SERVING THE QUANTITATIVE FINANCE COMMUNITY

ppauper
Posts: 70239
Joined: November 15th, 2001, 1:29 pm

### The amazing Ken Jennings

QuoteOriginally posted by: AGatISCWill his overall record fall as well?http://tv.msn.com/the-wrap/jeopardy-rec ... =28103well spotted, AGatISC !he has a ways to go yet as he first appeared on "Jeopardy!" on Sept. 13 He just won his 3rd game:Show #5978 - Wednesday, September 15, 2010ContestantsKatie Ganfield, a mom from St. Cloud, MinnesotaKevin Knudson, a mathematics college professor from Gainesville, Florida (University of Florida)Roger Craig, a graduate student of computer science from Newark, Delaware (whose 2-day cash winnings total $114,000)Scores at the end of the Double Jeopardy! Round:Roger Kevin Katie$13,800 $12,200$6,400 Final Jeopardy! RoundMONARCHS From 1513 to 1972, only men named Christian & Frederick alternated as rulers of this nation Roger got Denmark, the others missed with Sweden and AustriaFinal scores:Roger Kevin Katie $24,401$200 $6,400 Roger 3-day champion:$138,401

ppauper
Posts: 70239
Joined: November 15th, 2001, 1:29 pm

### The amazing Ken Jennings

sadly, Roger Craig only won 6 games and lost the 7th on Tuesday#5982 - Tuesday, September 21, 2010Roger Craig game 7.ContestantsJelisa Castrodale, a sportswriter from Winson-Salem, North CarolinaMatt Martin, a police officer from Arlington, VirginiaRoger Craig, a graduate student of computer science from Newark, Delaware (whose 6-day cash winnings total $230,200)Scores at the end of the Double Jeopardy! Round:Roger Matt Jelisa$22,000 $6,000$19,700 Final Jeopardy! RoundSPORTS & THE MEDIA On Feb. 8, 2010 the headline in a major newspaper in this city read, "Amen! After 43 Years, Our Prayers Are Answered" Only jelisa got "new orleans"Final scores:Roger Matt Jelisa $4,599$5,000 $39,399 New champion: Jelisa$39,399

rmax
Posts: 6080
Joined: December 8th, 2005, 9:31 am

### The amazing Ken Jennings

Ken is back! But this time versus Big Blue:Watson versus Ken

ppauper
Posts: 70239
Joined: November 15th, 2001, 1:29 pm

### The amazing Ken Jennings

QuoteOriginally posted by: rmaxKen is back! But this time versus Big Blue:Watson versus Kenoh wow, this should be interesting !

frenchX
Posts: 5911
Joined: March 29th, 2010, 6:54 pm

### The amazing Ken Jennings

So what is your bet ? The man or the machine ?

ppauper
Posts: 70239
Joined: November 15th, 2001, 1:29 pm

### The amazing Ken Jennings

QuoteOriginally posted by: frenchXSo what is your bet ? The man or the machine ? maybe rowdyroddypiper will take your wager (just don't expect to ever collect your winnings from him )

