WPS and the postseason, part one

A couple months ago, readers at The Hardball Times read my idea for a system to calculate the excitement level of baseball games. The WPS Index (introduced here) basically counted up the Win Percentage Added (WPA) of each distinct play in a baseball game into a Win Percentage Sum (WPS), measuring how far the probabilities of victory swung. I made adjustments, adding the values of the three most exciting plays in the game and the final play, and tested the index on a famous list of exciting games here.

The method is quick and a little dirty, but it holds up pretty well, and gives regular fans a tool to study games themselves.

That doesn’t mean I can’t keep using it, though, when I’ve got a good excuse. October is just such an excuse, with the playoffs right upon us. For anybody with a game-excitement system, it’s a natural time to look, not only at the games to come, but to postseason games of the past.

That’s what I’m going to do. It’s not precisely a new idea—Chris Jaffe did a very similar thing last year with his great-games system—but two views are better than one, right? Consider me Mr. Parallax.

I will be going into a bit more detail than Jaffe in one area. While he counted all playoff rounds in his summary of the best postseason series, I’m going to handle them separately. This week, it’s the League Division Series, along with the “tiebreakers,” those games that used to be called “playoffs” back in the days without divisions, iPhones or sliced bread. Next week, I’ll cover the League Championship Series. Two weeks after that, just in time for Game One, it will be the World Series.

Full disclosure

Before I get to the hot baseball-list action, I need to give credit where it’s due. While I did develop my WPS Index independently, I cannot truthfully say that I invented it. Five years ago, THT’s own Dave Studeman planted the flag on this territory, creating a Win Percentage game-excitement index that he wrote up for the 2007 edition of the Hardball Times Annual.

I am a bit new to the THT family, and I had not bought this annual or read Dave’s article before doing my WPS work (an error I have since corrected). I can plead ignorance, though that forgives only so much. I should know from previous experience that anyone who presumes to be a writer must be a reader: One must know the field, not least to avoid treading old ground, repeating old work. I hadn’t read as much as I needed to, and that’s my failing.

(Though why THT’s long-time readers didn’t alert me to my inadvertent plagiarism remains a mystery. Come on, people: catching our errors is half the fun!)

And while I’m talking about the Annuals, I should mention that not only was there a differing perspective on game-excitement calculations in last year’s Annual (in Sam Hendrickson’s article titled, naturally, “Quantifying Excitement”) but there will be yet more on this area in this year’s Annual, if this interests you as much as it does me. So that’s three Annuals you need to buy: 2013’s in a month or two, 2012’s and 2007’s now.

Maybe it can wait for me to finish this article. Or maybe not.

Ground rules, and the tiebreakers

There are two sources one can consult for the WPA data used to figure the WPS Index: FanGraphs and Baseball-Reference. I used B-R in my previous articles, and use it here, because its records cover many more years than FanGraphs. I said then that its WPA numbers weren’t quite as accurate, rounding to the nearest percentage point rather than to the nearest tenth, but made lemonade of this lemon by noting that the lack of decimals made scanning the inning-by-inning WPS numbers much cleaner and easier.

Since then, I’ve had a forehead-slapping moment. Baseball-Reference does provide the more accurate tenth-of-a-percent numbers: in its expanded box scores, it gives WPA+ and WPA- for each batter, counting up the positive and negative plays each one produced. Add up the positive values of team WPA+ and WPA- for both sides, and you get the base WPS Index (not yet counting the Best Plays and Last Play).

So for this three-part series, I will be using those more accurate numbers from B-R. When I want to show the progress of excitement throughout a game, I will use the older, full-point rounding method, which won’t mess up the numbers by more than a few points either way. (I will be using another method elsewhere, but I’ll tell you about that at the end of this article.)

