Network News

X My Profile
View More Activity
The new Washington
Post Weather website
Jump to CWG's
Latest Full Forecast
Outside now? Radar, temps
and more: Weather Wall
Follow us on Twitter (@capitalweather) and become a fan on Facebook
Posted at 10:45 AM ET, 02/ 4/2009

Backlash Hits AccuWeather After Lackluster Storm

By Andrew Freedman

* Cold for Now, Warm-up for the Weekend: Full Forecast *

One of the strongest winter storms in recent memory lashed the east coast yesterday with strong winds, heavy rain, and paralyzing snow.

Did you miss it?

I suppose you could be forgiven for not noticing it, since what may have turned out to be the blizzard of 2009 ultimately existed solely in computer model simulations of the weather. Despite their best efforts, the models with impersonal acronyms like "GFS" and "ECMWF" could not force the storm scenario they had predicted to come to fruition. A dusting of snow fell around Washington, and a few inches in New York and Boston, but that was far from the one to two feet that was projected at one point.

Keep reading for more on AccuWeather's hype-filled forecast, and the fallout...

The non-storm event provided an interesting case study in how communicators of weather information should balance the need to attract eyeballs, in the form of viewers and readers, with the uncertainties inherent in forecasting that can make even the best forecast go bust. Remarkably, despite an early consensus of the computer models that a major storm was likely on Groundhog Day (that consensus had crumbled by last weekend), most media outlets -- including the Capital Weather Gang -- limited their hype to a Category 2 situation rather than a Cat 5 "hype-cane."

Although there were probably others who leaned too heavily towards hyping the storm, the most notable example of a media organization that balanced the ratings/uncertainty equation in favor of hyping the storm beyond what was meteorologically justified was AccuWeather Inc., the Pennsylvania-based private forecasting company that has a reputation in the weather forecasting industry as frequently being bullish on predicted snowfall amounts.

On its Web site and in press releases, AccuWeather featured blaring headlines of a major storm that left little wiggle room for the company's forecasters to back away from their predictions as it became clear that the storm was not going to materialize as expected. For example, on Jan. 30, AccuWeather distributed a press release that warned of the likelihood of not just one "monster storm," but a string of them in the next few weeks.

Regarding the Groundhog Day event, the press release stated: "The storm coming Monday and Tuesday next week will track from the Gulf of Mexico to a position near Baltimore by Tuesday morning. The storm will take shape on Sunday, spreading rain over Louisiana and neighboring states. The storm will then race through the South on Monday (Groundhog Day) to reach the Northeast by Tuesday."

The use of the word "will," rather than "could" or "may," erroneously conveyed the message that the forecast was a sure bet. In that press release, AccuWeather did mention that there was uncertainty in the track of the storm, but it did not highlight the possibility that the storm would fail to get its act together until it was too far offshore to have much of an impact on the coastline.

Not surprisingly, viewers of AccuWeather's video blogs and visitors to its Web site felt cheated when the monster storm was a no-show. Sadly, these disappointed snow lovers took their anger out on individual forecasters, leaving one prominent forecaster at the company reeling from the backlash.

Henry Margusity, the colorful host of the "MeteoMadness" video blog for, declared that he would cease "vlogging" due to the emotions that it seemed to incite (welcome to the club, Mr. Margusity!). He quickly recanted that declaration after it generated a flood of supportive emails and comments, as well as a request from company management to continue such work.

"Look, we all get frustrated over the weather. I get frustrated. I probably have not slept in 3 days over this storm. The models were horrible, for reasons I cannot explain. It happens. But what I would prefer are people who look at the weather offer their opinions in a professional manner," Margusity wrote on Monday.

The fallout was so ugly that Margusity's wife had to urge readers/viewers to calm down in a special guest blog post:

...anger, disgust, and threatening statements were made, and by more than just a few. You don't have to agree with all of Henry's ideas; certainly I do not! Folks, express your disagreement politely, or let it go...

Before prospects of an epic storm fizzled, Margusity had been referring to it as a "Big Daddy" storm, but unfortunately for him and for snow lovers up and down the eastern seaboard, it turned out to be more of a delinquent daddy who left a few bucks on the kitchen table, ran out the back door and never looked back.

The swift and nasty backlash against AccuWeather for its several days of excessive hype illustrates the perils when forecasters don't clearly communicate the uncertainties of their forecasts. In Margusity's case, angry viewers felt he oversold the storm.

