Looking Back at the 2014 Postseason

Baseball is over for 2014, after a postseason closing that was a curious blend of the memorable and the forgettable. There were great games, and a World Series that went the limit, but perhaps you get the sense that it didn’t really come together. There are reasons for that impression, and I’ll be exploring that today, with some help from a creation of mine that people who follow my writing here are likely familiar with.

My previous two years writing at The Hardball Times, I did daily reports on postseason games, supplemented with a metric I called the Win Percentage Sum, or WPS. With the change in format that came to THT early this year, posting daily reports was no longer practicable. (It was also personally problematic, with up to four games a day plus writing time afterward straining what stamina I have.) I can, though, keep the WPS index alive with a post-postseason wrap-up, and at the same time shed light on the playoffs just past.

For our newer readers, you can read my original introduction to the WPS system here and here. Here I must include credit to someone else. THT’s Dave Studeman invented a system of measuring game excitement very similar to mine several years ahead of me, publishing it in the 2007 Hardball Times Annual. I was unaware of this when I produced my own system, and quite galled when I discovered I had poached on his turf. With his forbearance, along with his doubts about the new parts I accreted to his system, I continued to use WPS.

For those not inclined to follow the explanatory links, I’ll summarize WPS here. The base score is tabulated by adding together the change in Win Expectancy in the game created by every play, meaning every continuous action that changes the base-out state of the game. For example, if a player’s strikeout lowers his team’s chance of winning by three percentage points, that’s a score of three. If Kirk Gibson’s two-out two-run homer in the bottom of the ninth takes his team from a 13 percent chance of winning to actual victory, that adds 87 points to the WPS.

To this base score I add two supplements, meant to reflect a fan’s subjective enjoyment of the high points of a game. I take the three plays with the highest Win Expectancy changes, and add them to the base. I also add the value of the game’s final play, showing how much the result was still in doubt at the very end.

The final score that results can vary greatly. I’ve found that 300 is roughly the average score of a game, but they can range from below 100 to above 1000, both very rare but not unheard of. A game under 200 is reliably a bore (unless it’s your team winning), while I set the benchmark of a “great” game at 500 points.

Admittedly, there are holes in this system. It has no mechanism for measuring outstanding individual performances or milestones, save in how they affect the game’s competitive excitement. Albert Pujols’s three home runs in a World Series game barely interest the WPS system, as they came in a game already out of hand. Don Larsen’s perfect game scored an awful 135 exactly because the Dodgers had zero success against him.

What WPS measures is the competitive excitement of a game. It doesn’t measure the tension of a pitcher’s duel, unless the pitchers are pulling Houdini escapes to keep the string of zeroes intact. It doesn’t measure the historic import of arbitrary milestones: 3,000, 715, 0/0/0, all mean nothing to it. WPS likes a close game; it likes offense, or at least strong scoring threats. It doesn’t perfectly mirror what the baseball audience likes in a game, but it comes pretty close.

It is with this tool in hand that I examine the 2014 postseason. First, though, some calibration. I’ve already given a couple sentences to what the scores mean, but historical context can help fill out that knowledge. Counting this year, there have been 1,438 postseason games, and that’s a lot of context.

The table below shows the mean and median WPS values for each playoff round, plus the whole postseason taken together. Means are higher because of the scale: there’s much more room for a game to go above average than below. I have separated the League Championship Series into best-of-five (1969-1984) and best-of-seven (since 1985). I have also folded in the Wild Card knockout games with the League Division Series, as six games isn’t really much of a sample.

Mean and Median WPS Values by Postseason Series

Round	# Games	Mean	Median
WS	641	315.9	287.9
LCS(7)	334	336.7	313.85
LCS(5)	125	325.9	300.4
LDS+WC	338	337.4	308.35
All	1438	326.7	297.9

An explanation for the underperformance of World Series games is run environments. The WPS system likes high-scoring games that get the Win Expectancy needle constantly moving. Many World Series games were played during deadball eras, first in the 1900s and 1910’s, then later in the 1960s. The LCS round debuted in 1969, the year the mound was lowered and baseball emerged from its second scoring drought.