With that explained, I will move straight into the most exciting “tiebreaker” games ever played. There have been 19 such games in major league history, from the two-game Dodgers-Cardinals series that decided the 1946 NL pennant to the single game in 2009 that decided the AL Central champion. Anybody who read my WPS survey of the major leagues’ “20 Greatest Games” list may be able to guess the winner here:

No. 1. 10/6/2009, Tigers at Twins (AL Central Tiebreaker)—916.6 points

Game       1   2   3   4   5   6   7   8   9  10  11  12    F
Tigers     0   0   3   0   0   0   0   1   0   1   0   0    5
Twins      0   0   1   0   0   1   2   0   0   1   0   1    6

WPS        1   2   3   4   5   6   7   8   9  10  11  12
Tigers     5  27  37  10   6   4  20  59  76  62  15  70
Twins      9   5  29   6   6  44  61  12  29 129  15  38
WPS Base: 771.6  Best Plays: 116  Last Play: 29  Grand Total: 916.6

By WPS Index, this is one of the greatest games ever played, at least that doesn’t run 18 innings and tire everybody out. It stands almost 100 points better than No. 2, the 13-inning San Diego-Colorado NL Wild Card tiebreaker in 2007, that made “Rock-tober” a household word for a few weeks.

A Hardball Times Update
Goodbye for now.

In case you’re wondering—and few baseball fans wouldn’t—”The Shot Heard ‘Round the World” finishes a fairly distant fourth, at 537.4 points. Ahead of it is the second game of the 1959 playoff that clinched the Dodgers’ pennant win over the Braves, a bit of a forgotten classic. Down 5-2 in the bottom of the ninth, L.A. rallied for three runs to tie, then won in the last of the 12th with a two-out rally capped by Felix Mantilla’s throwing error trying to nip Carl Furillo on an infield hit.

The least exciting tie-breaker game on the list is the first game of the 1962 playoff between San Francisco and Los Angeles, an 8-0 snoozer limping in at 133.1 points. The other two games of that series were strong enough, though, that the 1962 series has the best combined score of all four three-game NL playoffs, edging 1951 by a 1127.3 to 1106.4 score. (B-R doesn’t have WPA numbers for the 1946 NL playoff, ironically because the games are considered regular-season affairs, but the Cardinals’ mini-sweep wouldn’t have beaten the ’62 showdown.) If you want to know more about the 1962 series, be sure to look at Chris Jaffe’s THT Live entry today about its concluding game, played half a century ago today.

Game by game

Now we get to the League Division Series. There have been 68 of these series since they were instituted in 1994 (and began actually playing them in 1995). However, the strike season of 1981 cobbled together a split-season format with an extra round of playoffs to determine division champions. I am counting those games as League Division Series, making 72 in all.

I’ll start the overview with a quick list of the least exciting LDS games of all time. I meant to make it a top five, but it goes to six for a reason I will promptly explain. (I won’t provide links for the games: I have some mercy.)

   Year/Game     Teams     WPS Score
6. 2002 Gm.2  Atl over SF    145.1
5. 2005 Gm.1  ChA over Bos   141.5
4. 2011 Gm.1  TB  over Tex   133.7
3. 2002 Gm.4  SF  over Atl   125.1
2. 2006 Gm.4  Det over NYY   119.6 [Series clincher]
1. 2010 Gm.1  Phi over Cin   110.2*

In my very first WPS article, I laid out some of the weaknesses the system has, and one of them is that it makes no adjustment for historic achievements. This is how Roy Halladay pitching the second no-hitter in postseason history ends up counting as the dullest LCS game ever.

In one sense this is fair: giving 2010 Roy Halladay a four-run lead after two innings takes most of the suspense out of the day. Still, the game fell through a loophole, and you are fully justified in mentally erasing it from this list. You could plausibly do the same for No. 4, when Rays rookie Matt Moore wowed fans by shutting down the Rangers. (At least he wowed me.)

As for the remaining games, two come from the same series, the Giants and Braves in 2002. Ironic how a series that went to a winner-take-all fifth game could be so dull in getting there. Also, the worst non-Halladay game ends up being an elimination game, the Tigers snoring the Yankees out of October ball, 8-3. One can argue that late games in a reasonably close series are inherently more exciting than the early contests, and should be judged differently. I will return to this theory later.

Now for the good stuff. Here are what the WPS Index judges the 10 best LDS games ever. I’ll give just the list and links for 10-6, with the full treatment for the top five.

