.552

0.6

0.494

0.358

0.364

-0.037

-0.194

-0.236

-0.156

勇士

0.29

0.419

0.5

0.444

0.556

0.599

0.154

0.137

0.099

0.130

So what does this all mean?Well, looking at the data team-by-team and season-by-season shows a tremendous amount of variance between Spring Training and the regular season so predicting regular season results from Spring Training W-L records is not recommended.然而, there does look to be some consistency in the maximum variance teams can expect from Spring Training into the regular season.

The most any team has outperformed their Spring Training stats over the last three years is the Dodgers in 2019 – Spring Training W-L was 0.483 while regular season W-L was 0.654;the most a team has underperformed their stats is the 2018 Orioles – Spring Training W-L was 0.586 while regular season W-L was 0.290 (as they intentionally tanked in the year).But the standard deviation of the variance itself is less than 0.1 so there is a nominal floor and a nominal ceiling of where a team is expected to end up (above or below) their Spring Training record – if the Dbacks walk out of Spring Training with a record of 0.290, the statistical data strongly suggests they aren’t ending the regular season at 0.654.

这是否意味着我们都应该恐慌?并不是所有的 - 还是有留下春训,其中首先将得到比他们现在更多的出场时间了几个星期。请记住,预测从春训记录中的常规赛战绩是统计不健全的事。But if the Dbacks continue to keep up their losing streak and their Spring Training ends with a catastrophically bad W-L record, the last three years of statistical data across all MLB strongly suggests that their regular season record is not one which will take them to the post-season.

(敬请期待 - 要对工作在ST和常规赛的玩家级别的其他统计模型!)

球队

2017 STW¯¯-L

2018 STW¯¯-L

2019 STW¯¯-L

2017和RSW¯¯-L

2018和RSW¯¯-L

2019和RSW¯¯-L

亚利桑那

0.5

0.5

0.4

0.574

0.506

0.525

亚特兰大

0.29

0.419

0.5

0.444

0.556

0.599

巴尔的摩

0.533

0.586

0.414

0.463

0.29

0.333

波士顿

0.563

0.71

0.414

0.574

0.667

0.519

小熊驰

0.419

0.576

0.594

0.568

0.583

0.519

智白袜

0.485

0.571

0.414

0.414

0.383

0.447

辛辛那提

0.457

0.345

0.296

0.42

0.414

0.463

克利夫兰

0.515

0.594

0.548

0.63

0.562

0.574

科罗拉多州

0.516

0.414

0.483

0.537

0.558

0.438

底特律

0.4

0.464

0.5

0.395

0.395

0.292

休斯顿

0.5

0.7

0.6

0.623

0.636

0.66

堪萨斯城

0.531

0.552

0.6

0.494

0.358

0.364

洛杉矶天使

0.6

0.394

0.484

0.494

0.494

0.444

洛杉矶道奇队

0.514

0.531

0.483

0.642

0.564

0.654

迈阿密

0.433

0.536

0.536

0.475

0.391

0.352

密尔沃基

0.5

0.613

0.576

0.531

0.589

0.549

明尼苏达

0.552

0.5

0.519

0.525

0.481

0.623

纽约大都会队

0.469

0.357

0.448

0.432

0.475

0.531

纽约洋基队

0.727

0.581

0.63

0.562

0.617

0.636

奥克兰

0.471

0.467

0.636

0.463

0.599

0.599

费城

0.452

0.433

0.5

0.407

0.494

0.5

匹兹堡

0.613

0.367

0.5

0.463

0.509

0.426

圣地亚哥

0.344

0.6

0.621

0.438

0.407

0.432

旧金山

0.543

0.484

0.448

0.395

0.451

0.475

西雅图

0.576

0.533

0.476

0.481

0.549

0.42

圣路易

0.714

0.567

0.444

0.512

0.543

0.562

坦帕湾

0.429

0.467

0.433

0.494

0.556

0.593

德州

0.515

0.267

0.433

0.481

0.414

0.481

多伦多

0.4

0.438

0.483

0.469

0.451

0.414

华盛顿

0.433

0.433

0.586

0.599

0.506

0.574