Highs and Lows

Now that you have some idea of what the numbers mean, I can start throwing them at you. And I may as well start at the top, with the most exciting game the postseason gave us this year. You’ll get two line scores with each game I present, one for the runs, and one for the WPS score as it’s accumulated inning by inning. I took the Win Expectancy data from FanGraphs, which comes in tenth-of-a-percent increments. The line scores round them off to whole numbers, but I do the calculations by the tenths, not that it changes the totals dramatically.

(Added note: I have since discovered that Win Percentage Added numbers have changed since the original day-after postings, meaning WPS scores would be slightly different if I tabulated them now. The changed numbers are due, I believe, to adjustments for run environments, going from generic to more specific. I’ll have to grit my teeth and go with the numbers I took down in the immediate wake of events. The differences will not be massive.)

It will surprise few who saw it that the best postseason game of 2014 was the very first.

Most Exciting Game of the 2014 Postseason:
American League Wild Card Game

Game     1  2  3  4  5  6  7  8  9  10  11  12   F       
A's      2  0  0  0  0  5  0  0  0   0   0   1   8
Royals   1  0  2  0  0  0  0  3  1   0   0   2   9

Game     1  2  3  4  5  6  7  8  9  10  11  12
A's     30  9 16  6  7 72  1  1 12  13  19  42
Royals  28  5 37  4  4  3  4 88 48  28  28 112

WPS Base: 615.5  Best Plays: 101.5  Final Play: 39.8  Grand Total: 756.8

For its first two years, the Wild Card game was a bit of a fizzle. It provided two average games, one slightly below-average one, and an outright snoozer in last year’s NL showdown. (That Reds-Pirates game was memorable for the amazing vocal enthusiasm of playoff-hungry Pittsburgh fans at PNC Park, but the game itself was too lopsided for it to engage neutral viewers.) That string of so-so was shattered by the greatest Wild Card game in the short history of this playoff round.

The game did have its dead spots: the fourth and fifth, when nobody got on base, and the four half-innings after Oakland’s five-spot put the A’s up 7-3. Those were easily overcome by the big swings in fortune this game provided. Multiple lead changes, along with multiple extra innings, are the classic recipe for piling up WPS points. Perhaps the ideal example of this is Game Six of the 2011 World Series—and not coincidentally, this Wild Card game was the best postseason contest since that game.

Do please note the combined score for the 12th inning: 154 points. That’s half the excitement of an average game right there. That’s more excitement than some entire games provide. Speaking of which …

For the least exciting postseason game of 2014, it looked for a long time like we’d have symmetry, with the Giants’ 8-0 blanking of the Pirates in the NL Wild Card game taking the dishonors at 156.4 points. This would also have reflected the symmetry of Madison Bumgarner’s postseason, as his first and last starts were four-hit shutouts that got boring fast. But the Giants did us one worse, and it ended up worse for them, temporarily.

Least Exciting Game of the 2014 Postseason:
World Series, Game Six

Game     1  2  3  4  5  6  7  8  9   F
Giants   0  0  0  0  0  0  0  0  0   0
Royals   0  7  1  0  1  0  1  0  X  10

Game     1  2  3  4  5  6  7  8  9 
Giants   4 14  8  1  0  0  0  0  0
Royals 12 52  2  0  0  0  0  0  0

WPS Base: 95.1  Best Plays: 29.1  Final Play: 0.0  Grand Total: 124.2

In football, a team can go up seven in the first quarter of play, and there’s still a good chance you’ll have an exciting game. That happens far less in baseball, and it requires the trailing team to make some kind of comeback, preferably fast. “Never” doesn’t qualify. Look at that wall of zeroes for the back half of the game. No sugarcoating that.

I intended to list the most exciting inning of the postseason here, but you’ve seen it already: the 12th inning of the A’s-Royals game. The 112 points put up by Kansas City in the bottom half looks like it would stand as the best half-inning as well, but it doesn’t. Outpacing it is one of the biggest comeback innings—or from the opposite viewpoint, one of the biggest meltdown innings—ever seen in postseason play.

Most Exciting Half-Inning of the 2014 Postseason:
Top of the 7th, NLDS Game One, Cardinals at Dodgers—133.3 points

The playoff immolation of Clayton Kershaw was like—well, various real disasters come to mind, but comparing a pitcher getting lit up to something that kills people lacks proportion. Except in one sense: you cannot believe it, and you have to keep watching, if only to convince yourself that it’s truly happening.

