The Mariners, The Rangers and the L Word
The M’s finished off their series in Arlington today against the first-place Rangers. The fact that the Rangers are in first is one of those things I find fascinating given their pre-season projections, their injury issues and the months and months that saw Shin-Soo Choo and/or Elvis Andrus hitting like replacement level players. The more you look at the Rangers’ campaign, the more fascinating it gets.* There’s been a lot of talk this year about variance – about a team’s base runs not correlating all that well with their actual W/L record (or run differential), about a team’s performance with runners in scoring position or close games, or any number of reasons that traditional pythagorean records have failed in their assessment of a club. To fans of said clubs, this is either evidence that advanced metrics have, in the charitable version, holes and blind spots, or, in the less-charitable (and less-credible) version, biases against this or that specific team. Because the Mariners are where they are in the standings, and because they seem to be there rather frequently, I thought it’d be a good idea to take a closer look at how and where the Rangers have excelled – to take a forensic look at *which* advanced metrics lead us furthest from their actual record, and how the M’s look in comparison. The point here isn’t to make the Rangers look bad – this is an attempt to figure out what’s going on. I’ll admit I’ve tended to watch the Rangers when they face the M’s, and that view hasn’t done a lot to illuminate the Rangers’ strengths. I’ve watched a limited set of games, so I’ll turn to some less-limited numbers.
The first thing we have to acknowledge is that there are far, far more ways to be “clutch” than are commonly identified. A few years ago, the Orioles won far more games than their overall WAR would’ve implied, and the reason was largely due to a great bullpen performance that produced an insane record in 1-run and extra-inning ball games. At the time, sabermetric fans scoffed, but O’s fans thought that the combination of a great tactical manager in Buck Showalter paired with a preternaturally dependable closer (Jim Johnson, who is just as remarkable in more recent years, but for much worse reasons) meant that this variance from their pure pythagorean record (the O’s won 11 games more than their pythag) wasn’t so much luck as a specific kind of skill. The next year, the St. Louis Cardinals actually *under*performed their pythag, but their incredible run differential seemed out of proportion to their overall batting line. By position player WAR, they were so-so, and though they were above-average in park-adjusted batting, they lagged clubs like the Angels and A’s. Despite all of that, they outscored the Angels and A’s by 50 and 20 runs…despite the fact that AL clubs get to use DHs. The Cards scored 77 more runs than 2nd place in the NL (the Rockies) because their were out of their minds with runners in scoring position – they put up an unreal .330/.402/.463 line that lapped the field. We haven’t seen anything like it in recent years, and the fact that it was “unsustainable” won’t take away the NL Pennant the cards won that year.**
When we talk about teams “beating” their lines or run differentials, these are the teams that leap to mind – not only are they recent, but they follow some established, identifiable pathways to improbable win totals. But this year’s Rangers simply don’t look like this. They’re 11 games over .500, but their pythagorean record is 74-74. Thanks to a sub-par pitching staff, the Rangers have given up exactly as many runs as they’ve scored, and before this series began, they were 7 runs in the red. So, is this a Baltimore thing where they’ve won close games and had tons of comebacks? Short answer: no. The Rangers do not have an amazing record in 1-run games, and they don’t have tons of comeback wins. On those measures, they look remarkably like the M’s. It’s not Cardinals-style performance with runners in scoring position either. The M’s have been awful, but when you park adjust, the Rangers have actually fared worse. If you prefer RE24, which just tallies the change in run expectancy for each event, both teams are solid, but the M’s pitching staff fares (predictably) worse.