ppauper
Posts: 70239
Joined: November 15th, 2001, 1:29 pm

IBM's supercomputer Watson wins Jeopardy01/17/10 IBM's supercomputer competed with former Jeopardy champions in a trial run. Watson will face off for the $1 million grand prize, in February. trackstar Posts: 26770 Joined: August 28th, 2008, 1:53 pm ### The amazing Ken Jennings I did watch Jeopardy to see how Watson would do last night. He (it?) was far in the lead for awhile there, but one of the humans made an adjustment in timing, it seemed to me, and then pulled even.Jeopardy! champs aim to take down IBM Watson computer - Scientific American Feb 14 Traden4Alpha Posts: 23951 Joined: September 20th, 2002, 8:30 pm ### The amazing Ken Jennings QuoteOriginally posted by: trackstarI did watch Jeopardy to see how Watson would do last night. He (it?) was far in the lead for awhile there, but one of the humans made an adjustment in timing, it seemed to me, and then pulled even.Jeopardy! champs aim to take down IBM Watson computer - Scientific American Feb 14These contests always amuse me.Watson will double in speed in 18 months, people will double in speed in perhaps 180,000 years (and that's assuming that smart people have more babies than non-smart people). trackstar Posts: 26770 Joined: August 28th, 2008, 1:53 pm ### The amazing Ken Jennings Actually, the human exhibited adaptive behavior, as well as increasing his own clockspeed. (This was Brad Rutter, not Ken Jennings.)Watson did not do well at the extrapolation of odd fact assortment-to-decade and the humans did well at this. ("What is the 1980s.")Several times saw Watson had the correct answer in the top position on his bar chart, but his confidence level was below his answer threshold. So the humans jumped all over that category.I'd have to watch again, more carefully, to see where else Watson was lagging, but there were a few places.If the tech team lowers Watson's answer threshold (ito confidence level), he would thrash them though. It seemed to me that more often than not, his answers were correct, even when he did not put them forward. **Going to refresh an old thread with media coverage of this match up and AI:Ghosts in the Machines **PS: You seem to have alot of confidence in Moore's Law. Last edited by trackstar on February 14th, 2011, 11:00 pm, edited 1 time in total. Traden4Alpha Posts: 23951 Joined: September 20th, 2002, 8:30 pm ### The amazing Ken Jennings The other issue if replicability. Once we have 1 Watson, we can make 1000, a million, or a billion more Watsons just as good as the first Watson and the cost per copy will decline significantly as the volume increases. Probably about 95% of the manufactured Watsons will be as good as the original -- the other 5% having some sort of manufacturing defect of varying levels of repairability.In contrast, try replicating a Brad Rutter or a Ken Jennings. The cost per copy isn't going to decline much with volume, it takes 33 years to make one, and the yield rate is horrible. Unless you believe that genetics predetermines 100% of performance, very few of the Brad Rutter or a Ken Jennings clones will be as good as the original. trackstar Posts: 26770 Joined: August 28th, 2008, 1:53 pm ### The amazing Ken Jennings We take different angles on what is interesting about this contest.I am interested in the strategy of the game and how often the people who are really good at it are simply mining their own strengths versus exploiting their opponents' weaknesses. And I suspect that a computer would be much better at calculating that matrix (in addition to everything else) than people thinking on the fly with tight time contraints.Maybe ppauper could tell us what is typical in the contest. Do people usually pick off the lower priced questions in all categories or do they tend to work up through the levels of single categories?Watson was definitely trying to take away the low end at first. I was amazed when he hit ~$3,000 and Ken and Brad were still in the low hundreds. But the rout did not continue that way.I'll have to watch again though. And certainly will tune in tonight!***Perhaps Jeopardy should be redesigned to provide a different range of game strategy challenges.
Last edited by trackstar on February 14th, 2011, 11:00 pm, edited 1 time in total.

trackstar
Posts: 26770
Joined: August 28th, 2008, 1:53 pm

### The amazing Ken Jennings

I will give you two simple examples.Let's say there are six categories in our modified Jeopardy and it is set up like Trivial Pursuit.You go first and pick the Science catagory.I happen to be good in Science and great in History.If I have reason to think that you are bad in History, once you get an answer wrong, I might try to take the rest of Science away from you before exploiting my strength in History.I don't think players generally do this - too risky and decisions have to be made too quickly. But in a longer, more elaborate game, it could make sense.**In Trivial Pursuit, you could play a version of the game where you limit the number of pie slices for each category.So aside from the natural tendency to go for the slices in your stronger areas first, you can also try to take the rug out from other people. 6 Players, only 2 slices available and I know that you are good in the Sports category? Let me go for Sports now and maybe you will be stuck trying for a Literature slice later on.
Last edited by trackstar on February 14th, 2011, 11:00 pm, edited 1 time in total.

Posts: 23951
Joined: September 20th, 2002, 8:30 pm

### The amazing Ken Jennings

I agree about the dynamic/adaptive strategy angle. Although the ground rules of a game may be fixed, a player's strategy must be a function of other's players' strategies, strengths, and weaknesses. At the beginning of a game, these opponents' characteristics may be unknown. So part of the early phases of game play would be devoted to exploring the opponent's behaviour while obfuscating one's own behaviour.Another interesting aspect of this game is that it really is a game designed by humans for humans. All the word play and cultural allusions seem biased to favor human contestants. I'd even wonder if the human question designers for these rounds knew that Watson would be a contestant and how that influenced their question construction activities. Did the game designers cheat? (Note: there's a higher-level game being played here in which the Jeopardy producers wish to maximize ratings, not create a fair test of who's best. I'd wager that the question designers have intentionally patterned the questions to minimize the chance that either humans or Watson keeps the lead for long.)Anyway, I think I know what would happen if the tables were reversed and a best-of-the-best human tried to play a game designed by computers for computers -- "Quick! Find 1 billion double-precision square roots!" Perhaps some Babylonians would win.

ppauper
Posts: 70239
Joined: November 15th, 2001, 1:29 pm

### The amazing Ken Jennings

Show #6086 - Monday, February 14, 2011The IBM Challenge.Jeopardy! Round only.ContestantsBrad Rutter, a $3.2-million winner from Los Angeles, CaliforniaWatson, a deep question answering system from IBMKen Jennings, a 74-game champion from Seattle, WashingtonScores at the end of the Jeopardy! Round:Ken Watson Brad$2,000 $5,000$5,000 The key thing in this show is often timing: If all 3 know the answer, you have to get the buzzer in first.And then there's hunting for the double jeopardy squares: when you get a question right, you get to choose the next answer, and if you hit the double jeopardy square you get to answer it unopposed and make a lot of money