Despite AccuWeather's hyped storm coverage, this was an extremely difficult storm to forecast. Everyone made mistakes, including those of us at Capital Weather Gang. In the next few days we'll feature a post explaining why the computer models had such a hard time handling this storm in particular, which may be the subject of several meteorology Ph.D. dissertations in the next few years.

In the meantime, we're interested to hear your views on what you'd like to see forecasters do to more effectively convey the degree of uncertainty involved in a given prediction.

By Andrew Freedman  | February 4, 2009; 10:45 AM ET
Categories:  Freedman, Winter Storms  
Save & Share:  Send E-mail   Facebook   Twitter   Digg   Yahoo Buzz   StumbleUpon   Technorati   Google Buzz   Previous: Forecast: Too Cold? Just Wait for the Weekend
Next: PM Update: Cold Winds With Snow Showers


Live by models, die by models...

Posted by: steske | February 4, 2009 11:03 AM | Report abuse

I don't feel sorry for AccuWeather one little bit. They overhype EVERYTHING. They are the weather equivalent of the worst tabloids around. No wonder they don't have much cred.

Posted by: Wallpass | February 4, 2009 11:18 AM | Report abuse

One fault with the CWG predictions (and you guys do a pretty good job) was that precipitation estimates all implied snow accumulation. IIRC it was 35%

Posted by: RMVA | February 4, 2009 11:21 AM | Report abuse

my message was somehow truncated. you gus gave odds of 35% of

Posted by: RMVA | February 4, 2009 11:25 AM | Report abuse

it happened again. web server problems, seemingly on your end. you guys gave 35% odds of less than 1 inch of snow, but no odds for the complete bust - zero accumulation - we got in Alexandria.

Posted by: RMVA | February 4, 2009 11:28 AM | Report abuse

Accu-weather is predicting a 95 degrees and thundersorms for Sunday.

Posted by: oldtimehockey | February 4, 2009 11:47 AM | Report abuse

RMVA: If you read the comments on a couple of these threads, you would have seen some people (well, okay, me) predict 0 accumulation. I pointed out that the weather temperature predictions preceding the storm were WAY off, consistently off and consistently below what the high/low temperature actually was. That was a tell-tale sign that the models were probably junk and we would get more rain than anything and not even much of that.

In the CWG's defense, they did post in the comments section that the reason they didn't produce an accumulation regional map (showing regions of the area and how much snow they would be getting) was because they didn't think it was likely we would get over an inch. And comparing that with the early predictions... it was VERY clear which direction the momentum of this storm was going. It was fizzling out... kinda like this whole winter so far.

Posted by: prokaryote | February 4, 2009 12:04 PM | Report abuse

I think the measured, conservative tone one finds in the official NWS forecasts are the way to go. Their three-day forecasts are usually pretty accurate, and they issue watches or warnings as needed 24 hours or so in advance. Beyond three days, they couch their long-range forcecast discussions with enough equivocation to avoid any premature hype. Very rarely, such as in the March '93 "Storm of the Century," do the models come to an agreement on a a big storm more than five days out. With winters around here, you pretty much have to preface every forecast of a snowstorm with "Depending on the exact track of the storm and the amount of cold air in place ..."

Posted by: jmbethesda | February 4, 2009 12:04 PM | Report abuse

RMVA said: "you guys gave 35% odds of less than 1 inch of snow, but no odds for the complete bust - zero accumulation - we got in Alexandria."

Actually, CWG's mets' posts were for 70% chance of 0-3"

That call nailed it for the area.

Posted by: Terp4Life2 | February 4, 2009 12:25 PM | Report abuse

As much as I feel bad for Henry over at Accuweather, I think he probably just needs to grow up. Seriously, its the internet, and people write all sorts of terrible stuff because of the anonymity. Then to have his wife writing on his blog about how his feelings were hurt strikes me as terribly lame... That being said, I think he is an entertaining guy who needs to read the CWG website more often!

Posted by: stewardsteven | February 4, 2009 12:41 PM | Report abuse

Seriously, Henry needs to grow some thicker skin. It's the internet. He and accuweather set themselves up for it with their ridiculous hypecasting. He's probably a nice guy, but seriously grow a pair dude.