Kershaw carried a 6-2 L.A. lead into the seventh, and departed two outs and five runs later, before reliever Pedro Baez gave up the three-run bomb that made it 10-6, St. Louis. The swing from massively ahead in a late inning to massively behind insured a big WPS number, but compounding it was how Kershaw nearly saved himself. A 6-3, bases-loaded strikeout of Pete Kozma and a 6-4, bases-jammed strikeout of Oscar Taveras (it still hurts to write the name) produced sizable pendulum swings, which only made the final swing to the Cardinals that much bigger when it happened.

That half-inning also contains the biggest Win Expectancy swing, and thus the biggest single-play WPS score, of the postseason. Matt Carpenter’s two-out, bases-clearing double turned a 6-4 Dodgers edge into a 7-6 Cardinals lead. Depending on when you looked it up on FanGraphs, that play was worth 48.8 or 50.7 percent of a win, half a game’s worth with one swing of the bat. And he didn’t even need a homer.

The Longer View

Taking a step back to look at the playoff series themselves, we find a peculiar pattern that holds for the first two rounds after the Wild Card games, then reverses itself for the Fall Classic. The divisional round introduced the pattern, which I have to describe as fine wine in cheap mugs.

WPS Values, 2014 Division Series Round

Game	DET/BAL	KC/LAA	STL/LAD	SF/WAS
Gm. 1	277.4	460.6	467.0	368.0
Gm. 2	412.6	476.0	347.5	686.8
Gm. 3	415.3	215.5	343.2	251.5
Gm. 4	–	–	370.3	329.2
Total	1,105.3	1,152.1	1,528.0	1,635.5

Of the 14 games, only three came in below the median WPS for playoff ball. The lone “great” game was the 18-inning epic between the Giants and Nationals in Game Two, but three others made close runs at my arbitrary 500-point cutoff. Kansas City played, and won, its second and third extra-inning games in a row, and extra-inning affairs rarely disappoint on the WPS scale.

This was a really good set of games. It’s the series they were played in that were no great shakes: two sweeps, and nothing going the whole five. We could be forgiven for feeling shortchanged, even if San Francisco and Washington did play five games’ worth of innings.

The WPS totals for the series end up pretty middling. Out of 84 LDS ever played, they tally at 21st, 33rd, 52nd, and 56th. If we used a Championships Added multiplier, the range would fall, but the simpler additive model may reflect our experience better. The series were too short to be very good, but the games were too exciting for any of the series to be very bad.

The pattern, slightly altered, maintained itself for the Championship Series.

WPS Values, 2014 Championship Series Round

Game	KC/BAL	SF/STL
Gm. 1	596.0	204.6
Gm. 2	448.7	583.4
Gm. 3	229.4	365.0
Gm. 4	290.1	365.0
Gm. 5	–	485.2
Total	1,564.2	2,003.2

The NLCS listings for Games Three and Four are not a misprint: they scored exactly the same. There’s no granularity in the rating system that would cause any clumping. It’s just a heck of a coincidence.

This round had a higher proportion of subpar games, three out of nine, but also produced two great games, plus a third that was awfully close. The series themselves starved us worse this time around, going a combined one game over the minimum.

By WPS totals, The Giants and Cardinals had the 23rd most exciting League Championship Series out of 58 that have been best-of-seven, with the Royals and Orioles coming it at 43rd. Again, adjusting for Championships Added would send both series dropping.

Seen through another lens, though, they look much better. Seven League Championship Series have been four-game sweeps, and the Kansas City-Baltimore series has the second-best WPS rating among them, behind only the 1995 Atlanta Braves’ blanking of the Cincinnati Reds. Of the 15 Championship Series to be decided 4-1, San Francisco and St. Louis played the most exciting of them all.

And it could well have been better still. Three of the NLCS games were walk-off wins, including Travis Ishikawa’s Game Five homer that got Joe Buck quoting other people’s classic calls once again. All three walkoff victories, though, came with no outs. The vector from the start of the inning was straight to victory, with no resistance. It’s the reversals of fortune, along with the sense that things are still unresolved, that makes a baseball game exciting.

