How Game Seven Saved the 2014 World Series
The 2014 World Series was an enjoyable contest. Not only did the Royals provide a fresh face and a different type of team than we’ve generally seen in the Fall Classic, but Madison Bumgarner gave us a performance for the ages. And, for the first time since the two Wild Card games, we actually got a winner-take-all contest; the first two rounds of the 2014 playoffs provided little in the way of drama about the outcomes of the series, as the winners combined for a 20-3 record in the division and league championship series. But, while the World Series gave us that elusive Game Seven, we also have to acknowledge that series had one of the lowest totals of in-game drama of any World Series in history.
One of the neat things about Leverage Index, besides giving a numerical representation of important situations, is that we can look at the average leverage index for an entire game and get a feel for how dramatic the game was as it went on. In a close contest where the lead is regularly changing, the average leverage index can push near two, as it did in Game Two of the NLCS; the aLI for that game was 1.94, meaning that each play carried something close to double the weight that it would have had in an average game. The 18 inning contest between the Giants and Nationals had an aLI of 1.81, so that was basically two full baseball games worth of drama similar to what you’d find protecting a ninth inning lead.
The World Series didn’t have any games like that. In fact, until the final game, it didn’t include a contest where the aLI was even over 1.0. The first six games were mostly blowouts, with the exception of game three, the one in which the Royals took a 1-0 lead in the first inning, built it up to 3-0, and then held on to win 3-2. Besides that we, had 7-1, 7-2, 11-4, 5-0, and 10-0 contests; the aLI of game six was a paltry 0.25, the third lowest number for a game in World Series history.
So, where does this series rank in terms of average leverage index for all of the games combined? Let’s take a look.
| Year | aLI |
|---|---|
| 1989 | 0.58 |
| 1968 | 0.63 |
| 1937 | 0.69 |
| 1960 | 0.72 |
| 2014 | 0.73 |
| 1977 | 0.74 |
| 1965 | 0.74 |
| 1903 | 0.76 |
| 1945 | 0.78 |
Even with a competitive and tension-filled final game, the aLI for this World Series was the fifth lowest in history; without that game, it would have been the third lowest.
1989, of course, is mostly remembered as the “Earthquake Series”, as a quake prior to game three forced a significant delay, and the World Series got stretched out over two weeks. The A’s won all four games, with the closest being a 9-6 victory in the Game Four finale. The highest aLI for any of the four games was .902, with the other three all falling at .50 or under. Not only did the A’s sweep the Giants, but they did so in about as boring a way as one can complete a sweep.
1968 was a bit more similar to 2014, as it went the full seven games, but included contests like a 13-1 drubbing in Game Six that was 12-0 by the end of the third inning. Like this year, though the final game provided some intrigue, as it was scoreless into the seventh inning before the Tigers pulled away and eventually won 4-1.
But besides that, nearly every other World Series in baseball history has provided as much or in-game drama as the 2014 version, and before the finale, those would have been the only two match-ups to produce a lower aLI than this year’s World Series. Thankfully, though, we got very good final game, and Madison Bumgarner’s performance single-handedly cemented this series place in history.
If you’re wondering about the flip side of things, here are the 10 most competitive World Series match-ups by aLI.
| Year | aLI |
|---|---|
| 2005 | 1.49 |
| 1991 | 1.41 |
| 1980 | 1.38 |
| 1924 | 1.37 |
| 1941 | 1.35 |
| 1915 | 1.33 |
| 1995 | 1.30 |
| 1992 | 1.30 |
| 1975 | 1.29 |
Yep, the 2005 World Series that saw the White Sox sweep the Astros was the most competitive game set of contests in history. The White Sox won every game, but they won 5-3, 7-6 (9th inning walkoff), 7-5 (in 14 innings), and 1-0 in the finale. Like the early rounds of this year, the games themselves were terrific, even if the outcome seems lopsided based on the number of games won by each side.
This year provided a pretty stark contrast between two extremes; great games resulting in short series, and then mediocre games ending with a tense final contest. Either way, though, it was an enjoyable postseason.
*Thanks to the tremendously useful Play Index at Baseball-Reference for the data.*
Dave is the Managing Editor of FanGraphs.
Interesting that three of the top ten most competitive World Series match-ups involved the Braves during the height of their run with Glavine, Smoltz, Maddux (only there for ’95 surprised me) and the rest. They may have only won it once, but they sure did give us some exciting baseball to watch.
The only way it could’ve been better, of course, is that we didn’t have an aLI of 0.00 in 1994.
