Pebble Hunting: The No... (08/06)
the archives are now free.
7 comments have been left for this article. BP Comment Quick Links
| |
|
|
August 12, 2014 Pebble HuntingThe Unlikeliest Part of the Rangers' Collapse
Two years and two months ago, the Rangers looked so invincible that it inspired me to write about the ways that seemingly invincible civilizations nevertheless collapse. Of the seven that I attempted to apply, a case could be made that five are relevant to the Rangers:
The last one is the big one, the one that every post-mortem of the 2014 Rangers will focus on, and the justification for some springtime-optimism pieces come February. The Rangers have already lost 1,500 days to injury, 40 percent more than any other team. But while the Rangers have undoubtedly been crushed by long absences of Martin Perez, Derek Holland, Prince Fielder, Jurickson Profar, Geovany Soto, Neftali Feliz—deeeep breath—and others, the effects have been multiplied by the other collapse factors, particularly the inability to develop and/or hold onto prospects. Here are the 20 Rangers prospects who were most highly regarded in, at the time, a very highly regarded system:
Plenty of these guys might still turn into something, either with the Rangers or elsewhere, but for the purposes of 2014 it won’t be much help; the 20 have combined to produce 2.5 WARP for this year’s Rangers, either in performance or in trade returns. Compare that to, say, the Angels (Trout, Richards, Cron, plus trades for Huston Street and Jason Grilli) or the A’s (Donaldson, Norris, Gray, plus trades for Jaso, Gentry, Callaspo, Lowrie).
Now, just to reiterate: The Rangers had a very, very good farm system, even if it ultimately provides (or, to date, has provided) little value. What made the Rangers’ collapse so inconceivable was that they appeared to be the best team in baseball and have one of the best farm systems in baseball. Either is enviable. Which, though, is more enviable?
Of course, being good now is best for now-based reasons. Bird in the hand and all that. But being good this year presumably also has some positive correlation to being good next year, and maybe the year after that, and maybe the year after that. A team that is good has probably been put together by men in suits who are good at putting together teams, and will continue to be good at it. A team that is good collects more revenue, which it can then spend on players (or nice suits). A team that is good might benefit from the increased trade value of its players. A team that is good might have developed a particular style of play that gives it an advantage, or it might even develop better chemistry, that element being, as Paul Depodesta once said, “a three-game winning streak.” For much of our lifetimes, from the early 1990s to the latter half of the 2000s, we saw teams that were good stay good for a long time (Yankees and Braves especially, but Twins, Giants, Angels, Red Sox) and teams that were bad fall into spirals they couldn’t escape (Royals, Pirates, Orioles, and for a long time the Tigers).
On the other hand, being good costs high draft picks, costs prospects in let's-get-better-now deadline trades, and might boost the cost of keeping one's own players, who are by definition doing well. So, to the question: Is it surprising that the Rangers suck so bad because they were so, so good just a couple years ago? Or is it surprising that the Rangers suck so bad because their farm system was so, so good a couple years ago?
Here’s what I did: I took every team since 2006 that won 90 games in a season, then looked to see how likely they were to win 90 games three years later, four years later, and five years later. I did the same for teams that won 72 or fewer games, which is the same-sized population on the sad end of the winning-percentage curve.
Then, I took all the teams that were ranked in the top and bottom five farm systems in baseball since 2006. Again I looked at their Year+3, Year+4 and Year+5 records to see how much more likely each type of team was to win 90 games in the medium-term future.
Here are the teams that start out good at the major-league level:
These results are probably a bit cleaner, narrativewise, than we have any right to expect from such a limited sample. But the progression largely makes sense: Three years after a team is good, there is a fairly substantial residual effect. The same applies for a team that is bad. As more time passes, the two groups merge, and by Year+5 they are distributed almost exactly as teams randomly assigned 90-win seasons would be. (About 25 percent of teams win 90 games in a given year.) It’s somewhat interesting that bad teams seem to recover faster than good teams get worse, though it’s possible that this is just a sample issue. We’re looking at only two extra good teams, out of 41, reaching 90 wins.
How about the likelihood of winning 72 or fewer from each starting line?
The results aren’t as nicely ordered, but the basic point is clear: If you want to be good in three or four years, it helps to be good now. Beyond that, it gets close to random. So we know that the Rangers, just three years after winning 96 games and representing the American League in the World Series, weren’t likely to be this bad; they were likely to have some remnant effects of their success; but they also weren’t immune, with Suck chances about 70 percent as good as Awesome chances. Now, to their farm. We’re going to call them a top-five farm system, for simplicity. We actually ranked them sixth at the time, and now we’re calling them a top-five system, so you might wonder how this is simple. For simplicity, I’m not going to answer that question.1 Here are each type of system’s chances of producing a 90-win winner three, four, and five years down the road:
And the chances of producing a 90-game loser:
Here we see a much bigger difference, and we see why the Rangers’ collapse is especially noteworthy. Practically no teams produce 90 losses three years after having a top-five system,2 and nearly half top 90 wins. Last one: What about teams that, like those Rangers, were both very good and had very good systems? There aren’t enough in the recent past to draw any sort of conclusions at all, but going back further introduces (a) less authoritative organizational rankings and (b) a different era of parity, or non-parity, than the one we’re looking at. Still, what the heck. Here are all +3, +4, and +5 outcomes for the 31 teams that, between 1986 and 1999, or between 2006 and 2012, won 90 games and were named a top-five organization (by Baseball America or, after 2007, by us) in the following winter:
So, again, the Rangers aren’t the first that this has happened to. But you can appreciate just how unlikely it is to see them this bad, this soon. It’s a challenge to go from very good to very bad that quickly, but it’s especially difficult to blow a top-five farm system. Fans of the Cubs, Twins and Astros can take heart; it is, the record shows, more promising to have a good system than a good team.
Sam Miller is an author of Baseball Prospectus. Follow @SamMillerBB
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The Stash List: 17th Edition
A very interesting article. I'd suggest you would strengthen your argument adding a 2nd variable, payroll, into your approach. There is a strong positive correlation between payroll and wins. There is also a strong positive correlation between payroll this year and payroll last year (big market teams remain big market teams, and vice-versa). I suspect that if you added this as a factor, your regression results would become even more stark. If you spend like the Yankees and Dodgers, you are going to be much more likely to win 90 games irrespective of farm system. If you are 2nd or 3rd quartile payroll team like the Rangers, I would expect that the quality of the farm system is significantly more meaningful in predict t+3 or t+5 results.
Glad you brought it up. Originally intended to, but payroll is a bit trickier because payroll often responds to winning; teams that expect to be good spend more, teams that expect to be bad spend less. You're right that it would have been a more complete look at it to figure out the right way to incorporate it here, but I was aiming to keep it simple. But you're right!