<p>
IBM is donating Watson’s winnings to charity.</p>
<p>The other two contestants are donating half their winnings. </p>
<p>1st - $1,000,000
2nd - $300,000
3rd - $200,000</p>
<p>There was also a nice article in the magazine about Watson a while back: <a href=“http://www.nytimes.com/2010/06/20/magazine/20Computer-t.html[/url]”>http://www.nytimes.com/2010/06/20/magazine/20Computer-t.html</a> They talk a bit about how he makes his choices and why he sometimes go wrong.</p>
<p>IBM programmers should have given Watson a female personality, like the SAL-10000 supercomputer in the movie ‘2010,’ the sequel to 2001: A Space Oddesey.</p>
<p>It’s clear that Watson has been programmed to play the “Chuck Forrest Bounce” strategy - a style of play named for a great champion from the early days who originated it. It’s where you randomly jump from category to category rather than methodically working your way through each category. The theory is that when you are in control of the board you play the bounce so that only you know which category is coming next, which (hopefully) keeps your opponents always slightly off-balance.</p>
<p>Back in the day, I didn’t like to play the the bounce strategy, if for no other reason than it makes for such an ugly, missing-teeth looking board.</p>
<br>
<br>
<p>Or the Star Trek computer, which was voiced by Mrs. Gene Roddenberry - the same actress who also played Nurse Chapel in the original Star Trek series.</p>
<p>It’s also clear that, 99% of the time, Watson can simply buzz in faster than a human being.</p>
<p>H & I figure the IBM researchers are in seventh heaven. Being a bit on the nerdy side ourselves, we understand … it must be AWESOME to have your work pay off this way!</p>
<p>We know a couple of folks who really were ON Jeopardy. They said…the hardest part is the “buzzing in” timing. Watson clearly is programmed to buzz in…and quickly. I have to believe those other two contestants know most of the answers…heck…WE knew a lot of them. They just can’t buzz in as quickly as Watson.</p>
<p>I am enjoying this Infomercial for IBM…and yes…that is what it is. BUT it’s actually quite fascinating. I’m looking forward to tomorrow.</p>
<br>
<br>
<p>This is correct. Due to all the tests and screening, to get on the show in the first place, nearly all players are pretty knowledgeable. In a regular Jeopardy game I’d estimate that for about 75-80% of the clues ALL THREE players know the correct response. What differentiates a big champion from a first round loser is often a matter of button timing.</p>
<p>When I played I had pretty good timing in the regular season and Tournament of Champions. But when I got called back for a special tournament 11 years later I got crushed on the the first round. I still had all my knowledge but my button timing was off. You don’t have to be off by much, just fractions of a second, to turn yourself from a serious Jeopardy animal into a first round corpse.</p>
<p>The IBM guys in the audience crack me up. They get so excited when their “boy” gets one right.</p>
<p>I watched Round 2 last night and became bored after a few minutes, mostly I think because of the “buzzer issue.” Like someone else said, it’s all a matter of timing. And if that’s the case what is Watson truly accomplishing? Also, the first night infomercial didn’t bother me but last night the ad naseum “IBM atop Mount Olympus” horn-tooting was annoying.</p>
<p>My H works in the human-computer interaction field (some of his colleagues work with Watson) and has worked on the “buzzer” issue. Apparenlty, Watson is programmed to hesitate a fraction of an instant depending on his confidence level. For example, if its confidence level is 98%, there is almost no way a human can beat it to the buzzer. But as confidence level goes down, there is a little more wiggle room for humans. After all, we are subject to sweaty fingers, twtitchy muscles, etc. Watson presses the buzzer at the same rate every time.</p>
<p>Once they got Watson good enough at understanding the questions that it could answer most of them correctly, a huge advantage shifts to the computer because of its advantages with the button.</p>
<p>As someone stated earlier there are lights on either side of the category board that illuminate to signal the players that Alex has officially finished reading the clue and the button are enabled. But the button enabling is not done by some machine. It’s done by a human sound engineer who flips a switch when he hears the last syllable of Alex reading the clue. I and most other Jeopardy champions I know never paid much attention to the activation lights. In fact in the heat of the battle I didn’t even see the lights. Instead you just try to get “in synch” with the rythmn of the guy who is flipping the switch - hitting your button a split second after he does. Sort of Zen button pushing - “Be the Button.” If you wait to see the lights you will usually be too late. </p>