| Team | OPS w/RISP | RISP wRC+ | One-run wins | Comeback wins | Bat. RE24 | Pitch RE24 |
|---|---|---|---|---|---|---|
| Mariners | 0.685 | 81 | 27-24 | 29 | 31 | -97 |
| Rangers | 0.694 | 87 | 26-20 | 27 | 41 | -42 |
From the data in the table above, I think you’d expect the Rangers to lead the M’s, but you’d probably expect both teams to be below .500. Looking at the *contextual* stats, neither the M’s look particularly good. That can happen to good teams, and it’s been a big reason why the A’s have struggled despite a GOOD run differential. So let’s take a look at some overall numbers as well:
| Team | Runs Scored | Runs Allowed | wRC+ | Pos. Player WAR | Pitcher ERA- | Pitcher FIP- | Pitcher fWAR | Pitcher rWAR |
|---|---|---|---|---|---|---|---|---|
| Mariners | 597 | 667 | 101 | 14.7 | 107 | 105 | 9 | 9.6 |
| Rangers | 666 | 666 | 95 | 17.1 | 100 | 105 | 10 | 13.7 |
The thing that jumps off the page to me is how similar the M’s look. Neither team looks particularly good, mind you, and at least the Rangers have an ~70 runs-scored advantage, but these teams look like they’d be closer than the 8.5 game advantage the Rangers actually have. This isn’t to take anything away from the Rangers or to insinuate the M’s have been horrifically unlucky. What it DOES say is that we’ve looked at a lot of numbers and *still* can’t quite identify why the Rangers are 12 games over .500. By ERA they’re not so great; they look better when you park adjust (which ERA- does), but if you do that, you have to park adjust their overall batting line, and that makes them look somewhat feeble.
Like a lot of people, including former co-blogger Dave Cameron, have been talking about this a lot, and the conclusion that he’s come to, and the conclusion that Rangers fans really, really hate hearing, is that it’s purely luck. What the data imply is that the Rangers’ ‘skill,’ if you will, is that they score runs in games when they need them, and then don’t score (or do much of anything) in losses. When you put it like that, it sounds crazy, and even looking at their batting and pitching lines in wins and losses, there’s no evidence that they perform any different than other teams. By the numbers, there’s no obvious way around the conclusion that their peripheral numbers don’t support their actual win/loss record. But that’s not interesting. So let’s do this instead: if there WERE some non-luck reason(s) the somewhat similar-looking M’s and Rangers should be 12 games over and 6 games under .500, respectively, what would it(they) be?
1: The park adjustment is missing something.
This is a flavor of measurement error; that, somewhere in the process of neutralizing and contextualizing all of these stats, Fangraphs and Bbref screwed something up. In this case, there’s a big difference in the raw stats and the park-adjusted ones. Maybe the park factors have missed something important – they may still be relying on pre-dimension-changed Safeco numbers, for example, or Texas’ may be inflated by a couple of scaldingly hot years in the recent past. Thanks largely to Nelson Cruz, the Mariners actually have a higher ISO than the Rangers – are the M’s getting more credit for that than it’s worth? Always possible, but this seems extremely unlikely. Or rather, if it affected their batting line, you’d figure it’d affect each team’s pitching WAR as well.
2: A low run-scoring environment has put a premium on things like base-running and the bullpen.
The Rangers are a good base-running team; they’ve earned nearly 12 runs above average in base-running according to Fangraphs, good for first in the AL and 3rd in baseball. The M’s, meanwhile, have the worst mark in MLB, at nearly 22 runs below average. That’s a big gap – it’s essentially as big a gap as you can get, and it just doesn’t seem like it’s anywhere close to big enough. Could there be some sort of reinforcing effect at play? A pitcher’s park PLUS bad base-running makes it exponentially more difficult to score runs, or something? Maybe, but it seems like that would show up either in the park factors themselves or in the base-running value measures.
As far as the low run-scoring environment goes, I wondered if it wasn’t a particular problem for the M’s. Recent research by David Smith concludes that a shockingly high chunk of what we call home field advantage stems from the fact that home teams outscore their opponents by a lot in the first inning. The theory is that pitching first is something of an advantage, because even if you give up some runs, the opposing starter has been sitting on the bench for a long time – essentially, the home starter goes right from warm-ups to the mound, while the visitor doesn’t exactly know when he’ll be called upon to pitch. This shows up in high run-scoring enviroments as well as low, and it persists even if the home team has just traveled. So: if the M’s were really bad at scoring overall, might they not enjoy this benefit? Is that what’s causing this gap between batting RE24 and runs? Er, no. The M’s haven’t scored a ton of runs in the first inning at Safeco, but they’ve given up even fewer. Their differential looks exactly like it should, only the numbers are lower – which is what we’d expect.