Agreed that no-WS 1994 sucked, but if it had been played the Braves would likely not have been a part of it, as they trailed the Expos handily when the season was cancelled. (Not to take anything away from their otherwise stellar run in the 90’s.)
You are right that the Expos led the Braves by 6 games when the strike occurred, but the Braves also led the Wild Card race by 3 games over the Astros.
There was a serious chance that WS would have been between the Expos and the Blue Jays. (Montreal had a terrific team and were running away with the NL regular season, tho Toronto was just more in the hunt.). No way of even beginning to guess how competitive such a WS would have been, but the big talk at the time was about TV ratings for it. Expo fans and writers still feel the MLB owners deliberately picked that season to lock out the MLBPA due to the prospect of a WS that was doomed to flop with American TV.
This would be a great theory if there had been a lockout and not a strike in 1994.
A rose by any other name …
Sometimes what’s called a “strike” is in fact a lock-out.
The MLB owners wanted a salary cap and a sole spokesperson in collective bargaining, being the commissioner. Four problems, in descending order of significance:
1. there was already a collective bargaining agreement in place;
2. factor 1 ORDINARILY would mean the NLRB would step in to enforce the CLB, but … baseball, and rich owners, and their connections to Congress;
3. historically of office of commissioner was located in some largely undefined space BETWEEN the owners and players; IOW this would put an end to the pretense that led to the owners picking the likes of Bart Giamatti, Faye Vincent or even Bowie Kuhn and expecting them to understand their real role was to look neutral while screwing the players; and
4. the entire basis for the owners’ complaint for WHY then needed a salary cap was unsupported by evidence and largely fraudulent and otherwise a blizzard of accounting inventions (I imagine Dave Cameron could explain this way better than I can.).
I’m going to hold onto the little bit of hope that you actually understand all this better than your comment suggests. Feel free to disabuse me of this generous concession.
This would be a great theory if the Jays had been good in 1994. They finished 55-60 and were 12 games out of the wild card and 16 games out in the division. There was zero chance they would play the Expos in the WS.
I’m thinking it probably happened before you started watching MLB. The BJs were a major power in those days, having won back-to-back WS just a few years before, and were quite competitive thru the playing of that dramatically abbreviated “season”. Indeed, it was SO shortened, it meant each game took on something like the inflated (not artificially so, but inflated nonetheless) leverage of a post-season game. A team had to be really, really bad to be considered ‘out of the hunt’ that season, and notwithstanding any impression from the final standings, the BJs were NOT really, really bad.
There’s also this tendency, which I’m not identifying especially with younger fans because it’s found all over the place, of looking back at past events and thinking they were somehow ‘justified’ by realities that were obscured by the them unfolding in real time, or otherwise ‘fated’ to turn out that way.
Yet, you and I are communicating here at FanGraphs, which, if it stands for anything, stands for the proposition that chance plays an vastly under-appreciated and misunderstood role in producing results in PAs, innings, games, weeks, months, seasons and even groups of seasons. That each of us IS here, communicating, suggests at least some receptivity to that philosophy (It’s certainly possible that you’re actually just a troll — they’re known to exist online! — but I’m granting you the assumption that you’re not trolling, but instead you just weren’t watching [Cue old-timey banjo picking music.] … in the PAST … when the 1994 season actually happened, and thus have failed to account sufficiently for the possibility that the past, like the present, as well as the future, rolled out (rolls out, will roll out) with a lot of chance and human choices (both misconceived and prudent) involved.
Thus, it may SEEM obvious in retrospect, but it sure didn’t seem so obvious at the time, and for what were considered at the time quite sound reasons. Two consecutive World Series championships and the core of the team that produced them still being around are not nothings.
I forgot to add that some non-trivial part of MY perspective on that season, not just now but at the time, comes and came from having been an avid real-time watcher of the 1964 National League pennant race.
Check it out: the Phillies were 7 games ahead of the field with 10 days left to the end of the season. When the actual season had actually ended and the dust had actually settled, the Giants had finished just 2 games back of the Phillies, the Reds had caught them, and the Cardinals had not just caught but PASSED them, to actually win the actual pennant.
Being of a certain age, that winter I played out that season on APBA and Strat-O-Matic cards (No PCs in those days.), and found the Reds winning on APBA, and the Phlllies winning on Strat-O-Matic. Friends who did the same told me of similar outcomes, PLUS of the Giants winning, and even the Braves, by far the most awesome hitting team of that season. As I grew older and put away childish things, the only remnant I kept from those APBA/Strat-O-Matic days wwere the player cards from that 1964 season. David Halberstam, an excellent non-insider reporter and writer, wrote a superb book that generally covers that season, called “October 1964”. Read and learn.