<p>A very common penalty that humans suffer is ringing in too soon, which inactivates the button for half a second. And half a second is an eternity in Jeopardy land. Watson, on the other hand, never jumps in too soon. He is responding to the first electrons that signal that the buttons are activated. Huge advantage to Watson.</p>
<p>One thing that human players have to adapt to is when they apparently change the person activating the buttons. You can come back after lunch and suddenly the rythmn has changed. What used to be a successful ring-in is now too soon or too late. Watson adapts to this instantly and automatically. Human players struggle with this all the time. That’s why you sometimes see huge shifts in momentum from one player to another - the timing on the button activation has changed slightly and now the other guy has the better hang of it.</p>
<p>It looked like Ken Jennings got the buzzer thing down tonight. I really do love watching him…he is so bright…and witty. Watson was fun but I was glad to see he didn’t totally swamp the humans tonight.</p>
<p>It was fun to watch…and it is fascinating.</p>
<p>coureur
I appreciate your insights. I hope you won a lot</p>
<p>Watson won</p>
<p>[This</a> is Jeopardy! - Carnegie Mellon University](<a href=“http://www.cmu.edu/homepage/computing/2011/winter/watson.shtml]This”>This is Jeopardy! - Carnegie Mellon University | CMU)</p>
<p>[Watson</a> triumphs in Jeopardy’s man vs. machine challenge - Computerworld](<a href=“Watson triumphs in Jeopardy's man vs. machine challenge | Computerworld”>Watson triumphs in Jeopardy's man vs. machine challenge | Computerworld)</p>
<p>
</p>
<p>thanks! I knew he got the clues electronically, but I wasn’t aware of the lights and how he was “plugged in” to the actual game.</p>
<p>To make a fair match, they need to put more delay/randomness on Watson’s buzzer.</p>
<p>Also, there are certain types of questions that are harder for Watson to answer. They need to include more of “those” questions!</p>
<p>I was amazed at the speed Watson could find the answer. But it’s no fun to watch a computer play–no nerves, no personality, no sense of humor, no conversation. . .</p>
<p>I saw coureur’s explanation play out clearly tonite when Ken had great button timing in single Jeopardy, but then experienced great frustration trying to buzz in during double Jeopardy, when the lights person must have changed. </p>
<p>Was anyone else totally tickled by the Simpsons reference Ken made in his Final Jeopardy response? I just love the guy! (Can still make myself laugh out loud by thinking about his famous “ho” response.)</p>
<p>I would have found Watson much less annoying if he had been been given a sexy female (Marilyn Monroe) or sexy male (Sean Connery) voice. Guess the IBM’ers don’t have a sense of humor.</p>
<p>Anyone have a theory on Watson’s odd betting amounts?</p>
<p>Brother forwarded this: [Jeopardy</a>! Champ Ken Jennings - The Washington Post](<a href=“http://live.washingtonpost.com/jeopardy-ken-jennings.html?hpid=talkbox1]Jeopardy”>http://live.washingtonpost.com/jeopardy-ken-jennings.html?hpid=talkbox1)</p>
<br>
<br>
<p>To clarify, when a player loses the ability to time the button just right, it’s not always due to a change in the person hitting the activation switch. It’s often due to the player himself losing his timing. You see that happen sometimes when a player gets rattled by missing a big Daily Double or something or gets ruled wrong on a debatable answer. Another advantage to Watson - he never gets distracted or upset.</p>
<p>Where players especially suspect a change of the engineer activating the button is when there is a noticeable change in the button timing from one game to the next. You might find yourself thinking “Hey, in the last game if I rang in right on the last syllable of Alex’s last word I was hitting it just right. Now I’m consistently jumping in too soon. Now I have to wait a half a beat. Something has changed.” In regular Jeopardy they tape five games, a whole week’s worth, in one day - about 30 minutes apart and with a lunch break between games three and four. If you are the defending champion you rush back to the Green Room between games to change to new clothes and come out and pretend it’s the next day.</p>