Our April Fool pranks are a tradition that has been carrying on for well over a decade. We have always attempted to make them do what they are supposed to do: fool as many people as possible and be entertaining. Of course it is trivially easy to fulfil the first criterion: "ChessBase to manufacture bottled drinks" is something nobody can easily check. But it is not funny.
We believe we have succeeded in the second criterion fairly well in the past: we produced a gaudy, blinking news page (for a full day) last year, or told of the discovery that Matt Damon was the second cousin of Magnus Carlsen the year before. However, the task of actually fooling people has become progressively more difficult. Armed insurgents, equipped with advanced Google searches and massive forum discussions, attack our news page every April 1st and then send letters or post forum messages telling everybody what the April joke was and how easily they had recognised it.
So we started hiding the prank behind fake jokes, i.e. publishing reports that sounded fairly outrageous but were perfectly true. Take for instance April 1st 2010: we started with an article telling about how grandmasters were worried about the Large Hadron Collider at CERN in Switzerland/France. Later in the day we reported that authorities in Reykjavik had decided to exhume Fischer body to extract DNA in a paternity suit. Both reports were true. Then finally the Carlsen-Damon report. It was interesting for us to see the successive tsunamis of letters strike our mail box after each report appeared on the news page.
But things are getting more difficult, as the insurgents arm themselves with ever more powerful tools. This year, when we published our first fake April Fool report, virtually everyone had googled Capablanca and billiard to find the story was a hoax – but one not perpetrated, just reported by us. The yarn was invented in 1950, and the game which was supposedly won by Capablanca was played by Hoffer in 1880. Everyone found that out from this detailed account, published on Edward Winter's Chess Notes page.
The second report on April 1st was about how you can use Fritz 13 and Let's Check to monitor what the international chess community was cooking up, and where you could expect your opponents to come up with stunning novelties. Perfectly true – everything the article said had its basis in reality.
And then we published the press release of the Danish auction house Bruun Rasmussen, which will be selling the historical 1972 Fischer-Spassky chess table 1972 for over 300,000 dollars. As luck would have it the auctioneers had not yet set up a separate page for this item – or it was too difficult to find. So many readers fell for that, believing the story was a fake. Not so.
So what did we do this year to throw the predators off our tracks? Well, publish the real joke on April 1st at 23:55h, i.e. five minutes before midnight – in Pago Pago. On the one hand we have lots of friends and fans in the capital of American Samoa, and on the other we have always felt sorry for the inhabitants of this Pacific island, which is always the very last place on earth to celebrate the New Year – or in 2000 to enter the new millennium.
Publishing the article when it was 23:55h in Pago Pago, however, did not override out content management system's fixation on European Daylight Saving time, to which it currently adheres. For it we had published the report on April second, and shucks to the Pago Pagoians and everyone else. Note that our previous stories, to which it attached the April 1st publication date, appeared when it was already April 2nd in New Zealand or Petropavlovsk (lots of friends and fans there as well).
In order to compensate for this slightly devious trick we did put in some glaring pointers: that our interview with Vas Rajlich, in which he claimed to have busted the King's Gambit, was conducted one day after his move to Budapest on March 31st (wink, wink, nudge, nudge), and the PGN of his analysis was also set to that date. Many readers picked up on that. In addition, after consulting computer expert friends, we "preposterised" the original interview, claiming that IBM had provided their Watson super cluster to Rajlich for over four months to calculate the King's Gambit. "But that gives it away immediately," said Vas Rajlich, who had helped us to craft the prank.
Apparently not, or not at once. We realised this when international journals and TV stations started contacting us for statements and interviews. One New York reporter from a prestigious scientific publication was definitely annoyed: "This interview is over," he said, when we broke into his first question ("Tell me about the history of the King's Gambit"). "Congratulations, you fooled me completely. And I got up at 04:30 a.m. to do the interview." We still feel bad about that. The BBC reporter, on the other hand, laughed heartily over the explanation.
Finally we started to get emails and Skype messages from very strong players with messages like: "Fred, I know you've probably been asked this quite a bit recently – but the Be2 not Nf3 stuff is April's Fool's joke, right? No-one is insane enough to actually do that, I assume (and hope)." Another very strong young GM wrote: "Hi Frederic, I always appreciate the ChessBase April Fool's reports. I'm eager to test out 1 e4 e5 2 f4 exf4 3 Be2! equalizer" (plus smileys). How had he known?
Note: The chess historian Edward Winter has provided (factual) background
reading on the opening 1.e4 e5 2.f4 exf4 3.Be2 and its origins here.