    Year/Game     Teams     Inn. WPS Score
10. 2009 Gm.2  NYY over Min (11)   706.4
9.  1995 Gm.1  Atl over Col (9)    706.6
8.  1995 Gm.5  Sea over NYA (11)   722.7 [Clincher]
7.  2004 Gm.2  NYY over Min (12)   743.5
6.  2000 Gm.3  NYM over SF  (13)   779.7

The 2009 Twins-Yankees game was a wild one, not least because of a blown fair/foul call in the 11th that kept a lid on a Minnesota rally. The Braves-Rockies game is the only one in the top 10 that ended in regulation, Atlanta staving off a bases-loaded rally in the ninth. The decisive Yankees-Mariners game in 1995 is on the “20 Greatest” list, but WPS thinks there was a better one in that series.

No. 5. 10/3/2003, Giants at Marlins (NLDS Game 3)—792.8 points

Game       1   2   3   4   5   6   7   8   9  10  11    F
Giants     0   0   0   0   0   2   0   0   0   0   1    3
Marlins    2   0   0   0   0   0   0   0   0   0   2    4

WPS        1   2   3   4   5   6   7   8   9  10  11
Giants    12  19   7   6  19  41  40  25  35  31  49
Marlins   23   4   9   6   3  29  13  23  24  44 133
WPS Base: 597.8  Best Plays: 122  Last Play: 73  Grand Total: 792.8

Game Four of the series made MLB’s “20 Greatest” list, but this game outstrips it by 300 WPS points. From the tying rally in the sixth onward it’s consistently exciting, but not really extraordinary until the 11th. Coming from behind in an extra inning to win is rare, and you’ll find few more exciting examples than this.

Ivan Rodriguez not only smacked the two-out, bases-loaded, two-run single to win this game, but he made the game-saving tag at the plate to win the game and series the next day. Showoff.

No. 4. 10/1/2003, Red Sox at A’s (ALDS Game 1)—815.6 points

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

WPS        1   2   3   4   5   6   7   8   9  10  11  12
Red Sox   15   5   9   6  29  17  63  32   4  15  31  50
A's        5   9  40   6  19   5  43  15 109  26  37  57
WPS Base: 663.6  Best Plays: 118  Last Play: 34  Grand Total: 815.6

The WPS system loves see-saw innings, momentum swinging back and forth. The A’s ninth was a glittering example, running up a huge half-inning score on just one run. The extra innings built the tension, each frame coming a little closer to a breakthrough than the last, until Ramon Hernandez’s three-on, two-out bunt single broke the deadlock. For one game, Billy Beane’s you-know-what worked in the playoffs.

No. 3. 10/9/2005, Braves at Astros (NLDS Game 4)—921.2 points

Game       1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18   F
Braves     0  0  4  0  1  0  0  1  0  0  0  0  0  0  0  0  0  0   6
Astros     0  0  0  0  1  0  0  4  1  0  0  0  0  0  0  0  0  1   7

WPS        1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18
Braves     5 19 46  4  8  6  3  2  8 37 52 33 14 57 26 14 38 26
Astros     5 10  4  4 19  4  8 36 69 29 15 15 20 15 30 15 15 48
WPS Base: 756.2  Best Plays: 123  Last Play: 42  Grand Total: 921.2

Nothing can redeem the middle-inning blahs like lots of extra frames. The Astros proved it in Game Six of the 1986 NLCS, losing a sizable lead to the Mets and bowing in 16, and they demonstrated it again here, this time coming back against Atlanta and finally getting the walk-off homer from Chris Burke. Note that the Braves had the two most exciting extra innings, without managing to push one across. WPS likes offense, but you don’t always need to score to rack up the points.

No. 2. 10/4/1995, Mariners at Yankees (ALDS Game 2)—927.3 points

Game      1   2   3   4   5   6   7   8   9  10  11  12  13  14  15    F
Mariners  0   0   1   0   0   1   2   0   0   0   0   1   0   0   0    5
Yankees   0   0   0   0   1   2   1   0   0   0   0   1   0   0   2    7

WPS       1   2   3   4   5   6   7   8   9  10  11  12  13  14  15
Mariners  5  16  23   5  14  37  40  21  32  29  15  54  15  15  43
Yankees   7  13  15  13  30  41  34  16  15  26  15  82  30  35  48
WPS Base: 783.3  Best Plays: 108  Last Play: 36  Grand Total: 927.3

Game Five of the series would eclipse this game in fans’ memory, which is a pity. This classic had almost no dull innings, a Ken Griffey Jr. homer in the 12th followed by a last-lick Yankees comeback (and WPS underrates that moment, due to a run and out on the same play), then the walk-off dinger in the 15th. There was even more historic import that WPS can’t tabulate: Don Mattingly’s only postseason homer, and the world’s introduction to Mariano Rivera, Reliever of Doom. This is the third 1995 LDS game in the top 10 … but not the last.

