How to win over Wall Street with big data

This year there is a popular phrase, "in front of big data, you don't even have any underwear left." And football is the world's most advanced use of large data, the most successful areas. In the just-concluded Brazil World Cup, including Baidu, Microsoft and other Internet giants have been through large data business, predicted a considerable part of the results of the game. Some have even ventured to comment that, in the face of big data, internet companies are winning over Wall Street, which is represented by Goldman Sachs and Deutsche Bank. The 2014 World Cup will be a new starting point: The big data business formally moves from pure technology to the practical application of everyday life.

What model is critical?

This year's World Cup did not "Octopus Emperor", but there are CCTV beauty host "Squid Liu" and large data companies "to gamble." Finally, the math model designed by science and technology defeated CCTV Beauty's good luck, CCTV sports channel anchor Liu YIXI in the forecast lost to Baidu.

According to the reporter understand, 2014 Brazil World Cup, Baidu, Microsoft, Google and other domestic and foreign technology giants and big investment bank Goldman Sachs, Deutsche Bank and even Bloomberg, have launched large data forecasts.

Internet companies, in addition to Yahoo, almost total victory. Baidu, Microsoft, Google predicted the correct total of the top 16 (Google only forecast the eight strong), as well as the eight strong, Microsoft, Baidu for all the four strong, Google in the four strong forecasts of the failure; in the semi-final, Baidu and Microsoft even accurately predicted Brazil's results against Germany. Yahoo, which had always been accurate, was thought to be using models to promote its business this year, leading to a majority of errors.

By contrast, Goldman Sachs, Deutsche Bank and Bloomberg's three traditional Wall Street firms are far too wrong. The producer of Goldman's World Cup report this year, Kevin Daly, a major economist, led Goldman's quantitative analyst to Kevin Daley a model of the formal international football matches since 1960. But Goldman Sachs ' eight-strong forecasts were wrong in three (Costa Rica, Belgium and Colombia), and one of the top four (in Spain) had the right rate of 37.5% for the group match.

Bloomberg's approach, similar to Goldman Sachs, simulates 10,000 results on the basis of FIFA scores of national teams, and is similar to Goldman Sachs, which says that Spain will be in the final with Argentina after a penalty shoot-out and eventually lose to Brazil, but the reality is that Spain has already gone home.

Deutsche Bank's model is even more outrageous, and its model is said to have combined the FIFA rankings, historical record, player composition and odds of betting, which results in Brazil's 8-strong, and England's title.

Year or large data

Industry analysts believe that the limitations of cloud computing capabilities, Goldman Sachs, Deutsche Bank, Bloomberg, represented by Wall Street companies to use the data is not large enough, is the cause of their calculation errors. In fact, looking back at the World Cup, the predictions of several big investment banks are guesswork. The last Morgan Stanley said England would win the championship, and UBS said Italy could enter the top four, far from the actual results.

"The big data is not the technology, but the application." "said Zhang, chairman of the media consultancy. And the industry generally believe that, with the large data business, the 2014 World Cup will become the domestic big data business officially mature starting point.

According to Baidu related people, at present, Baidu large data has been in the tourism, education, race industry has tried many times, some products are expected to be commercialized in the year. Following the World Cup, Baidu's big data forecasts will also be launched at the box office forecasts, real estate forecasts, financial forecasts, SME index forecasts and other products, and in mid-September to launch the forecast open platform, the largest open Baidu in the large data field data and technical capabilities, to promote the rapid development of the whole large data industry. Baidu has just dug from Google Brain three top experts in artificial intelligence Wunda, and announced that "Baidu Brain" has reached 2~3岁 intelligence.

Accuracy rate of World Cup forecast

Baidu: Forecast in 16 World Cup knockout, only in 3, 4 finalists, finally achieved 16 15 of the results, the forecast accuracy rate is 93.75%. And this year frequently upset group race stage, Baidu to match the result forecast accuracy rate also reached 58.33%.

Microsoft: Microsoft has successfully predicted the outcome of the World Cup knockout 16 games. Even before Brazil's semi-final against Germany, it predicted that the Germans would sweep the home side of Brazil. This year, upset's group race phase, Microsoft reached 56.25% accuracy rate.

Goldman Sachs: Goldman Sachs predicts that Brazil will be the last to win the Cup, and now we see that Goldman's predictions are straying from the wrong track. Goldman Sachs's forecast for the World Cup is only 34% accurate.

View

Large data not to be trusted

Some local economists told reporters that the big Data system is no longer strong, can not avoid the "Black Swan" incident. Chen, Ph. D. In economics, argues that because the players are the top players at the top of the World Cup, the probability of their being out of whack is very low and large numbers are likely to be used. "But if it is a more complex economic estimate, the risk of failure of large data systems is much higher." ”

"In fact, behind every international football match there is a huge gambling group in the manipulation of large data can not be fully believed, otherwise you will lose miserably." "A senior local fan told reporters.

(editor: Mengyishan)