Use access key #2 to skip to page content.

RRobertsmith (58.73)

Watson vs Motely Fool dream team



March 25, 2011 – Comments (10) | RELATED TICKERS: IBM

The next IBM challenge:  Watson running Artificial Intelligence Stock Market Trading Software vs a Motely Fool team of TMFUltraLong, TSIF, and kdakota630.  Comment and help define the contest, weekly checkpoints? Is daytrading allowed?  (no!)  What about options? (yes!)  Margin trading allowed?  Should the computer be allowed a human team player/double checker?  (maybe you?)

 Is this like playing poker, where very few poker robots are profitable?

10 Comments – Post Your Own

#1) On March 25, 2011 at 4:40 PM, miteycasey (29.07) wrote:

Poker robots aren't successful because they are first level thinkers and don't have database backends. Put a database on the backend and let the bot make decisions against that and I think bots could beat most ssNLH and msNLH.

Report this comment
#2) On March 25, 2011 at 6:42 PM, kdakota630 (29.48) wrote:

I'm flattered that anyone would consider me part of any "Motley Fool Dream Team", but I don't think I'm in the league of TMFUltraLong and TSIF.

Reminds me of a story...

(I know you're all out there saying, "Jeez, not another story, Jimmy", but it's a good one.)

This is about 10 years ago now, but a friend of a friend of mine played 3 NHL games for the Pittsburgh Penguins, including at least one exhibition game when this story took place.

They had him at centre at the start of the game.  He looked to his left wing and there was Mario Lemieux.  He looked to his right wing and there was Jaromir Jagr.  They happened to be playing the NY Rangers, and he looked up to see that he was facing Wayne Gretzky, and at that moment the only thought going through his mind was:

"Can I have your guys' autographs?"

Report this comment
#3) On March 26, 2011 at 1:21 PM, TSIF (99.97) wrote:

Good story kdakota, we'd probably feel the same way if we won a dinner with Buffett and other respected analysts.  I certainly would have been asking the pros for thier autographs, but I'd like to think I would have waited until we'd leff the ice!

Thanks for the callout Robert, but I think you can do better recruiting your dream team.  Some of us players just jump mostly on high beta stocks for fun and risk is only CAPs points. Some of the better investors were recruited for experiments, Tigerpack is one example, and did quite well.

I get a few extra points because I put so much of my time/CAPs time into timing.  Of course in real life you don't have the seven day rule, so you wouldn't be stuck with a call that turned on unforseen events.  Overall, there are much better players here on CAPS, especially if you look at sectors.

I don't think a Watson would be so well at poker.  I've played a few computer poker games and some try to bluff, but not very well, and you can usually detect a pattern or hole to exploit.  Watson would do better at calculating the odds, but there is enough human influence for very good players to tilt things.

I'm not sure that Watson would do well at stocks either. Computers are VERY good at watching stocks and buying when something moves a certain percent with a certain amount of volume and selling on the reverse. They have no emotion in the trade and I think in this volitile market that strategy computerized does VERY well.  It might fall short when something rises on an earnings day in advance of earnings and then it gaps down due to a disappoint, but it would gain on a gap up scenerio and probably even out.

The market, however has too much phsycology for a computer to stock pick effectively.  World events and volitility we are having now, an oil well blowing, earthquake, middle east, oil, cause and effect, fear, etc. would all be difficult for the computer, but then again, it could probably tell you which company's will suffer most from a particular event.  All in all, computers do decently in a stable market, but there are many more speculators and daytraders around that influence things in the short run. I think humans who like to gamble could beat a computer in the short run.

Good thoughts, and again thanks for the callout on your very first Blog!!!!

Report this comment
#4) On March 26, 2011 at 1:26 PM, TSIF (99.97) wrote:

PS. kdakota630, build a time machine, go back and make exactly the same calls, at exactly the same time, but close your losers when they are negative 100 instead of being "stubborn"....  :)   and you'd have my play blown out of the water.  (of course, I'd borrow your time machine and upcall all the picks you are now down over 100 and maybe TCK and I'd be kicking too!!!  :)

Report this comment
#5) On March 26, 2011 at 1:36 PM, NOTvuffett (< 20) wrote:

The problem with time machines is you eventually run into the Daleks, lol.

Report this comment
#6) On March 26, 2011 at 2:14 PM, TMFUltraLong (99.61) wrote:

No no no...the problem with time machines is you never would have built one in the first place if you didn't have brutally negative picks... therefore... you can never reverse the negative picks by going back and changing them as there would never have been a time machine in the first place if you were never wrong.

Ya catch what I'm dishin'..... =)


Report this comment
#7) On March 26, 2011 at 2:38 PM, kdakota630 (29.48) wrote:


I understand.  It's like the lack of logic between the first two Terminator movies.  The first one is fine, based on man designing computers that got too smart, except that the second movie says that they designed the smart computers based on technology found from the first terminator which could never have happened without that technology that couldn't have been designed without it.


I just wish I had the foresight to stick with my real-life TCK pick.  I forget where I bought it, but I'm pretty sure it was in the $5-7 dollar range, and sold at pretty much the break-even point, I think at a small loss once fees are accounted for.  Too bad because even without my lucky timing with it in CAPS I'd still have been looking at an 8-bagger.

Report this comment
#8) On March 26, 2011 at 3:03 PM, RRobertsmith (58.73) wrote:

You might like this article in the April Scientfic American (pg 19 & 20) about Watson and a company marketing a AI (TSRI) that up dates software (a software programming robot, or agent that updates software).  Makes me feel like I was scooped, or at least like we are working along the same lines.   

Anyone who deals with a ultra large (static or nearly static) database of information will be endangered in the near future, with the advent of quantum computers and Watson style operating systems. 

   This would include doctors, lawyers, and stock traders (fund managers).  But the computer won't have the gut feeling about the amount of risk to take, or the amount of panic to ignore in the news, in order to trade on the news.  Image a doctor like House, with access to Watson and a world view of all risk/reward ratios when treating a patients

Report this comment
#9) On March 26, 2011 at 3:07 PM, TMFUltraLong (99.61) wrote:

Actually I was making a very poor Time Machine (2002) reference....but Terminator logic works too....


Report this comment
#10) On March 27, 2011 at 1:34 PM, TSIF (99.97) wrote:

I think I would rather have a Watson, who might not take the risks to find the disease than a House who hacks you apart piece by piece until sometimes the cure is worse than the disease.

So lets see on the time machine. If I had one, I'd be rich, but if my motive for having one was to get rich, then because I got rich I would not have invented the time machine....hmmmmm...but would I still be rich...and if I was rich, would I screw my life up and wish I was poor....hmmmm....maybe I was rich and I invented the time machine and screwed my life up so I'd be poor....I'll have to remember where I parked it the last time I used it and where I put the key.

Report this comment

Blog Archive

March (1)

Featured Broker Partners