When the final points were tallied, Pitt shattered Carnegie Mellon University by almost $5,000… When the final points were tallied, Pitt shattered Carnegie Mellon University by almost $5,000 in a round of a game similar to “Jeopardy!” and triumphed in a game of wit and trivia.

Well, technically Pitt came in second place, but no one expected the humans to win, anyway.

The victor in the matchup was Watson, a supercomputer that launched to fame after it dominated a special “Jeopardy!” game last month. IBM trained the machine to participate in the game against two “Jeopardy!” champions, Brad Rutter and Ken Jennings.

Similar to an oversized tablet, Watson is a 2-and-a-half by-1-foot screen with a globe as an avatar. It is able to speak its answers in a computer-animated voice. IBM brought the machine to CMU’s campus yesterday, its first stop on a brief college tour. More than 300 people watched as three representatives from Pitt and CMU squared off against Watson in the mock “Jeopardy!” game.

But in an ironic twist, the focus of the event was not the machine’s ascendance toward world domination. Rather, the all-day symposium celebrated the humans who created the machine — which then went on to defeat the humans.

The trivia game was the culmination of the day, which began with a conference on innovative IBM Watson technologies.

Chancellor Mark Nordenberg kicked off the event, remarking on IBM’s technological prowess.

He also lauded Pitt and CMU’s symbiotic relationship and how it enabled the two universities to foster knowledge and innovation.

But the partnership didn’t stop the game from getting a little competitive.

After the Pitt students were introduced at the start of the game — all are members of Pitt’s Honors College — Randy Bryant, dean of CMU’s School of Computer Science, brought in some good, old-fashioned trash talk.

“We like to think our entire university is an honors college,” he said.

But the Pitt students prevailed. Brian Sisco, Richard Kester and Danielle Arbogast represented Pitt, quickly answering the few questions they could buzz in time for.

Watson answered most of the board, but there was one category where Pitt completely dominated the competition: Wine, White Wine.

Luckily, Pitt’s break came during the second round, where the value of each question doubled. The category asked about different grapes and regions involved in white-wine production.

Pitt swept the board — beating Watson — even snagging a Daily Double, where it could wager as much money as it wanted to.

With $9,800 in the bank, Pitt made a relatively cautious gamble of $800. In the end, it was a smart move, as the students could not identify what blanc the Marseille region is known for.

“I really need to drink more,” said Kester, a senior neuroscience and history major. After some deliberation, the group went with, “What is rum and coke?”

Wrong.

Before the event, Kester said the group didn’t undergo much preparation for the game. All are officers of Pitt’s Quiz Bowl, a group that participates in academic-oriented trivia tournaments.

But even with the group’s experience with trivia, Kester didn’t have high expectations for the team, especially after watching Watson play “Jeopardy!” last month.

“I saw him play Ken Jennings. My heart sank a little,” he said.

Kester said that the questions in “Jeopardy!” typically aren’t that hard — it’s retrieving the information quickly that’s the challenge.

He added that the challenge wasn’t about trivia, it was about buzzing.

Watson receives the text of the question the same time it appeared on screen for participants to read. While the humans read the words and searched for the information, Watson would take the text and run it through a series of algorithms to determine the answer.

It will only answer questions it is confident in: During the game, Watson’s confidence level was represented by a bar graph on a large screen behind the players. Typically, Watson would only answer if it was more than 50 percent sure the answer was correct.

David Ferrucci, IBM principal investigator and “father” of Watson, explained the process in a lecture before the game.

Using the question, “In cell division, mitosis splits the nucleus and cytokinesis splits this liquid cushioning the nucleus,” Ferrucci broke down how Watson tackles the question.

First, Watson goes through all the text it has in its memory and comes up with a list of possible answers. In this case, Watson came up with organelle, vacuele, cytoplasm and plasma. Using key words in the question, Watson then scans for phrases or sentences that would locate the correct answer.

In its search, Watson found, “Cytoplasm is a fluid surrounding the nucleus.”

But Watson’s weakness is the ability to completely understand the nuances of language. Whereas a human would understand that in this context, a fluid is a liquid, Watson does not.

So, it runs an additional algorithm to see if a fluid is a liquid.

Although Watson is notorious for defeating “Jeopardy!” champion Ken Jennings, it didn’t begin as a super-intelligent machine.

In its developmental stages, Watson once answered, “This Frenchman was the father of bacteriology” with “How tasty was my little Frenchman?”

But it didn’t make any goofs like that in yesterday’s matchup. In fact, the computer answered so many questions in the first round that moderator Eric Brown, an IBM researcher, was forced to slow down Watson’s response time.

He also started reading the questions more rapidly, giving the humans more time to buzz in the answer. The humans were only allowed to buzz after he had read to a certain point.

Even with the help, both Pitt and CMU fell quickly to Watson. It would frequently answer entire categories, like Classical Music, Romans and O Gods, causing the CMU students to throw their hands up in frustration.

The last question’s category was Characters in Classical Literature. The query was, “The first person mentioned by name in ‘The Man in the Iron Mask’ is this hero of a previous book by the same author.”

All three got the question right: D’Artagne.

Even though Watson had a healthy lead heading into the final round, it still wagered $15,099, launching it to a grand total of $52,199. Pitt made the smallest wager of $1,937, bumping it up to $12,937, and CMU still came in last with $7,463, almost doubling what it had with a $3,063 wager.

Junior Doug Slocum, who is also a member of Pitt’s Quiz Bowl, was at the event to cheer on his school. He said he was proud that Pitt got second, even though he wasn’t expecting it.

“I expected them to get shlanked by Watson,” he said.