Wizard
Administrator
Wizard 
  • Threads: 1491
  • Posts: 26435
Joined: Oct 14, 2009
July 13th, 2010 at 10:35:41 AM permalink
Somebody asked me about Paul the Octpus' record at handicapping the World Cup. I've also heard mention of him several times on the news.

Paul's record is 12 correct picks and 2 incorrect picks. The probability of getting exactly 12 out of 14 correct picking randomly is combin(14,12)*(1/2)^14 = 0.56%. The probability of getting 12 or more correct out of 14 is (1+14+combin(14,2))*(1/2)^14 = 0.65%. He was not given the choice of picking a tie, and there never was one in the games he handicapped. I'm not sure how his record would have been depicted if there were any ties, but I suspect it wouldn't have been included.

My strong opinion is this is just dumb luck. I also finding it a sad commentary that this is getting as much publicity as it is. I know Paul was mentioned on the ABC World News at least twice. He probably got more coverage than some ongoing civil wars in Africa. I'm thinking of doing the same experiment with my dog the next football season. If he does badly, nobody will care, but think of the free publicity he might get if he falls on the right tail of the bell curve.
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
Nareed
Nareed
  • Threads: 373
  • Posts: 11413
Joined: Nov 11, 2009
July 13th, 2010 at 10:53:14 AM permalink
Quote: Wizard

He was not given the choice of picking a tie, and there never was one in the games he handicapped. I'm not sure how his record would have been depicted if there were any ties, but I suspect it wouldn't have been included.



I started hearing about this animal only when the second round started. In the WC the second round, quarter-finals, semi-finals and finals are playoff games: someone has to win. If they're tied in regulation play, they do two overtime periods of 15 minutes each. If they're still tied, they do series of five penalty shots each until one team comes out ahead. There was at least one such tie, but the official result is X team won 5-4 and team Y lost. the score sometimes is given with an asterisk indicating penalty shots, and someitmes it's written like X 1 (5)-1 (4) Y

Quote:

My strong opinion is this is just dumb luck.



What else could it be? I did hear the "predictions" in advance, so at least it wasn't a hoax.

I'm sure dozens of other people had other animals try to pick winners and didn't get so many results right.

Quote:

I'm thinking of doing the same experiment with my dog the next football season. If he does badly, nobody will care, but think of the free publicity he might get if he falls on the right tail of the bell curve.



Great idea. But maybe you should restict it to the playoffs, when there are less games and no ties. You can tie in Football, it's just very rare.

What kind of dog do you have? A nice, telegenic pooch would be best.

Better yet, get together with other dog owners and ahve each dog predict some games. Get enough dogs and one is bound to be "right" much of the time.
Donald Trump is a fucking criminal
Wizard
Administrator
Wizard 
  • Threads: 1491
  • Posts: 26435
Joined: Oct 14, 2009
July 13th, 2010 at 11:08:10 AM permalink
Quote: Nareed


What kind of dog do you have? A nice, telegenic pooch would be best.





I have a Havanese.

If I do this I'll state that in the unlikely event of a tie then it won't count. I'll also have him pick against the spread. I suppose I could trick him into picking likely favorites straight up, as may have been a factor with Paul, but I might have my actuarial membership revoked.
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
boymimbo
boymimbo
  • Threads: 17
  • Posts: 5994
Joined: Nov 12, 2009
July 13th, 2010 at 11:27:23 AM permalink
Mr. Wizard, breathe. Take a deep breath, then breathe again.

You know that your dog cannot predict football games. Neither can coins, rocks, mountains, fish, mice, rodents, birds, marsupials, dolphins, or apes.

You're already famous enough. Don't bring your pretty dog into it.
----- You want the truth! You can't handle the truth!
Wizard
Administrator
Wizard 
  • Threads: 1491
  • Posts: 26435
Joined: Oct 14, 2009
July 13th, 2010 at 11:32:19 AM permalink
Quote: boymimbo


You're already famous enough. Don't bring your pretty dog into it.



Fame is addictive; you can never have enough.
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
boymimbo
boymimbo
  • Threads: 17
  • Posts: 5994
Joined: Nov 12, 2009
July 13th, 2010 at 11:44:40 AM permalink
So is gambling.
----- You want the truth! You can't handle the truth!
Nareed
Nareed
  • Threads: 373
  • Posts: 11413
Joined: Nov 11, 2009
July 13th, 2010 at 11:45:35 AM permalink
Quote: Wizard

I have a Havanese.



Cute. Just make sure he doesn't show teeth in the official photo (unless the picking method involves the dog chewing the hell out of a miniature team mascot).

Quote:

If I do this I'll state that in the unlikely event of a tie then it won't count. I'll also have him pick against the spread. I suppose I could trick him into picking likely favorites, as may have been a factor with Paul, but I might have my actuarial membership revoked.



Ties are so rare in the NFL some players don't even know they exist. Whole seasons can go by without a single tie.

Trickery can be accomplished easily and surreptitiously. Just put a morsel of his favorite food near/on the favorite team. Or have your wife or kids handle the dog while you just record the results.
Donald Trump is a fucking criminal
konceptum
konceptum
  • Threads: 33
  • Posts: 790
Joined: Mar 25, 2010
July 13th, 2010 at 3:07:10 PM permalink
My suggestion is thus:

1) Take pictures of your dog in positions where it looks like he is reading a) mathematical/probability texts, b) handicapping texts and/or websites, and/or c) sports teams and players information/websites.

2) Have dog pick sports teams to win. Maybe the first week of NFL results.

3) Send pick results, along with photographs, to mailing list of thousands of people. Do not send every game, but just rather one game results to each person, but randomly select who gets which game results. Keep track of who gets what results.

4) For the games the dog "picks" correctly, repeat picks for next weeks games. Send to those who got correct results from 1st weeks games.

5) Repeat for a 3rd week.

6) After 3 weeks, you will have presented a certain number of people with your wonder pooch who has correctly picked 3 winning games for the first 3 weeks. For a subsription to the wonder pooch's next picks, please send $25 to ...

7) ?

8) Profit!
Wizard
Administrator
Wizard 
  • Threads: 1491
  • Posts: 26435
Joined: Oct 14, 2009
July 13th, 2010 at 3:32:20 PM permalink
Quote: konceptum

My suggestion is thus:...



Of course, that is what many handicappers do.
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
rxwine
rxwine
  • Threads: 209
  • Posts: 12166
Joined: Feb 28, 2010
July 13th, 2010 at 5:09:57 PM permalink
Quote:

I have a Havanese.



Well, I know dogs play poker. I've seen the painting!
There's no secret. Just know what you're talking about before you open your mouth.
dwheatley
dwheatley
  • Threads: 25
  • Posts: 1246
Joined: Nov 16, 2009
July 13th, 2010 at 5:27:44 PM permalink
That wiki article has a section on how some scientists believe that octopi are attracted to vivid colours (although they are colour-blind) and horizontal shapes. All the prediction the octopus made involved countries with vivid horizontal flags.

Does your dog have a predilection towards certain animal mascots?
Wisdom is the quality that keeps you out of situations where you would otherwise need it
pacomartin
pacomartin
  • Threads: 649
  • Posts: 7895
Joined: Jan 14, 2010
July 13th, 2010 at 6:07:25 PM permalink
I have heard that picking a hedge fund might be better done by your dog. The reality is that hedge funds that are losing money tend to be terminated. So the remained have almost always done better than average. Picking the one with the highest return over several years is statistically a sure way to pick a mediocre performer for next year.
  • Jump to: