Announcement: Be excellent to each other.


Caravel Forum : Other Boards : Forum Games : Puzzle "tag" (Don't forget to read the rules in the first post!)
<<64656667
Page 68 of 88
69707172>>
New Topic New Poll Post Reply
Poster Message
stigant
Level: Smitemaster
Avatar
Rank Points: 1182
Registered: 08-19-2004
IP: Logged
icon Re: Puzzle "tag" (0)  
Well, first of all, Beethro may have secretly known everyone had blue eyes and threw in the green for fun.

Think of the implications! If everybody had blue eyes and he said "Blue and Green eyes" and everybody took that to mean that there was at least 1 person with green eyes, each person would immediately conclude (incorrectly) that they were the ones with green eyes. The next morning, everybody would show up to commit sepuku and as they were dying, and saw everybody else around them with blue eyes dying they would realize their mistake. Only to then conclude that they would have to commit sepuku the next day instead.

How sad.

____________________________
Progress Quest Progress
03-27-2008 at 05:50 PM
View Profile Send Private Message to User Show all user's posts Quote Reply
TripleM
Level: Smitemaster
Rank Points: 1379
Registered: 02-05-2003
IP: Logged
icon Re: Puzzle "tag" (0)  
Really really easy (found in the same source as my last problem):

You've been granted magical powers to become unbeatable at tennis. But you're in the final against Federer, and you know that your powers won't last the whole game. What do you want the score to be when your powers leave you?
03-28-2008 at 12:13 AM
View Profile Send Private Message to User Show all user's posts Quote Reply
Tahnan
Level: Smitemaster
Avatar
Rank Points: 2460
Registered: 11-14-2005
IP: Logged
icon Re: Puzzle "tag" (0)  
6-0, 6-0?
03-28-2008 at 03:30 AM
View Profile Send Private Message to User Show all user's posts High Scores This architect's holds Quote Reply
stigant
Level: Smitemaster
Avatar
Rank Points: 1182
Registered: 08-19-2004
IP: Logged
icon Re: Puzzle "tag" (0)  
I guess I would want the score to be 5 games to love in the third set (me having won the first 2 sets) and the game score would be 40-love, with me serving since that would give me an immediate 3 chances to win the match with a single point followed by 4 chances to win the match with a single game, followed by some tiebreaker opportunities, and then 2 more sets to get a single set.

Or maybe I would want it to be 40-30, 5-4 Federer in the final set so that he'd put me out of my misery quickly. I mean, if I lost in this scenario, it would be because Federer is pretty darn good. But if I still managed to lose the scenario above, I'd be pretty darn embarrassed.

6-0, 6-0?
The scores in the first 2 sets matter little as long as you won them.

____________________________
Progress Quest Progress

[Last edited by stigant at 03-28-2008 03:31 AM]
03-28-2008 at 03:30 AM
View Profile Send Private Message to User Show all user's posts Quote Reply
TripleM
Level: Smitemaster
Rank Points: 1379
Registered: 02-05-2003
IP: Logged
icon Re: Puzzle "tag" (0)  
Nope.
03-28-2008 at 03:45 AM
View Profile Send Private Message to User Show all user's posts Quote Reply
golfrman
Level: Master Delver
Avatar
Rank Points: 165
Registered: 07-10-2006
IP: Logged
icon Re: Puzzle "tag" (0)  
40-love?

____________________________
Jesus is Lord and Messiah!
03-28-2008 at 04:52 AM
View Profile Send Private Message to User Show all user's posts Quote Reply
Tahnan
Level: Smitemaster
Avatar
Rank Points: 2460
Registered: 11-14-2005
IP: Logged
icon Re: Puzzle "tag" (0)  
I'm not clear on this. I know how I play tennis. Once my magical unbeatable power leaves me, I will lose every subsequent point, period, and at that point it won't matter what the score was up to that point. So, er, what?

Edit: hey, wait, I take it back--can I be playing this Federer?

[Last edited by Tahnan at 03-28-2008 07:27 AM]
03-28-2008 at 07:25 AM
View Profile Send Private Message to User Show all user's posts High Scores This architect's holds Quote Reply
TripleM
Level: Smitemaster
Rank Points: 1379
Registered: 02-05-2003
IP: Logged
icon Re: Puzzle "tag" (0)  
Even you have a nonzero probability of winning each point :P Federer could double fault, for example.

Afraid not.. its the real Roger Federer.

So.. no takers?
03-30-2008 at 03:00 AM
View Profile Send Private Message to User Show all user's posts Quote Reply
stigant
Level: Smitemaster
Avatar
Rank Points: 1182
Registered: 08-19-2004
IP: Logged
icon Re: Puzzle "tag" (0)  
Well, am I totally off base? Or is there some other situation, similar to the one I outlined but with a slightly higher probability of me winning?