The walkoff games missed chances to raise the excitement level. By playing out too quickly, the Giants’ clincher fell short of my “great game” threshold that it would have reached with even one out intervening before Ishikawa won it. (By now, you’re probably expecting me to complain that Bobby Thomson should have gotten himself out so it could be Willie Mays firing the Shot Heard ‘Round the World with two outs instead of one. I wouldn’t go that far … though it would have added a good 60 points …)

So for two rounds, we had short series full of very good games. That changed for the World Series. We got to Game Seven, but the route was a lot less scenic than we would have hoped. Or to revive my earlier metaphor, it was a crystal goblet full of Two-buck Chuck.

WPS Values, 2014 World Series

Teams	Gm. 1	Gm. 2	Gm. 3	Gm. 4	Gm. 5	Gm. 6	Gm. 7	Total
Giants/Royals	159.1	278.9	257.1	343.2	174.0	124.2	323.3	1,659.8

The Giants and Royals combined to lay three rotten eggs in Games One, Five, and Six, along with two other subpar tilts. The two games that get above the mean don’t do so by much. This probably surprises some fans, who would argue that both were, or at least Game Seven was, great. Putting aside the obvious subjectivity of “great,” my argument is that they’re not teeming with the excitement that WPS is best at measuring.

For examples, I’ll dip back into the ALCS. The third and fourth games finished in the sleepy 200s, despite both being tight 2-1 affairs. The problem is that they were too tight. The Royals and Orioles combined for a mere 19 hits over those two games. A modest rally could have swung either game, but the pitchers were choking off those rallies. Tension, yes; excitement, I say no.

Similar things happened in the World Series. In Game Three, the teams combined for 10 hits, right in line with the ALCS contests. In parallel streaks lasting from the second to the sixth inning, Tim Hudson for the Giants and Jeremy Guthrie of the Royals retired 12 and 11 consecutive batters respectively. This interleaved into a string of 20 straight batters retired in the game. Later, the two relief corps combined to get the final 15 batters of the game out.

These streaks happened with one-run leads on the board, so suspense was there, with the potential for great excitement. Nevertheless, these stretches of pitching dominance put the game in stasis for long periods, including what could have been a nail-biting ending. It was a good game, but couldn’t quite launch itself to something better.

Game Seven had similar problems. After some early offense, the two bullpens threw a combined 13 innings that went five batters over the minimum. No reliever needed to throw to more than four batters in a frame, not one was charged with a run, and Bumgarner centered his relief effort on a string of 14 straight batters retired. A game with the potential to be a classic was put in a headlock, and only a fielding gaffe with one out to go jolted it out of what would have been a sub-300 WPS score.

The overall WPS score for the series is a mediocre 69th out of 110 World Series played, raised that high solely by its length. A better comparison uses the 36 World Series that went to a decisive seventh game. (The 1920 Series also ended in seven, but is excluded here because it was best-of-nine.) In that group, the 2014 Series comes in 32nd, or to look the other way, it was the fifth least-exciting seven-game World Series ever. (Rock bottom was the 1965 Dodgers-Twins series, at 1484.8.)

Past Time

There is one other matter I will cover regarding the 2014 postseason. This has been a burr under my saddle ever since I began doing postseason write-ups, and I didn’t exactly hold my tongue about it the previous two years, so why would I start now? Especially since MLB itself has had to take notice of it, and is doing some preliminary work in the Arizona Fall League to produce a remedy. I am talking about the length of games.

It’s no news flash that games are getting longer. It’s perhaps a bit more of a news flash that games during the postseason have recently been substantially longer than those in the regular season. Lengthening games aren’t so much of a chore when start times are at seven o’clock or thereabouts. When the start gets pushed back past eight in the evening, Eastern Time, and the game itself gets longer, it becomes a trial. Worse, it shuts off viewing by children, the next generation of fans, who are getting sent to bed mid-game, assuming they haven’t learned not to bother starting to watch in the first place.

One could say this reflects an East Coast bias. I prefer to say that it reflects an Eastern Time Zone perspective. I grasp that MLB may want to get part of the World Series broadcast into Pacific Zone prime-time hours, for the extra ad revenue. (A similar motive explains why the Super Bowl, once an afternoon affair, now extends to 10 o’clock Eastern time). Nevertheless, it adds up to fattening themselves on the seed corn of their eastern fan base, of which I’m part.