No. 1. 10/3/1995, Red Sox at Indians (ALDS Game 1)—932.9 points

Game      1   2   3   4   5   6   7   8   9  10  11  12  13    F
Red Sox   0   0   2   0   0   0   0   1   0   0   1   0   0    4
Indians   0   0   0   0   0   3   0   0   0   0   1   0   1    5

WPS       1   2   3   4   5   6   7   8   9  10  11  12  13
Red Sox   5  10  30   6  11   4  10  52  36  55  46  53  31
Indians   5   5   6  26   7  58   5  33  15  38  80  75  57
WPS Base: 760.9  Best Plays: 126  Last Play: 46  Grand Total: 932.9

The very first day of (non-strike) LDS play produced its most exciting game, though by a slender margin. There were dry spells as late as the seventh, but the eighth onward was almost continuous action and tension, with but a single 1-2-3 inning. Tim Naehring and Albert Belle traded homers in the 11th for a big WPS spike, but the 12th, with twin rallies that fell shy, was just as exciting. The winning blow came from backup catcher Tony Pena, the walk-off home run that is so common when the WPS numbers start reaching the edge of outer space.

1995 was an amazing year for the League Division Series, spoiling fans for the future. Too bad I was on strike that season. (Turnabout is fair play, right?)

Series by series

For the series numbers, I am following the simplest method: add up each game and present the total. Once again, I’ll start with the least exciting series, for completeness’ sake and to get them out of the way early.

   Year     Teams      WPS Score
5. 1997  Atl over Hou    722.7
4. 2008  LAD over ChN    712.1
3. 2000  StL over Atl    686.8
2. 1998  NYY over Tex    679.4
1. 2005  StL over SD     640.6

All three-game sweeps. Only one of the advancing five teams would make it to the Series, that being the short-list “Greatest of All Time” contender ’98 Yankees. The least exciting four-game LCS was the 2004 Cardinals-Dodgers series; the dullest five-game set was—no, not the 2002 Braves and Giants, despite the two horrid games listed above—the 2001 Mariners-Indians series. (Their Game Two finished just off the worst game list at #7.)

We can do better than this. Heck, seven single games above did better than any of these series. Let’s see what the cream of the LDS crop provides. Again, the lower half of the top 10 gets just the list and link, while the upper half gets a little more.

    Year     Teams      WPS Score  
10. 1981  NYY over Mil    1737.4  (five games)
9.  1996  NYY over Tex    1751.9  (four games)
8.  2009  Phi over Col    1780.3  (four games)
7.  2004  NYY over Tex    1807.9  (four games)
6.  2003  Flo over SF     1877.9  (four games)

It’s surprising how many four-game series made the list: six out of 10 overall. Four-game LDS are more common than five-gamers, by 24 to 20, but this goes well beyond that. We’ll talk about this matter later. The best three-and-out in the LDS was Cleveland-Boston in 1995. Yes, 1995 again.

No. 5. Red Sox vs. Angels, 2008 (four games)—1993.8 points
338.6 / 500.6 / 680.2 / 474.4

How does a four-game series beat most five-games series in a quantitative measure of excitement? By setting “above average” at its absolute floor, and heading up from there. Three of the games were tied through eight innings, the other a one-run game at that point. Throw in a 12-inning contest and a walk-off win to clinch the series, and you end up not pining for Game Five.

No. 4. Astros vs. Braves, 2004 (five games)—1998.3 points
242.6 / 611.9 / 294.2 / 621.4 / 228.2

Next question: How can a five-game series chart high on this list with three mediocre games? The other two need to be great, and they were. Atlanta’s Game Two comeback included a 100+ WPS eighth to tie, and a Rafael Furcal walk-off four-bagger in the 11th to win. They did it again in Game Four, a three-run rally knotting the game in the sixth, and Furcal again involved in the go-ahead tally, this time scoring it on J.D. Drew’s single. Houston’s victories lacked such drama, but there were three of them, and the Braves’ comebacks went for nothing—except a great WPS score.