Posted by: tjack55 | February 4, 2009 12:50 PM | Report abuse

Accuweather made a choice a few years ago to go to a "hype" type site. They used to be reasoned and were cautious in making their predictions. But when they took Joe Bastardi off of free radio and tv and put on their premium (pay) site they showed their true intentions. They are no more than the weather version of the National Enquirer.

Posted by: MKadyman | February 4, 2009 12:51 PM | Report abuse

The problem: The models are finite and include only the most important atmospheric data inputs.

The Earth's atmosphere is a system encompassing a nearly infinite set of input variables. No computer this side of the Pearly Gates is capable of processing a set of infinite variables.

The result: GIGO/Garbage in, garbage out. Even when models "agree" the input data may have been in error from the get-go.

Today's computer models sometimes are no better than the Weather Bureau meteorologists of a generation gone by.

Consider the message currently on the signboard in front of Central United Methodist Church in Ballston: It reads:
"Prayer is a relationship, not a wish list."

Unfortunately, we all have our meteorological wish lists, myself included. When the weather does not fit our wishes, we go after the messenger. AccuWeather is no more accurate than its input source [probably NOAA].

Posted by: Bombo47jea | February 4, 2009 12:56 PM | Report abuse

prokaryote, you are correct on some comments saying it was unlikely we would get over an inch. However, when one probability weights the CWG forecast, the probability weighted snow accumulation was over 2". Does not seem consistent.

Terp4Life, I guess my issue was how it was phrased.

Posted by: RMVA | February 4, 2009 12:57 PM | Report abuse

RMVA -- The chance of a goose-egg was supposed to be implied by the chance of "Less than 1 inch." Maybe in the future we'll consider more explicit phrasing (i.e., "0-1 inch"). As it turns out, many metro area locations received on the order of a half-inch of snow yesterday (see totals here), with most accumulation occurring only on grassy areas, as we noted would likely be the case in our forecast.

The bottom line, though, is that when dealing with minimal accumulation potential, the science just isn't good enough (and may never be) to confidently say whether a specific location will end up with a half-inch or nothing at all.

Posted by: Dan-CapitalWeatherGang | February 4, 2009 12:57 PM | Report abuse

... phrased.

Posted by: RMVA | February 4, 2009 12:58 PM | Report abuse


Thanks. My preference would be to include a probability for negligible accumulation. the phrasing of less than one inch implies that there will be accumulation.

Also, how does one consider that the probability weighted snowfall from the forecast was over 2 inches?

Posted by: RMVA | February 4, 2009 1:00 PM | Report abuse

Some people can never be happy. We actually UNDERhyped the Philly snow amount yesterday - our map had them in the 1-3, they got 8. There is one thing we can agree on - this storm was insane as far as the models are concerned. I posted a lengthy blog entry yesterday at going in-depth about model inaccuracy, addressing some points that Bombo mentioned above.

Posted by: weathermatrix | February 4, 2009 1:12 PM | Report abuse

I'm curious as to when the last time the forecast establishment underestimated the size of a snowstorm. I'm sure it's happened, but I can't recall when. I can recall numerous examples of catastrophic snowfall forecasts that ended up, let's just say, quite a bit over the top.

Posted by: shoveit | February 4, 2009 1:13 PM | Report abuse

RMVA -- Not sure where your "over 2 inches" comes from ... that wasn't even a category in the forecast we're talking about. We said there was a 70% chance that accumulations would be 3 inches or less (that includes zero), broken down by a 35% chance of 1 inch or less (that includes zero) and a 35% chance of 1-3", and that the greatest accumulations would be on grassy areas. Not a perfect forecast, but not bad considering quite a few spots in the area picked up a half-inch.

Posted by: Dan-CapitalWeatherGang | February 4, 2009 1:16 PM | Report abuse

I saw that Philadelphia got a whopping 8.5 inches. What explains such an anamoly?

Posted by: irish64 | February 4, 2009 2:02 PM | Report abuse

I am so sick and tired of people bashing Accuweather. Here is a novel idea, if you don't like thier site, don't go to their site, if you don't like JB, don't pay for his services. At least Accuweather takes a stand, and doesn't call 50% chance of rain a forecast. Either it rains or doesn't rain.