So, the bullpen. Both of these clubs have poor overall marks for bullpen performance. The Rangers’ ERA and FIP are slightly higher, but they have a small lead in WAR (1.6 to 1.3) thanks to park adjustments. Neither team is above average in any of these stats, and both are in the bottom half by strand rate, too. The distribution of pitcher WAR is different, with the M’s having both the most valuable reliever (Carson Smith) as well as the *least* valuable relievers (Joe Beimel, Fernando Rodney). The M’s have given more innings to guys having awful years, but it’s not a huge gap. Still, this is an area where depth could come into play. The M’s haven’t found a good replacement for either Mark Lowe or Charlie Furbush – Beimel was supposed to stand in for the latter, and that really didn’t work out.
3: Depth and batting order black holes.
For years, the M’s have under-scored their base runs, and for years, the M’s have struggled with offensive sinkholes towards the back of their batting order. Mariner catchers this year are hitting an unfathomable .157/.208/.260, good for a wRC+ of 29. Remember 2013, when the M’s had to rush Mike Zunino to the majors because no one could hit? They posted a wRC+ of 63 that year. This is historically awful, but could it account for the gap between RE24 and runs? Maybe, but the Rangers have struggled with some black holes (though not quite as black) too. Their CFs rank 29th in baseball in wRC+ thanks to the implosion of Leonys Martin. In fact, their OF as a whole ranks 29th, as Jake Smolinski, Ryan Rua and Ryan Strausborger have all scuffled. The Rangers have just one OF with more than 2 WAR (Choo), while the M’s have 4.
But what’s helped the Rangers is that they’ve been better about swapping out ineffective players with good ones. They didn’t go from Zunino to Jesus Sucre, they went from Carlos Corporan to Robinson Chirinos. Jake Smolinski (and Joey Gallo and Ryan Rua etc.) weren’t cutting it, so the Rangers traded for Will Venable, and they sat Martin in favor of Delino DeShields. This hasn’t always worked (Venable and Josh Hamilton haven’t done much, and Drew Stubbs has been awful), but it’s worked enough. This contrasts with the M’s, who still haven’t found a real starting catcher, and whose bullpen has gotten noticeably worse in the second half, not better.
The bullpen is perhaps a better example of this, as the Rangers have figured out who to turn to in high leverage situations. Lefty Sam Dyson and righty Jake Diekman have *only* pitched in the 2nd half, while Tanner Scheppers and Ross Detwiler haven’t been able to do more damage. That’s probably helped, but it’s worth remembering that even in the 2nd half, when the team is a combined 16 games over, they still don’t have a great run differential. Thanks to their recent sweep of Houston, it’s finally in positive territory, but it’s not great. Still, one could argue that the M’s – as a team in a low-scoring park – have been especially hurt by their bullpen’s second-half struggles. That would imply that the Rangers, with THEIR home park, haven’t needed their pen as much, which makes it harder to argue bullpen depth and sorting out the good from the bad *isn’t* the cause of their surge.
4: Good players underperforming and then bouncing back.
Many Rangers fans argue that their seasonal stats look bad due to the first half, when Adrian Beltre and Shin-Soo Choo were both hurt and bad. Prince Fielder carried them, but could only do so much. That’s similar to the predicament the M’s were in, where Nelson Cruz was dominant, but had no one to drive in thanks to Robinson Cano and Kyle Seager’s struggles. In the 2nd half, Choo’s been the best hitter in the AL West with Beltre not far behind, but Fielder’s regressed significantly. Meanwhile, the M’s got solid production from Cano and Seager, but Cruz hasn’t really dropped off. At least on the batting side, the M’s have benefited from (positive) regression and their offense has been one of the league’s best, particularly once they started getting production out of the lead-off spot – it may seem hard to remember, but the M’s OBP from their 1st hitters has impacted by several starts from Rickie Weeks and Logan Morrison.