Toronto trailed Cleveland by 12 games in the WC race at the time of the strike, with Baltimore and KC also between them and a playoff spot. Pointing out extreme outlier comebacks to try to justify that the Jays had a “serious chance” to return to the playoffs is intellectually dishonest.
The early 90’s Blue Jays were not. The 1994 team, sans several major contributors from 1993 (Fernandez, Henderson, and Ward) and with others way off their previous level of performance (Olerud, Alomar, White, Guzman, Stewart) were a group on the decline. That’s why they were 55-60 and that’s why the club was noncompetitive over the next two seasons.
Would be interested in what the Royals’ aLI was over the WC game, DS, and CS. Probably something similar to the 2005 WS.
just curious, what about ranking WS by TOTAL LI (tLI?), summing together the LI of each game played… would give more weight to 6 & 7 game series that may have had 1 dud and multiple great games.
How did you find all this data so quickly with the Play Index? Looking over the Index, I’m not sure how you found all the World Series data so quickly. I was actually putting together a spreadsheet showing average aLI for the World Series’ myself (more for personal curiosity than anything else). Long story short, knowing how to work the Play Index would be a huge time saver.
This is a great article nonetheless, and I’m glad someone else discovered how incredibly entertaining the 2005 World Series secretly was.
It was a little tricky. I used the game finder, then the filters to limit to WS games only, and then only display the home team so I’d only get one record per game. That gave me the ability to see all 641 WS games, including their aLI. I exported that to Excel, did a text-to-columns function to remove the year from the date column, then sorted/subtotaled by year.
A measurement that rates a 4-game sweep as far more dramatic than a 7-game series is flawed. To properly measure the level of drama in a series, you need to: (1) multiply the leverage of each game by the leverage of the game in the series, and (2) express it as a sum rather than an average.
I explicitly said this measure was looking at the most competitive series, not the most dramatic. Doing what you suggest would be of interest as well, but it wasn’t the point of the post.
Yes, your post was very clear and I didn’t mean to suggest it was misleading or wrong. I just don’t think average in-game drama is a very good proxy for the overall level of drama of a series.
It appears that LI aggregates in-game occurences for single games, but does not really quantify the urgency a team would feel to win said games, such as elimination games, or game 7.
Anti-Fan Graph anecdotes but a lot of the true drama of a post season race is the manic-depressive time between the games.
Supporting a Game 2 winner feels like you are on top of the world for the off-day until Game 3 punches you in the gut and you are certain all is lost. A close victory in game 5 gives you hope over the off-day but you are filled with the ‘what-if’-we-don’t-win-Game-6′ dread
This series was somewhat like the 1960 series (ranked right beside it) in that game 7 was the big savior of an otherwise uneven series. Then again 1960 had 5 Hall of Famers involved and much better teams overall.
Yeah, the 1960 and 1968 World Series stood out to me too, as 2 classic 7-gamers. I think the fact that the Pirates got blown out 3 times by that Yankees juggernaut heightened the drama of their eventual victory.
Memories
From the corners of our minds …
“much better teams overall”.
Um …
That Yankees team was mostly a blend of vin tres ordinares (including the preposterously over-rated Bobby Richardson, an icon for his birthstate, no mere slouch but MLB’s answer to who’s the greatest ever Hanger On), made to appear vintage due to the blinding presence of Mantle, Maris at his best (better even than 1961), and the Ghost of Clemente.
I’ve gone thru this exercise; you should too, anyone can. It takes a bit time with Baseball_Ref, but it’s instructive. Go thru one of those dubious position-by-position analyses the MSM does at Series time, and you’ll find the 2 current teams are at much the same level as those two 1960 teams.
Examples:
Gordon’s by far the best at LF.
Great as Maris was that year, LEGITIMATE no slouches are Pence and Clemente, and Pence and the Clemente of THAT year at least are basically a wash. Okay, KC’s not so hot.
Best catcher is Posey, mask-down; Perez arguably second (Berra was on his downhill slide.
[Fun Fact: 37 Perez’ in MLB history, yet one Bumgarner, Fidrych, Moustakas, Pence, etc]
Shortstop clearly goes to Dick Groat, but it’s really hard to see how today a guy who finished no better than 8th in WAR would win his league’s MVP, and Crawford, Infante and Kubek are indistinguishable.
[Jeopardy entries for category Oh Those Baseballing Crawfords:
1. A. Sam Q. Who was the best Crawford?
2. A. Brandon Q. Which Crawford did the best Tommy Lee Jones?]
Hoak had his Best Evuh year at 3B, but a better player than Panda? Not. Yankees had the Lesser Boyer at third (tho not the Least Boyer, that was Cloyd). Moose, OK, feckless here; best nickname?
You think Maz at second = Contest Over; not so! Do the math. Skowron had a strong year, but I’d still take Hosmer OR a healthy Belt over Diplodocus Moose the Foorst; and any of them over that cheap laffs of a Dr. Strangeglove the Pirates had at first.
All 4 pitching staffs were pretty okay, but none dominated their leagues, and otherwise, since the dramatic change in emphasis towards the bullpen, it’s tough to fine tune distinctions … other than, of coure, Bumgarner.
We don’t actually KNOW how many HoFers will come out of this WS. Right now, SURE there were 5 from THE PAST, and there are zero NOW, but d’uh, that’s all for in [Cue eerie music.] the Future. For sure Bochy, likely one maybe both of the firm of Bumgarner and Posey; beyond that, I expect Dave Cameron to assign to Cistuli for the general amusement of we patrons.
Anyway, I dispute the “way”, and to some extent even the “better”. I WATCHED that WS, in real time, and there’s simply no comparing how freaking PONDEROUS the typical player of that era was.
I stop reading Fangraphs for a few weeks and I come back to this act? I see you play both the role of Abbott and Costello. Stop engaging in revisionism–unlike other aspects of history recent baseball history is quite well documented.
Oh, you hear that? It’s a train whistle. The Nonsense Train is about to leave the station. Next stop is any other baseball blog. Au Revoir.
Obviously a brain-slip, but the Royals won game 3 in exactly the way you describe, not the Giants
oops, fixed some time between me opening this article and reading this article. carry on….
It’s funny seeing 2005 as the top considering how “boring” the teams were.
Has it perhaps occurred to anyone, that the biggest thing wrong with this WS was Fox.
Agreed, Fox was pretty bad.
Come on. Don’t be shy. Let’s just get it out there. At the center of the heart of Fox’s failure to announce a baseball game is Joe Buck.
I think you mean FAUX, friend! hawhawhaw
Warmest regards,
Obummer
Fox has me baffled.
The pay millions for the rights and then stick the NLCS on a channel no one knows or cares about and then feel the best way to market it is to tell viewers they are idiots for not knowing about FoxSports1
For the World Series they seem to have a pre-written narrative for each game and stick to it no matter what happens on the field that night. I’m sure they have done some market research that states the ‘average’ World Series viewer wants to hear this but I end up watching Fox
Can you somehow convert game win expectancy to series win expectancy? Such a conversion would make high leverage moments in low leverage games less meaningful, like game 4 when up 3-0.
I’d love to see a version of this where the aLI is weighted by some measure of series leverage. Even if you used a coin flip model to get series leverage it would still be very informative. For instance, it’s likely that 2005 would be bumped off the list, while this year would probably move up quite a bit. It would much better reflect the actual suspense in the games.
I know I’m a little late, but I was curious about the interaction between great games and great Series as well (2005 was a lousy Series with great games, 1960 was actually kind of lame because of all the blowouts), so I did some quick-and-dirty math combining the aLI of each World Series game with Sky Andrecheck’s Championship LI (thanks as always to the Play Index at B-R). I wrote a blog post about it here (http://excessivelylogical.blogspot.com/2014/10/what-were-best-and-worst-world-series.html), and there’s a link there to the full dataset.
Good job not providing the LI for game 7. Not like anyone would’ve been interested in that…
1.34
One thing about the 89 World Series that made it even more lopsided was the week layoff between Game 2 and Game 3.
It allowed the A’s vastly superior 1 and 2 starters to pitch in all 4 games. Dave Stewart v Scott Garrelts and Mike Moore v Reuschel and Robinson
The 89 A’s were still a much better team overall than the San Francisco team but the earthquake certainly didn’t do the Giants any favors
Your point about competitive vs dramatic is well-taken, and I’m surprised there’s so many complaints about it. But I’m also surprised no one has pointed out the other, more-relevant-to-your-point issue: it’s easier to maintain a high average LI over 4 games than over 7. I’m not sure how to adjust for that; maybe you don’t, and you just group each series by the number of games played.