(Perhaps the real problem is the outdated concept of prime time having automatically inflated ad rates in an era of DVRs, streaming, and time-shifted viewing, but I will not tackle that subject here. Baseball, Shane. Write about baseball.)

In the 2014 regular season, the mean duration for a baseball game was 3:07:48. That’s up three and a half minutes from 2013’s figure of 3:04:14. To even out the effects of extra-inning games, for reasons which will be plain in a couple paragraphs, I also calculated game pace by time per inning. 2014’s games came in at 20.93 minutes per inning, up from 2013’s rate of 20.52.

For the playoffs, the numbers go up, and not by a little. The 38 games played in the 2013 postseason took an average of 3:23:32, more than 19 minutes longer than in the regular season. Rendered by minutes per inning, the postseason pace was 23.06 minutes per six outs, beating the pre-October number by 2.13 minutes. I noted at the time the glacial pace of certain Boston Red Sox pitchers, and having them go all the way definitely contributed to some long-term insomnia out east.

This year, the length of games got out of hand, at least by straight measure. The average postseason game lasted 3:37:53, half an hour longer than in the regular season and a quarter-hour longer than the previous year’s playoffs. A major reason, though, was a generous helping of extra-inning contests. The Giants-Nationals epic was one of six games, out of 32, to go past nine. That compares to one in the last postseason. A six-hour, 23-minute marathon of baseball is bound to slant the figures.

This is where the pace by inning comes into its own. Games in the 2014 postseason averaged 23.44 minutes per inning. This beats the regular season tally by two and a half minutes, but is a modest 23 seconds per inning slower than last year’s playoffs. Going by that pace, an average nine-inning game would have been three and a half minutes longer this postseason than last, which is almost exactly how much longer the regular season games were.

The extra time that games take in October can be attributed to MLB’s policy of adding commercials in the playoffs, to the tune of a full minute for each between-inning break. With games going longer by two and a quarter or two and a half minutes per inning, more ads explains most of that, possibly all if breaks for pitching changes are thrown in. (I’m not sure whether those were super-long breaks like between innings. I didn’t realize at the time that I needed to put a stopwatch on the commercials.)

So games are longer, postseason games are much longer, and more than once I was tempted to take a nap at the seventh-inning stretch. What does this do to the whole concept of game excitement, the things WPS is supposed to be measuring?

If the same stuff is taking longer to happen, it stands to reason that excitement is dissipated or diluted. Perhaps a more useful way to measure game excitement is to divide the WPS score by the duration of the game. So as a brief experiment, let me do that.

Take the combined WPS scores of all 32 games from this year’s postseason, and we get 11,561.3 points. The combined duration of the games was 6,972 minutes. That gives us a WPS rate of 1.658 points per minute. Do the same with last year’s playoffs, and we get 11,684.1 divided by 7,734 minutes for 1.511 WPS per minute. Longer games didn’t result in more boring games this year, because the games were simply better, and by a pretty wide margin (at least until the Series).

The best game of the 2014 postseason, the AL Wild Card, remains so by WPS per minute, barely. The Wild Card game comes in at 2.655 points per minute; its closest competitor, Game Two of the NLCS, had 583.4 points in 3:41 for a 2.640 pace. The Giants-Nationals marathon lasted 6 hours, 23 minutes, so its 686.8 points come in at 1.793 per minute. The worst game remains Game Six of the Series, dragging along at 0.618 WPS per minute.

For a comparison, I will offer the very first World Series. The eight games played in 1903 totaled 1,861.3 WPS points, a dullish 232.7 apiece that is low even for Series of the Deadball Era. The games, though, were rapid affairs, the longest taking 2:02 and the average settling in at 1:48. Do the division, and you find the 1903 World Series producing 2.154 WPS points per minute, half a point per minute better than the 2014 postseason.

You may take this as an indictment of modern game pace, or of the WPS per minute statistic: your choice. Indeed, if you’re as liable to be driven to distraction by the Royal Rooters’ constant renditions of “Tessie” as the Pirates were, you might well prefer a glut of Viagra ads.

And with that distressing image, I close my overview of the 2014 postseason. For those of you hoping for something on a more intimate scale, fret not. Brad Johnson will be along tomorrow to provide just that, and I am as eager to read his piece as I was to write mine.

References and Resources

Baseball-Reference and FanGraphs provided the vital statistics on the postseason, past and present near-past.