As for last week, CWG was initially forecasting a significant storm. Not here in DC, but the significant storm was implied by your initial forecast of rain and warm (otherwise there would be no rain and warmth). Everybody knows Henry is Henry. I happen to like JB, he does good work and that is why I pay for the pro site. Does he get over excited sometimes, sure, anybody who is passionate does this, but if you take the time to listen what he says (not what you want to hear), you might learn something.

Posted by: mciaram1 | February 4, 2009 2:09 PM | Report abuse


January 25-26, 2008. They thought the low was going out to sea as late as the night before - only one area forecaster, on the 11pm news, said it might change direction - and we ended up with at 8 inches of snow, I think.

Posted by: MKoehl | February 4, 2009 2:34 PM | Report abuse

mciaram1 -- Thanks for the comment. Re: "At least Accuweather takes a stand, and doesn't call 50% chance of rain a forecast. Either it rains or doesn't rain" ... sorry, that's not how our atmosphere and the models used to help predict weather work. Sometimes conditions and model information are such that there's only a 10% chance rain will fall, sometimes 30%, sometimes 50%, sometimes 70%, etc. And sometimes one portion of a forecast area receives measurable precipitation while an area only several miles away doesn't. The job of meteorologists is to accurately assess and communicate the chance of rain, snow or what have you. While some outlets communicate forecast uncertainty differently than others, virtually every outlet regularly uses percentages when it comes to precipitation forecasts, including AccuWeather.

Posted by: CapitalWeatherGang | February 4, 2009 2:34 PM | Report abuse

The blog entry about "Secrets of Weather Forecast Models, Exposed" was incredibly informative and enlightening, btw.

Posted by: Terp4Life2 | February 4, 2009 2:40 PM | Report abuse

This was a very difficult to forecast storm. I think people underestimate the amount of uncertainty weather can carry even with our fancy models. If weather was as easy as taking a look at a model and making a forecast, it would not be the least bit interesting. That bit of uncertainty, which does fluctuate like it did with this storm, is the reason I follow the weather. It is the reason I read Henry's blog. That tantalizing bit of "anything can happen" keeps me hooked. And Accuweather likes to overinflate the uncertainty to sensationalist levels, but that is why we have NWS and Capweather - to give us more realistic ideas.

Henry M has been doing his blog for a while - I don't think this was about being thin skinned. I think there were some legitimate threats to make him consider quitting his blogging. I mean it was a FORECAST - he did not say anything about anyone personally - no racial, sexist, religious etc comments. These people threatening him need to get a life and realize that their actions were completely unjustified. There are more important things to get angry about - not a stinking forecast.

Posted by: dotyt | February 4, 2009 2:41 PM | Report abuse

About mciaram1's comment: It really helped me to learn that the forecast percentages mean that there is a certain percent chance of precip falling WHERE YOU HAPPEN TO BE STANDING RIGHT NOW. Even if it's almost certainly going to rain in your region, the chances of it falling on YOUR house and affecting YOUR plans are much lower. But it might fall on your friend's house a mile away. It's just impossible to predict with complete accuracy whose house is going to get rained on and whose isn't. I don't know anything about forecast models, but basic statistics tells you that much.

Personally I think discussion about uncertainty should be open and frank. CWG did not forecast a significant snow event a week ago. They mentioned what the models were saying, but added quite strongly that we should take anything the models say so far out with more than a grain of salt. I would much rather have that kind of forecast than one that says, "even though I have no scientific way of knowing if northwest Washington DC is going to get an inch of snow, I'm going to say it is just because I need to take a stand one way or the other." That does me absolutely no good.

Posted by: LaurainNWDC | February 4, 2009 3:17 PM | Report abuse

I disagree regarding how uncertainty should be handled. 50% chance of showers is not useful to the public. What would be more useful would be hit and miss showers, less than .10" where it rains. Another example is widespread showers, Periodic showers, or Period of Rain. What does 90% chance of periods of rain mean anyway? That is a stand. I understand that models have probabilistic output. It is the job of the forecaster to analyze it and make a forecast. Otherwise, what value does the forecaster provide over the model.

BTW taking stand does not mean issuing some grand useless statement. Accuweather said there would be a big storm that could have a sig impact, but they did not issue useless snow forecast out front. They did say there was uncertainty and details would be forthcoming. I understand that they were wrong. For better or worse, the reality of the world is that information is valuable, and the sooner that you get info out the more value you generate especially for people in the energy sector. If you are repeating what everybody else is saying you are not providing much value nor are you differentiating yourself. That is the pressure that exist as a for profit. BTW I think that people confuse what Accuweather as a whole says versus what JB or Henry think. I suppose that is the risk of having people within your organization blog.