No. 3. Braves vs. Rockies, 1995 (four games)—2122.7 points
706.6 / 529.2 / 601.0 / 285.9

Yes, 1995 again. (And spoiler alert: It’s gonna happen again.) The first three games were super back-and-forth tussles, noteworthy in that each time it was the visiting team that pulled ahead to stay in the final inning. WPS is biased toward walk-off home team wins (apt, as fans are pretty biased toward them too), so these games racked up big numbers the hard way. Too bad Game Four ended up a Braves rout, a deflating end to what had been an incredible series. Still, it would have been tough to climb higher up this list: The margins at the top are awfully big.

No. 2. Red Sox vs. A’s, 2003 (five games)—2431.9 points
815.6 / 166.3 / 510.2 / 475.8 / 464.0

To complete an up-page thought, Billy Beane’s you-know-what generally works in the playoffs for two games. In this case, the first two. Boston came all the way back, but nothing was easy about it. An extra-inning walk-off win plus two one-run contests added up to three very tight, very good games. That plus the Game One classic puts this series squarely in second—though a distant second.

No. 1. Mariners vs. Yankees, 1995 (five games)—2770 points
407.7 / 927.3 / 259.6 / 452.7 / 722.7

This was close to a runaway, well ahead of all the other series. Game One was good, see-sawing until New York broke it open in the seventh, although Seattle did bring the tying run up twice in the ninth. Game Two, listed above, was awesome, and put the Bombers up 2-0. Game Three was the only fizzle, a bunch of 1-2-3 innings intermixed with Seattle eruptions in the fifth and sixth, Tino Martinez driving both. (No wonder the Yankees traded for him as soon as they could.)

Game Four was a slugfest. Seattle fell behind 5-0, hauled itself ahead with a four-run third and a pair of pickets afterward, then refuted the Yankees’ rally to tie with a five-run eighth. New York got Bernie Williams up in the ninth as the tying run, but he flied out to insure Game Five.

We’ve mentioned that Game Five already. Down 4-2 in the home eighth, Seattle pulled even with a Griffey longball and the patience to let a wilting David Cone yield a bases-loaded walk. Both teams mustered failed rallies in the ninth; Seattle threatened again in the 10th, but again couldn’t convert. New York’s walk-sac-single rally put the Yanks ahead in the 11th, but Seattle didn’t use even one out in replying, Edgar Martinez’s two-run double being game, set, and match.

Boy, they were really working to get disgruntled fans back that year, weren’t they?

Modification: if all games are not created equal

The indices for these postseason series, and the games themselves, are based on the assumption that each game builds upon an equal base of excitement potential, that Game One starts off even with Game Five. That’s not really the case, though. The games aren’t played in isolation. The teams’ advance toward a World Series title adds a level of excitement, and a system that purports to measure the excitement of baseball games is not complete if it cannot take this into account.

I’m not nearly the first baseball writer to make this observation. My earlier WPS articles mentioned Sky Andrecheck’s Championship Leverage Index, which melds the excitement of a game to the influence it has on deciding a champ. I said then that I might do something similar once October came around—and here it is.

I will measure the importance of a game in a playoff series by how much it moves the likelihood of a team winning that series, very much akin to what Win Percentage Added does within games. Multiplying that by the WPS Index of each game produces what I will call the Series Percentage Sum, or SPS. I’ll use that to re-examine series, and individual games, to show which are the most exciting in the greater context.

I’m not going to get too ambitious with this modified system. I’m going to compare apples to apples, LDS to LDS today, and won’t stretch it to compare different playoff rounds with each other. I could try, but any easy calculation along those lines would hang on the assumption that the only exciting thing about a baseball game is how influential it is in deciding a champion.

I don’t believe that. I believe there is some inherent value in a baseball game just for being a baseball game. How high that excitement value is, compared to the excitement of the title chase, I’m not going to try to calculate today. I’m not sure it ever can be, subjective as the matter is. Ten different fans would give 10 different answers, the casual ones leaning toward the playoffs, the more dedicated ones appreciating everyday games more.

There’s a real philosophical question here, one that deserves a fuller airing than this article can provide. Another day.