As the reactions started to pour in we started to panic: how could we explain the real reason why the Rajlich interview had to be fake? We asked our computer chess friends to help and got a couple of good pointers. Ken Thompson, one of the pioneers of these kinds of searches and database calculations, told us that the task was "too big by huge magnitudes. Especially the draw discussions (Be2), where all moves for both sides have to be expanded." Vas Rajlich, sitting in a "funky Internet café in Budapest", explained in greater detail:
It's reasonable to construct a search tree of around 10^18 positions using modern technology. The chess alpha-beta tree is thought to have at least 10^45 positions. The alpha-beta tree for the King's Gambit will be at most 10x to 100x smaller than that. So, we're still probably a good 25 or so orders of magnitude away from being able to solve something like the King's Gambit. If processing power doubles every 18 months for the next century, we'll have the resources to do this around the year 2120, plus or minus a few decades.
Actually Vas is being overly optimistic, and we are probably overly pessimistic when we say: it will not be possible in the course of this universe. The Rybka author added the following caveat:
You must remember that the tree for any specific 32-man position can be much smaller than we expect – if one side is immediately lost or if there is an immediate forced draw. Could the King's Gambit have a massively reduced tree? If the King's Gambit is winning for black, then this is theoretically possible. It is possible that after 2.f4 White is simply crushed instantly, no matter what he does. I very much doubt that this is the case. Everything in my chess experience tells me that any Black win would be long and tortuous. If the King's Gambit is a draw, though, then there are really no prospects for a massively reduced tree. In other words, it is (slightly) more preposterous to claim that 3.Be2 has been worked out to a draw than to claim that 3.Nf3 has been worked out to a loss.
There you have it – Peter Svidler and the other Super-GMs can sleep peacefully, knowing that something like this "solving chess" thing is not going to happen – at least not until next April.
In the following we give just a small sample of the letters we have received, chronologically, with the potentially embarrassing ones anonymized. The chess forums were full of wild and hostile postings, which you can read if you are brave. And just google "rajlich bust kings gambit" (without the quotes of course) and see where this story has been to around the world. Even Wikipedia has added a paragraph to its King's Gambit article: "In April 2012, IM Vasik Rajlich, a chess programmer who created Rybka, claimed to have proven to a 99.99999999% confidence level that Fischer's article was nearly but not entirely correct. Rajlich's program utilized IBM hardware similar to its Jeopardy winning Watson system and encompassed around 3000 cores. The search concluded that after accepting the gambit, black has a forced win against all white responses except the unusual 3. Be2, which allows white to hold on for a draw." What have we done, we ask ourselves. And why do we do it?
Douglas Wiggins, Boston,MA
You have no article for April Fool's day! That's usually a tradition for ChessBase!
N.G., Sussex, England
Jakovenko European champion? Nice April Fool's joke!
Anonymouss, Boston, MA, USA
Was fooled the past two years but was watching for it this time! Once again, well done.
I wonder if you guys will ever run out of ideas for April Fool. Chess + Snooker was very nice.
E.B., San Pedro, Laguna, Philippines
Clearly a fabricated news on Capa's billiards. This was the first time I heard of the chess-billiards match, though I read somewhere that Capablanca was a strong billiards player that if he had devoted more time to the game then he would have been a world champion. Chessbase rocks!
Dennis Cesar Caluban, Kuwait
Capablanca's Chess plus Billiards match is the April Fool's Day joke, as mentioned by the great chess historian Edward Winter's article "A Chess-Billiards Concoction". To be honest, I was fooled into 'watching' Topalov-Kramnik Friendship Match on Playchess Important Broadcast.
Richard Rose, Fairhaven, MA, USA
Nice try. The Capablanca/billiards story is a fabrication April fools. I would suggest you check out Edward Winter's Chess Notes site for all of the facts.
N.S., Swansea, UK
You got me thinking this time. Crystal ball Vs Chess & Billiards! I vote for Crystal ball as there is no way it can be true.
T.T.K., San Diego, California, USA
I suppose you have published an article about the supposed chess-and-billiards match between Capablanca and Hagenlocher only because it's April 1st. As is now well known, this story was a hoax, as for example can be read here: http://www.chesshistory.com/winter/extra/chessbilliards.html
Richard Reich, Fitchburg, WI, USA
The articles give it some verisimilitude, but this game has been known to be a hoax for a long time. Edward Winter wrote a nice expose, but of course you knew that. But still, it it more credible than the campaign promises of candidates for the US presidency!
K.N., Suva, Fiji
News Chess Match of the Century table up for sale? I reckon this is an April Fools Hoax. But nice try. Well done CB.
Mats Winther, Johanneshov
Concerning the solution of the King's gambit. Maybe it is time to make amendments to the chess rules. "Chess with extended castle" would make the King's gambit playable again:
Alessandro Mossa, Firenze, Italy
Eh, eh, eh, very clever: three red herrings on April 1st, and one big fat April's Fool on April 2nd. The technical gibberish is interesting, and portions of it even make sense, but, seriously, SOLVING the King's Gambit? Not in this world... I especially appreciate the humour of selecting 3.Be2 as the only forced draw. Congratulations, guys!