Yet despite the fact that the M’s have a higher wRC+ in the 2nd half, the Rangers have scored more runs. The same (or different?) effects that led the Rangers to yield more runs from each batting event has held true in the second half, despite the massive swings in the performance of several of their star players. Baseball!
5: Run scoring, and how runs translate to wins, is messy at the team level.
I’d love an actual explanation, and for those who think this answer is essentially hand-waving, you’re pretty much right. But if you’ve got other theories, I’d love to hear them. The Rangers have scored a higher percentage of their total runs in wins (73%) than the M’s (65%), but it’s hard to attribute that to skill. You can argue that the Rangers have simply been better at beating the teams they *should* beat, but the M’s and Rangers are both 4 games over .500 against teams with sub-.500 records.
Because the M’s record is so driven by their poor performance at home, and because this isn’t the first time the M’s have struggled at home, I kept thinking that Safeco was part of the reason why the M’s look so different from the Rangers. But look at the M’s run differential on the road: they’ve been outscored by 47 runs on the road…and yet they’re 38-36. The weird devil-magic that we ascribe to the Rangers seems to be the same stuff that’s allowed the M’s to post a .500 record despite a pythagorean record far, far worse than their home mark.
Apparently, this kind of thing happens every now and again, and because it does, it seems harder to credit the Rangers with some innate advantage in runs-per-win. That doesn’t mean this post is dispositive, and if you’ve got alternative theories as to what the Rangers have done, I’d honestly love to hear them. I’d hope teams have a better idea about why this happens, though I’m sure they’re also able to quantify just how much luck/sequencing goes in to a team’s record.
While luck and randomness seem to play a big role, I do wonder about park effects. In isolation, the effect may not be large enough to matter, but I wonder how large the effect can get once you layer on crappy base-running, a sub-par bullpen and the like. Again, this seems like the kind of thing that would be dwarfed by the overall talent of the club, but the point of this post was to call attention to just how close the M’s and Rangers seem to be at everything but the whole “wins and losses” thing. Yet another thing I’d ask a GM candidate who wanted to take the reigns.
The Rangers have outplayed the M’s, at least when the Rangers don’t actually PLAY the M’s. The gap doesn’t appear to be large or the result of clear and consistent advantages in true talent. That’s encouraging for M’s fans, of course – this could be us someday. On the other hand, the Rangers are doing this without Yu Darvish, Jurickson Profar, Matt Harrison, etc., and their surge is correlated with 2B Rougned Odor’s hot-streak. Both clubs are getting production from young players, and both teams are dealing with adversity. The Rangers have been fortunate in some ways this year, but that doesn’t mean the M’s will be favored next season.
* Your mileage may vary. Maybe I’m easily fascinated.
** This year, of course, the Cards are doing something similar, but on the pitching side. With men in scoring position, Cards pitchers have given up a .263 wOBA, a mark that’s above all of 2 qualified batting lines. With RISP, the Cards turn the entire NL into Alcides Escobar-with-flu-like-symptoms.
Comments
9 Responses to “The Mariners, The Rangers and the L Word”
Leave a Reply
You must be logged in to post a comment.
I don’t know if it’s luck, but there’s no doubt that regardless of how good the stats are, that baseball seasons — and here we do have good luck — will never unfold with predictive determinism. I enjoy reading these articles. At the same time I know games also consist of elements like mental errors, taking risks, base-running execution, outrageously over-the-top athletic performances, confidence, fear, hate, exaltation or rage. And it’s interesting to see, and reflect on how, performance deviates from what can be statistically expected. In fact, it’s sort of comforting. There’s an element of chaos in there somewhere as well as the human element which is always going to find a way to deviate from the standard model.
June sucked.