For now, my focus is the LDS, and this is the system I’m using. Games One and Two have equal values of .1875, meaning the winner’s chance of series victory rises by 18.75 percentage points. Game Three of a 2-0 series rates at .125; in a 1-1 series, it comes out as .25, as will any Game Four. Game Five, a coin flip for the whole ball of wax, naturally comes out at .5.

It’s fairly obvious from those numbers alone that the new method will heavily favor long series, especially those whose Game Fives are nail-biters. The theory works out in practice, as almost all the four-game sets are swept out of the top 10. Here is that new list, though not entirely new, especially at its apex.

    Year     Teams     SPS Score  Games  Prev. Rank (out of 72)
10. 2003  Flo over SF    432.25     4        6th
9.  2002  Min over Oak   439.9      5       27th
8.  2005  LAA over NYY   448.8      5       20th
7.  2011  Det over NYY   453.2      5       22nd
6.  2011  Mil over Ari   470.5      5       26th
5.  1997  Cle over NYY   473.8      5       11th
4.  2001  Ari over StL   502.4      5       15th
3.  2004  Hou over Atl   503.2      5        4th
2.  2003  Bos over Oak   598.8      5        2nd
1.  1995  Sea over NYY   757.3      5        1st

The Mariners and Yankees lap the field, as far ahead of second place as second is ahead of ninth. The 1995 Braves-Rockies series, previously No. 3, gets dumped all the way to 21st, partly from Atlanta winning the first two games, partly from Game Four (with the highest multiplier) being the only dog of the series as well as the last game. Two of last year’s Division Series make the list, and my eyeball test tells me they’re pretty worthy.

And if there’s one other formula for an exciting League Division Series, it appears to be “beat the Yankees in five.” (Gee, I didn’t know Excel spreadsheets were invented in Boston.)

We can apply the SPS system to individual games as well. The resulting top 10 list is a mix of the familiar and the new, with something very familiar still at No. 1. It is also as top-heavy with Game Fives as you would expect.

    Year/Game     Teams     Inn. SPS Score  Prev. Rank (out of 280)
10. 2002 Gm.5  Min over Oak   9    184.1       90th
9.  2000 Gm.3  NYM over SF   13    194.9        6th
8.  1999 Gm.5  Bos over Cle   9    196.5       82nd
7.  2003 Gm.3  Flo over SF   11    198.2        5th
6.  2002 Gm.5  SF  over Atl   9    198.4       78th
5.  2005 Gm.4  Hou over Atl  18    220.3        3rd
3t. 2001 Gm.5  Ari over StL   9    232       t-47th
3t. 2003 Gm.5  Bos over Oak   9    232       t-47th       
2.  2011 Gm.5  Mil over Ari  10    276.6       25th
1.  1995 Gm.5  Sea over NYY  11    361.35       8th

This new list shows a weakness of the SPS method: there have been only 20 LDS Game Fives, and almost half of them are on the list. A number of merely good games get elevated to greatness simply because they’re Game Fives. On the other hand, if you think a game shouldn’t have to go into extra innings to be considered highly exciting, this list is for you, with most of the newcomers finishing in nine.

The best of the new games was last year’s Milwaukee-Arizona clincher, with Carlos Gomez’s steal and Nyjer Morgan’s single winning the day for the Brew Crew. It gets my highly subjective thumbs-up as a great game, and with a WPS over 550, it gets that mildly objective imprimatur as well. (If I have to draw a line, I’d say greatness starts at 500.) The other newcomers, I wouldn’t swear to as “great.”

The new method still gives the 1995 Mariners-Yankees series the top slot, but this time it’s Game Five instead of Game Two, which drops to 15th. I can only imagine if the earlier contest had somehow been transposed to Game Five: There would be a flashing “Tilt” somewhere on my spreadsheets.

After this first pass with the SPS system, I confess myself not to be overwhelmed. The multipliers feel weighted too heavily toward the final game, no matter what the probabilities say. Maybe that intrinsic value of a game as a game, or a playoff game as a playoff game, would even things out. Maybe I should move part of the Game Five credit back to Game Four’s result, for making Game Five possible. (The most famous World Series Game Sixes—1975, 1986, 2011—all led to Game Sevens, which doesn’t feel coincidental.)

