So I've been slightly annoyed by the small contingent of people around the Interwebs suggesting that KC's spring training dominance this year has any significance whatsoever, so I decided to look into it. Then, I started seeing some pretty good teams (SF, Oakland, etc.) atop the ST winning percentage list, so I got curious and made a scatter plot of ST win% vs. Regular Season Win%, just to prove to myself that there isn't a correlation. This is probably incredibly stupid because I'm sure someone has done this before, but I'm avoiding doing actual homework so I did it anyway.

 

Now, of course, there are problems with sample size (3 seasons isn't nearly enough) and the fact that ST has so many fewer games played so the winning percentages on one axis are far more varied than on the other. However, since I went through the effort, I decided I'd post the results anyway:

 

 

So the slope of the linear regression is positive...but just barely so. The R-squared value is insanely low, too. Basically, even with the problems above, there is essentially no significance to winning in Spring Training.

 

But just to see if there was anything to be gleaned from the very top ST performers (since it looks relevant when you glance at the standings) , I took JUST the top 5 teams in each Spring Training and plotted them vs. their regular season win%. Now, obviously this introduces even MORE sample size problems, plus a selection bias. But, on the other hand, whatever.

 

 

Even lower R-squared value. Interestingly, though, a bimodal distribution split around .500 regular season win%.

 

So my acting theory, with nothing but really sh*tty evidence to back it up, is that two kinds of teams win in Spring Training: those who suck and thus have stocked up on great draft picks who get a lot of spring play (KC), and those whose talent is just so good that they can't lose even in Spring Training (SF). It remains to be seen how the Kings of the Cactus League, the 2013 Royals, fit into this hypothesis, but I'm thinking they belong to the former group.

Nice work.

FWIW, the identities of the lower distribution of graph 2:

 

2011 Royals

2011 Rockies

2010 Indians

2012 Blue Jays

2012 Mariners

2011 Twins

2010 Cubs

 

So not all perennial losers with highly-rated systems. I may be, and probably am, wrong about my working theory. Back to the drawing board.

Edited March 13, 201313 yr by ScottyDo

FWIW, the identities of the lower distribution of graph 2:

 

2011 Royals

2011 Rockies

2010 Indians

2012 Blue Jays

2012 Mariners

2011 Twins

2010 Cubs

 

So not all perennial losers with highly-rated systems. I may be, and probably am, wrong about my working theory. Back to the drawing board.

 

I think a lot probably has to do with the teams philosophy going into ST. Even if you're a pretty crappy MLB team without a great farm system but you say, have new owners or GM and/or manager on a hot seat you may want to create a buzz around your team in the local or national media. So, and yeah, I'm over simplifying this, a team would just basically a team can play their regulars later into games. Often, especially early on in ST, a team could probably do well playing roughly their regular 25 a majority of the game.

I think a lot probably has to do with the teams philosophy going into ST. Even if you're a pretty crappy MLB team without a great farm system but you say, have new owners or GM and/or manager on a hot seat you may want to create a buzz around your team in the local or national media. So, and yeah, I'm over simplifying this, a team would just basically a team can play their regulars later into games. Often, especially early on in ST, a team could probably do well playing roughly their regular 25 a majority of the game.

Of course, the real answer probably is "the most AB's anyone gets in spring training is ~75, and in 75 at bats the worst player in MLB can look like an all star, especially since some guys don't put their entire arsenal out for spring training".

Very interesting read. I am of the opinion that winning in Spring Training has very little to with winning in the regular season. FWIW, the Royals still have the best record in baseball, at 23-6.

Another part of the problem is that, for the sake of the study, the only constant you have is the theory of a 9 inning game and that the goal is to win. While they'll still play 9 innings (assuming no rain), game 1 of Spring Training is far, far different than game 25 or game 28 simply because early on you're trying to get guys some work in just to basically see how guys look while stretching them out into their roles while later you are trying to get more into game situations as well and, to some extent, trying a little harder to win those games. Really, the best games to look at in Spring Training are probably about the last 7 or so, and you'd have to look at Spring Trainings for the past 40 or 50 years to get anything close to a worthwhile sample.

 

Still, this is awesome.

KC finished 72-90. They added James Shields to pitch every 5 days. And Wade Davis who should contribute as well. So what kind of realistic turnaround are we talking here? 81-81?

KC finished 72-90. They added James Shields to pitch every 5 days. And Wade Davis who should contribute as well. So what kind of realistic turnaround are we talking here? 81-81?

They could do it. They could get another couple wins out of guys like Francoeur not being totally terrible, they could get a couple wins if they had their regular CF and C for most of the season, but the real key for them remains Moustakas and Hosmer. Those guys have the ability to be all stars; if they play like that, they're dangerous, if they muddle through another season, then .500 could be a ceiling.

 

Edit: Found it, here's a good summary. Lots of guys need to perform above their averages, but Hosmer and Moustakas are the guys who need to make jumps.