Posted by: mciaram1 | February 4, 2009 3:41 PM | Report abuse

For the first time in awhile, Margusity has actually been toning things down for most of this winter. He's actually been pretty good about not overhyping things (or wish-casting as he has been VERY prone to doing), and one cannot always fault the meteos for what can frequently turn into a business decision instead of one rooted in science. I'm pretty sure meteos don't run that place, but they do work there.

Therefore, I don't think it's solely on Margusity or Abrams or any of the other guys there. People blaming only Margusity are mistaking the trees for the forest. AccuWeather is in this to make money, and they make their money by at times overhyping events. You don't think that press release was written by a meteorologist do you? Was a meteorologist consulted? I'm sure. But based upon the wording, I can bet a meteorologist didn't have final editorial control over it.

Posted by: hobbes9 | February 4, 2009 3:43 PM | Report abuse

hobbes9: I bet you're right about to what extent a meteorologist was consulted prior to sending that press release out on Jan. 30. That same release talked about one to two feet of snow where it would stay all snow. Keep in mind who the audience was for that product: journalists like me, who wouldn't do a story on the storm if it didn't seem to be a big threat.

The discussion about probabilistic forecasting is interesting, especially since forecasters are increasingly relying on using probability to communicate their level of certainty of a forecast and make easily understood forecasts. To me, probabilistic forecasting is far more valuable than "taking a stand" and going all out for snow or no snow. Meteorology is a science, not a political campaign.

Posted by: Andrew-CapitalWeatherGang | February 4, 2009 4:25 PM | Report abuse

Before bashing Accuweather too much, i would make sure the forecasts here are accurate. There was a fiasco in the roads this am esp. in the east with lots of accidents when there was significant snow in the suburbs/exurbs.

Posted by: jojo2008 | February 4, 2009 4:38 PM | Report abuse

jojo2008 -- yep, this morning's dusting, to 2 inches in some spots northeast of D.C., was a blown forecast. Happens to everyone at some time or another, unfortunately. But a blown forecast is a different topic than Andrew's column, which is about hype and the communication of forecast uncertainty.

Posted by: CapitalWeatherGang | February 4, 2009 4:46 PM | Report abuse

RMVA: Flip a coin four times tomorrow morning. On any given day, the probability that you would at least 3 heads is 5 in 16 or about 30%. But sometimes you won't get 3 heads... that doesn't mean the probability was wrong. It just means it didn't happen.

Posted by: prokaryote | February 4, 2009 5:02 PM | Report abuse

Criticise Accuweather's forecasts all you want, but their weather girls are hot!

Posted by: bdeco | February 4, 2009 5:35 PM | Report abuse

I honestly think the problem is no longer with the weather forecasters, but with the general public. Granted I'm from back up North where it snows all the time, but the gripes about the forecasting are amazing to me. Complaining that the team didn't specifically say "0 inches"? Taking a shot because they missed snow showers that hit only isolated areas with an inch or two? You're asking the Cap Gang among others to be perfect in an everchanging, and what will always be an imperfect science.

For this site specifically, you have some dedicated people who focus specifically on the DC Metro area, and try to give the best explanation possible, rather than the one sentence forecasts you get from other outlets. The people who constantly come on here and find ways to bash the meterologists for slight errors in forecasting should really find something better to do. They can't control the weather, only report it. Models are wrong, people are wrong. It all reminds me of the Seinfeld episode where Jerry decides to go to the woman's work who was heckling him at his job of standup. Would any of us want someone critiquing us to the bone for every tiny mistake we make?

My point in all this is that people need to lighten up. It will snow when it snows. One or two inches won't paralyze the'll make it through some unexpected snow showers. These guys get it right most of the time, and when they are right, it's usually spot on. Personally I appreciate the details, whether they bring us snow, or 70 degree January days. All you guys have a pretty thankless job, so you get my thanks for me for being the best weather outlet I have seen to date.

Posted by: JBinAlxndra | February 4, 2009 7:18 PM | Report abuse

AccuWeather deserves every bit of piling on it is getting right now. Anybody remember Accuweather's attack on the National Weather Service's funding a couple of years ago? Accuweather has also not been shy about bashing its competitors publicly.

Posted by: Curmudgeon4 | February 4, 2009 8:04 PM | Report abuse


Thanks for the really nice comment. We're glad you find our content useful and we appreciate the feedback.

Posted by: Jason-CapitalWeatherGang | February 4, 2009 8:05 PM | Report abuse

I feel no shame in bashing AccuWeather for their hypecasting. I thought it was funny when people were posting here "Henry M says blah blah". Noise, nothing but noise.

It's also hilarious how AccuWeather lobbies to have the NWS eliminated. AccuWeather does not have the interests of the public in mind. Rather, their interests are only in making money. You might argue that it is in their best financial interest to be reputable. I have to ask: how does the National Enquirer stays in business then?

Posted by: JTF- | February 4, 2009 8:19 PM | Report abuse

I'm TOTALLY with you, JTF - (and in particular, Margusity) had their black eyes coming to them after their disastrous hype and wishcasting this past weekend. Mrs. Margusity signing on to beg for Henry's forgiveness on his blog was the cherry on top of the sundae....sheesh.

Posted by: VAStateOfMind | February 4, 2009 10:14 PM | Report abuse


in calculating the probability weighted average, I used ...

35 percent of 0 to 1 inch, assume 0.5 times 35 percent

35 percent 1 to 3 inches, assume 2 times 35 percent

20 percent 3 to 6 inches, assume 4.5 times 20 percent

10 percent 6 or more inches, assume 6 times 10 percent

Add all that up and its over 2 inches, probability weighted.

Posted by: RMVA | February 5, 2009 8:47 AM | Report abuse

Stu Ostro over at The Weather Channel has an interesting blog post on the need for forecasters to refrain from jumping on the hype bandwagon as much as possible. In particular, he discusses the early comparisons that were being made between the potential Groundhog Day storm and the Superstorm of '93, and shares an internal email he sent to TWC staff regarding that comparison.

Posted by: Andrew-CapitalWeatherGang | February 5, 2009 11:36 AM | Report abuse

RMVA: "Dan, in calculating the probability weighted average, I used ...
35 percent of 0 to 1 inch, assume 0.5 times 35 percent
35 percent 1 to 3 inches, assume 2 times 35 percent
20 percent 3 to 6 inches, assume 4.5 times 20 percent
10 percent 6 or more inches, assume 6 times 10 percent
Add all that up and its over 2 inches, probability weighted."

I'm not Dan, but since I am a mathematician and since the 2 inches comment was originally to me, let me explain.

Your 2 inches would be the AVERAGE expected snow accumulation over many realizations of storms with this exact spread of probabilities. The problem with relying on the AVERAGE (or mean) for a storm is that you will only get ONE realization of the storm, and the average tells you little about one realization, one storm.

So if the CWG gives the exact same snowfall accumulation predictions for similar storms, say, once a year for the next 20 years. On average, we would get 2 inches each year from only those storms (ignoring all other snowstorms). But some years, that total will be 6 inches, other years 3 inches, and many years under 1 inch (including 0 inches).

If I were to give you money of the next 20 days in these ammounts: 0, 3, 7, 1, 1, 2, 0, 2, 1, 3, 1, 1, 1, 6, 3, 0, 0, 2, 2, 4 (dollars). On average, I'm giving you $2 a day, but some days I give you nothing. Only 4 times do I give you exactly $2. And half the time I don't even give you less than $2. So if you had a $2 cup of coffee you wanted to buy every day for the next 20 days... I wouldn't be counting on me to give you enough every day. You can't rely on the average (mean) on any given day (realization).

JB: "I honestly think the problem is no longer with the weather forecasters, but with the general public."

Yes and no. I can't blame people who were taught math (and statistics) incorrectly. Those expectations are created from experience, often from poor forecasters or poor presentation of the forecast. One of the reasons I like the CWG is that they do give probabilities for accumulation, rather than simply giving a single number, which would probably be the mean, which I hopefully just explained is not that useful for any given storm.

Posted by: prokaryote | February 6, 2009 12:19 PM | Report abuse

The comments to this entry are closed.

RSS Feed
Subscribe to The Post

© 2012 The Washington Post Company