Series Percentage Sum is, at best, a work in progress. Perhaps it functions better encompassing a whole series rather than individual games, or maybe I just need better multipliers. We’ll see SPS again, but I’m not throwing out WPS in its favor.

Things to come

So we close this chapter of the WPS post-season review, with the year 1995 chiming in our heads like a particularly tenacious earworm. Next week, it’s the League Championship Series—but WPS will be back before then.

This series on postseason WPS came to mind because 2012’s playoffs are upon us, and instead of looking backward, I’m looking ahead as well. For the duration of the playoffs (and for any tiebreakers, which are still possible in a couple scenarios as I wrap up writing on Tuesday morning), I will be doing daily THT Live posts giving WPS breakdowns of the games, along with any other commentary that strikes my fancy.

Up until now, I’ve used Baseball-Reference as my resource for WPA numbers, but October baseball shifts things into a higher gear. B-R takes until the following morning to post the relevant data for the day’s games, and I need my analyses ready for you bright and early. Fortunately, FanGraphs gives play-by-plays in real time, including Win Percentage Added. This means I can finish crunching the numbers in time for our regular morning readers.

Call it corporate synergy, if you like. Call it borderline OCD, if you must: That’s probably not unknown in the field of baseball analytics. Heck, call it Melvin if that’ll get you reading. The point is, the WPS Recap will be here for you starting Saturday morning, or even earlier if the baseball deities throw Team Chaos a bone. (Some people were wishing so hard for a five-way tie, it seems cruel to deny them completely.)

P.S. Readers of the future, your links for the rest of this series are here for Part Two and here for Part Three. Readers of the present, don’t click those until I’ve actually written them, please.

References & Resources
Baseball-Reference was my source for all historical WPA data.

A writer for The Hardball Times, Shane has been writing about baseball and science fiction since 1997. His stories have been translated into French, Russian and Japanese, and he was nominated for the 2002 Hugo Award.
Newest Most Voted
Inline Feedbacks
View all comments
Marc Schneider
10 years ago

No offense and I generally agree with statistical analysis, but it seems to me that quantifying what is an “exciting” game is like quantifying who is a beautiful woman.  Do you really need statistics to do this and even if you can do it, why?

10 years ago

Another game that obviously seems to have slipped through the cracks is last year’s NLDS clincher by Chris Carpenter over Roy Halladay. Don’t even try to tell me that wasn’t an exciting game (or series for that matter). No lead changes, but holding onto a 1-0 lead for 9 innings is INTENSE.

Shane Tourtellotte
10 years ago

Marc:  One reason why is that it helps us recall games that might have slipped our memory, or introduces us to games we never knew about.  1995 has been such a revelation to me.

Todd:  Game 5 of Cards-Phils comes in at 553.2 points, 25th of the 280 LDS games played.  It’s legitimately great, but doesn’t crack the highest echelon because, frankly, Carpenter was just too good.  Exactly two Phillies got as far as second against him, and even a 1-0 game loses something when the trailing team almost never threatens to score.  If Roy hadn’t had that rocky start, though … whew.

10 years ago

The 2004 Texas Rangers fell short of the playoffs so I don’t see how they could have been in one of the most exciting series.

Also maybe when you weighting games in a series it looks like you’re giving more weight (.25) to a 1-1 series game than a 2-0 series game (.125). Seems like any potential elimination game should be weighted higher, and maybe a bonus for games in which a team staves off elimination?

Shane Tourtellotte
10 years ago

Oops.  You caught one.  The #7 Division Series should be the 2004 Yankees over the Minnesota Twins.

Technically, the bonus for staving off elimination is that you get to play another game that adds to the numbers.  It’s already counted:  the problem is whether some of it should count in another place.

10 years ago

This shows what any Mariners fan could tell you – 1995 was mind-blowing.  First, the Mariners had the greatest come-from-behind regular season comeback in baseball history.  then, they played in the greatest post-season series of all time.  Then … they ran into the Indians juggernaut, man that team was good.

All for a franchise that had never been to the post-season, only ever had two winning seasons (barely) in their entire history, and usually sucked rocks.  For big baseball fans like myself, we had never experienced anything remotely like it, and the entire town went nuts.

10 years ago

Hey Shane, didn’t know if you ever saw this one:


It’s my ranking of the tensest World Series ever.