On the English
Chess Forum a member reached the following interesting conclusion:
The first paragraph: "Fifty years ago Bobby Fischer published a famous article, "A Bust to the King's Gambit", in which he claimed to have refuted this formerly popular opening." Look for what stands out, people: the gag is in "popular opening", which is a perfect anagram of "April One Popgun". A popgun is a toy gun, a joke gun. A joke. It's an April First Joke!
This is fascinating stuff. I myself spent many hours thinking if chess could be solved. I am of the opinion that chess can be solved. Tablebases are the clear example that it is possible to solve chess. We have tablebases for 6 pieces. Guess what, in 2015 more or less will be the 7 pieces tablebases coming out. If we can improve from 6 to 7, why not in the near future evolve to 8, 9, 10 to 32 pieces tablebases?! It took us 16 years to solve checkers. The outcome was a draw. Chess will be the same, it will be solved, but it might take decades or even centuries.
What Rajlich has done is to use a confidence interval in which he can be "almost" sure that certain variation wins or loses. Thus, it is reducing the amount of work to be done dramatically. It is theoretically possible that there are some flaws, but I don't think so. This is fantastic because he is creating "tablebases" for certain openings, without having to use brute-force method to check every possible line.
This means that we will be seeing tablebases for many opening lines in the coming years. I can wait to see the results. This is absolutely fantastic. I think this is just a few small steps towards "solving chess". For people who fear this ever happening, don't worry we won't be alive till then. Even when it gets solved, people will still play it. Can you memorize a couple millions of lines? Nah-ah, so our great^10 grandchild will still enjoy chess just like us!
People like this who can't play but choose to kill the game are sickos! This is all very disturbing. How can he be smiling? I spit at the photo. This an evil beast. This is terrible.
Ed Gaillard, New York, NY
I don't know about other countries, but the USA has various holidays that are observed on the following Monday when they fall on a weekend. However, April Fool's Day is NOT one of them. Just thought you might want to know.
Francesco Tosi, Latina, Italy
Nice try, it almost got me, but there is a dead giveaway. "The next day, in spite of the bustle of moving boxes and setting up phone and Internet connections Vas, kindly agreed to the following interview, which had been planned some months ago... A nearly successful try, indeed.
E.W., Princeton NJ
Concerning your article on the "bust'' of the King's Gambit, your interviewer did not explore what may be the most surprising claim here. It is not surprising that a computer could determine the outcome of the very sharp lines that can follow from the King's Gambit in which one player has a forced win. What is surprising – enough so that I am a bit skeptical – is the claim that the computer has demonstrated that the line with 3. Be2 is a draw. Unless there is a demonstrably forced draw – one side has to force perpetual check, or something, but this was not claimed in the article – the best the computer analysis is likely to have been able to show is that the play after 3. Be2 is close enough that it was not possible to determine for sure the outcome with best play. Still, it is fascinating that the main line 3. Nf3 has been shown to be a loss for White (I find this highly believable) and that Bobby Fischer had the main points right.
Stanislav Tsukrov, Germany
The King's Gambit story is an April Fool's joke, right? Please tell me it is!
Okan Isbilir, Istanbul, Turkey
Is your article about the King's Gambit a post-April Fool's day joke? if so, congragulations! it is so amazing and believable!!
Omer Tunali, Ankara Turkey
But even with the evaluation function > 5.21 is decisive assumption it is impossible to prune the tree in such an early stage, using current technology. Perhaps Rajlich learned how to program quantum computers?!
Does it mean that the King's Gambit will be forrbiden in tournaments, to avoid quick draw?
Evan Katz, Manhattan, New York (Former Contributing Editor
of Chess Life Magazine)
As always, a hearty congratulations and thank you to everyone at ChessBase – and this year also to Vas Rajlich as well – for a very funny, highly amusing, and quite original April Fool's article regarding the alleged "solving" of the King's Gambit! The April Fool's article was extremely clever and greatly appreciated, even if it actually did come out on April 2, one day after April Fool's Day! Thanks again.
Yibing Fan, Toronto, Canada
Very good writing! At beginning I 99.9999999% believed it. And then those slight chance made me think: if the King's Gambit is solved, then there are only limited openings, and then after all chess is solvable, isn't it? Then I realized it's April 1st. Ha ha, very nice try! Anyway, this is best April 1st article. I am 99.9999999% sure many people will be fooled.
Martin Bennedik, Germany
Congratulations on being Slashdotted. However I am calling this your April Fools. I think it is possible to prove a win in some positions, but you also claim that a draw was proven. However, there should be too many positions in the drawing lines to prove a draw. In checkers this is different, as you can peek from the opening right into the endgame database – because every move brings you one step closer to the endgame. Not so in chess. You can make too many moves forth and back while still being in the drawing zone. There would be too many drawn positions even for the Watson Jeopardy cluster.