It’s easy to blame the manager, and of course hard to actually prove… but I wonder if some of this comes down to McLendon being a poor decision-maker with regards to when he sends runners, lineup ordering, and other “little things” that lead to the team underperforming their talent level.
I kind of suspect that just as in real life, the black hole at catcher is distorting the fabric of reality around it. There are so few teams in baseball history that have gotten so little production from a non-pitcher spot in the order that our ability to quantify the impact is necessarily limited, and it could be that having what amounts to an automatic out (a slow, double-play prone automatic out no less) in the lineup could have a more catastrophic effect than we’d typically assume, especially in the first half when few other players were actually hitting. Obviously they’ve been largely able to overcome that in the second half.
Of course that might just help to explain why the Mariners have underperformed this year, not why the Rangers have overperformed. My guess is that it’s a combination of a number of factors you’ve discussed: baserunning, bullpen effectiveness, maybe a few slightly-errant park adjustments, and of course a big ol’ helping of luck. For once it would be nice to see the M’s on the lucky side of things, but I’m not counting on it any time soon.
I agree with Westside Guy although I can’t prove it. I consider Lloyd to be among the most clueless tactical managers I’ve ever seen…having players sacrifice with Zunino on deck, handing out intentional walks like Marshawn hands out skittles, hitting and running with low-contact batters and low-percentage stealers, burning through the bullpen so that he has no options in an extra-inning game, etc, etc.
The road record versus run differential can be explained by six games where the M’s were outscored 82-12: 4/10 at Oak (0-12), 6/12 at Hou (0-10), 6/14 at Hou (0-13), 8/14 at Bos (1-15), 8/15 at Bos (10-22) and 9/17 at Tex (1-10). There have been no comparable home losses…the worst home loss was by 8 runs (8/8 vs. Tex, 3-11).
What I find a bit frustrating about this analysis is that you can see how small the difference is between success and failure, and that makes all the bad base running, the high-leverage innings pitched by bad relievers, the lead-off at bats by Rickie Weeks, the failure to find a backup catcher better than Jesus Sucre, all the more aggravating. There’s that nagging feeling that the organization does a lot of little things wrong, and those little things add up.
Part of the problem, to me, is that we try to get precision from less than precision. The W/L record is the result of everything that happens on the field and we try to explain it with a partial representation of what happens on the field.
Give the Rangers just 4 more losses and the Mariners 4 less losses and they have essentially the same record. Looking at team performance, there is a range of possible outcomes with the team W/L record being one of them.
I don’t like using the word luck because part of that luck is just the inability to identify relevant factors that influence results. Sometimes something stands out to explain an anomaly, but many times not, like the Rangers’ W/L record. It’s hard to fully explain something when not all the relevant factors are known.
Since “luck” or randomness is such a big factor in baseball the 162 game season would seem to be enough to even out that randomness.
Some of it is bad strategy. “The contact play” unless you are trying to avoid a DP would be an example for me. Major league infielders seldom throw the ball away.
Then there is poor fundamental execution. Not scoring a runner at third with less than 2 outs. Which to me is failure to adjust to a situation. A pop up and a groundball to a middle infielder playing at normal depth both show up as an out when figuring BARISP. One produces a run and the other doesn’t.
Base running, not so much stealing or hit and run failures. But the inability to score from second on a”normal” outfield hit. Or go 1st to 3rd on a ground ball single to right. A lot of that is the Jack Z, multi- DH roster.
This roster needs more Marte, and less Montero. Speed, the ability to catch and throw and put a ball in play are what this roster lacks.
Nice article and good follow-up comments. For me (and I assume others) things can be broken down in even simpler (albeit probably unfair) terms:
Brad Miller threw away 3 games with errant late inning throws … and Fernando Rodney.
I like Miller (but not as a SS) and I hated Rodney, so I’m not cherry-picking a player preference, just attempting to break things down to as fine of an “in game situation” point as possible.
The team is probably 5-6 wins behind what “could have been” if routine throws were made at a critical time of the game or if Rodney’s blown saves had been cut by one-third.