Frank Pi, Vienna, Austria
No! please!!, don't say this Rybka-King's Gambit story is a tardy April 1st joke. It has the date 02.04, but the game inside has 01.04!!?! don't let me down.
H. Steinenacher, Karl-Marx-Stadt, GDR
Surely you jest! To go from computers just having solved six piece endings to having them solve the entire King's Gambit! I believe this to be yet another clever hoax, though much more plausible than your first two. What other "fools" will you think up next, dear Chessbase??
Randy Hoch, Spring, TX USA
I must admit I'm stumped this year. You have three stories which could equally pass for authentic or a prank! Although the auction of the Fischer-Spassky chess table is corroborated on other sites, I have a difficult time believing historical items like this would be auctioned off. Of the two stories left, I give greater credence to the Capablanca chess/billiards face-off. So the April Fool's article is probably the Bust to the King's Gambit!
Mathieu Buard, St Etienne, France
Aren't you late for this year's April Fool? published 02.04.2012, but I can't belieave it's true while yesterday news are very credible... But still lot of fun to read! Keep on the good work.
Vincent Gagliano, Ft. Lauderdale, Florida, USA
This has to be the annual April Fool's day story. First, forget the King's Gambit. Imagine the upheaval in the chess world if Rajlich solved the Ruy Lopez (as in, beginning after Bb5) or the Open Sicilian, which are both the same number of moves in. It would make revolutions by Steinitz, Nimzovitch et al. look tame. Second, the square root of 10^100 (10^50) is a pretty big number in its own right, but even with 10,750,000 hours of single core time, 10^20 positions would be a reach, which is a tiny, tiny fraction of all possible positions.
Sam Trenholme, Puebla, Mexico
You really need to restrict posting of April Fools jokes to April 1; it's unethical and shoddy journalism to post an April Fools joke with a posting date of April 2.
Jimmy Liew, Puchong, Malaysia
Every ChessBase reader in the world was searching for the traditional April Fool article on April 1st. But ChessBase was sneaky this year and publish it on April 2nd! I declare "Busting the King's Gambit" the April Fool article for 2012!
Michael Viking, Portland, USA
The article is posted on 02.04.2012. I understand the premise, but people understand time. If it was posted on 2.4.12, it's posted on April 2nd and is therefore no an April Fools joke. Not sure about Germany, but in America, doing an April fools joke on April 2nd is a big faux pas...Great concept...too bad it came out the 2nd.
Pablo Pena, Irvine, CA
I find the accuracy of this article very suspect. Perhaps it's an April fool's joke! The alpha-beta model of pruning variations is hardly 100%. As I understand it, the only reason he gave an accuracy of less than 100 was because the cut-off of evaluations is around 5 and it's rare that the weaker side can draw such a position. However the computer is pruning by "best" moves until it reaches 5.16 (or thereabouts) but it's "best" moves can often be fallacious due to the "horizon effect" (which if it tried to remedy would result in the analysis of 10^100 moves). I can't tell you how many times I put in a position in Rybka that showed a 2.5 + for white only to have the computer play itself and completely flip-flip after say 10 or 15 moves (this is particularly true of dicey tactical positions as in the Dragon or Botvinnik SemiSlav).
Samuel Siltanen, Espoo, Finland
Thanks for the article "Rajlich: Busting the King's Gambit, this time for sure". It was a great April Fools' joke. As a computer programmer and chess enthusiast I know that chess could be solved in theory if we had enough computing power. However, it is not possible with the current hardware even with the obscure optimizations suggested by Rajlich in the interview. Also, if one can solve the King's Gambit, why could not one solve Najdorf as well or the initial position? This logical contradiction, in addition to the date of the interview, was what revealed that this was the April Fools' hoax in 2012.
Leonard, Blackburn, Champaign, United States (Associate Professor of
In your article "Rajlich: Busting the King's Gambit, this time for sure" Ralich states that "if Rybka is displaying +5.12 or more the outcome is 99.99999999% secure" (which is probably a guess, but I won't argue it). Then from this Rajlich erroneously concludes that "That is approximately the confidence number we give to our King's Gambit results: 99.99999999%." If their engine reaches +5.12 n times, then the probability that it is correctly concluding a forced win all n times would only be (0.9999999999)^n. So, the more times it is reaching a +5.12 position, the lower the probability gets that it didn't make a wrong conclusion somewhere. It would only take about 6.9 billion times of reaching +5.12 and ending the analysis before the probability sinks below 50% that it didn't make a mistake somewhere. Otherwise I enjoyed the article (and your website in general) immensely.
Ed Zator, Toronto, Canada
You suckered me again!
– Probably more to come –