____________________________
Progress Quest Progress
03-30-2008 at 03:07 AM
View Profile Send Private Message to User Show all user's posts Quote Reply
TripleM
Level: Smitemaster
Rank Points: 1379
Registered: 02-05-2003
IP: Logged
icon Re: Puzzle "tag" (0)  
There is another situation. Which gives you a much higher probability of winning.
03-30-2008 at 03:15 AM
View Profile Send Private Message to User Show all user's posts Quote Reply
zex20913
Level: Smitemaster
Avatar
Rank Points: 1723
Registered: 02-04-2003
IP: Logged
icon Re: Puzzle "tag" (+1)  
The only situation I can think of is being ahead 6-X, 6-X, 6-6 with tiebreak in your favor at 5-0. Of course, "much higher percentage" is only relative here.

____________________________
Click here to view the secret text

03-30-2008 at 12:34 PM
View Profile Send Private Message to User Send Email to User Show all user's posts Quote Reply
stigant
Level: Smitemaster
Avatar
Rank Points: 1182
Registered: 08-19-2004
IP: Logged
icon Re: Puzzle "tag" (+2)  
Actually, 6-0 in the tie break since tiebreakers are to 7 points, and that would give you 6 single-point options for victory. But then you don't have any single-game options for victory. Depending on your un-aided chances of winning a game, it might be slightly better to end up in the situation that I described. Besides, if you could get to 6-6 in the third set with your powers in tact, couldn't you have gotten to 6-0 with 6 more games worth of power left over?

____________________________
Progress Quest Progress
03-30-2008 at 06:02 PM
View Profile Send Private Message to User Show all user's posts Quote Reply
TripleM
Level: Smitemaster
Rank Points: 1379
Registered: 02-05-2003
IP: Logged
icon Re: Puzzle "tag" (0)  
Thats the answer. The chances of not losing 6 straight points far outweigh the chances of being able to win an entire game in the rest of the set - I'll leave it as an ObPuzzle for something to calculate more precise probabilities. (At least for anyone other than world class players.)

[Last edited by TripleM at 03-30-2008 10:32 PM]
03-30-2008 at 10:32 PM
View Profile Send Private Message to User Show all user's posts Quote Reply
Rabscuttle
Level: Smitemaster
Avatar
Rank Points: 2518
Registered: 09-10-2004
IP: Logged
icon Re: Puzzle "tag" (0)  
Tennis scoring is screwed up. I thought that games (or sets or whatever) went until someone was above the maximum points and two thingies up) That's why you get 5-7s and 6-8s.

Although I think I've also seen 6-5s. So, I have no idea. Except that tennis scoring is just plain nuts.
03-31-2008 at 01:21 AM
View Profile Send Private Message to User Show all user's posts High Scores This architect's holds Quote Reply
zex20913
Level: Smitemaster
Avatar
Rank Points: 1723
Registered: 02-04-2003
IP: Logged
icon Re: Puzzle "tag" (0)  
If it's the fifth set (or maybe just match point), you have to win by 2. Otherwise, you can win by getting to 7.

But I agree. 0-15-30-40. Who on earth came up with that scoring system?

____________________________
Click here to view the secret text

03-31-2008 at 01:37 AM
View Profile Send Private Message to User Send Email to User Show all user's posts Quote Reply
stigant
Level: Smitemaster
Avatar
Rank Points: 1182
Registered: 08-19-2004
IP: Logged
icon Re: Puzzle "tag" (0)  
I believe the point-scoring system was originally based on the minutes around a clock: 0 (12 0'clock), 15 (3:00), 30 (6:00), 45 (9:00). 45 has 3 syllables, so it got shortened to 40.

The last time I checked the rules (which admitedly was a while ago), tie breakers would be played at 6-6 in any set except the 5th (or 3rd set if you're playing best of 3). If the score is 6-5 then you play another regular game. If the person with 6 wins it, then they win the set 7-5. Otherwise, the score will be 6-6, and a tiebreaker is played. To win the tiebreaker, you must get to 7 with a lead of 2. So tiebreakers could go on inevitably, but in practice usually don't go past 10-8 or so. In the 5th set, you must get to 6 games with a lead of 2, so things can go on for a long time if nobody manages to break serve (or nobody manages to hold serve), but in practice, scores very rarely exceed 10-8 again (though this time we're counting games instead of points).

____________________________
Progress Quest Progress

[Last edited by stigant at 03-31-2008 10:11 PM]
03-31-2008 at 10:02 PM
View Profile Send Private Message to User Show all user's posts Quote Reply
stigant
Level: Smitemaster
Avatar
Rank Points: 1182
Registered: 08-19-2004
IP: Logged
icon Re: Puzzle "tag" (+2)  
The chances of not losing 6 straight points far outweigh the chances of being able to win an entire game in the rest of the set
Actually, with quite a bit of analysis, (which I'm happy to post but it will take some time to type it up into a form that will make sense to anybody else), this is not quite true.

I'll give the executive summary. I reduced the problem down to this: suppose your chance of winning a point off of Federer is X (0<X<1). To simplify things, we'll assume that X is unaffected by who's serve it is or other factors. Using Markov chains, we can find the probability of winning the match from any current state of the game. This is where the analysis is a bit complicated, becuase the state-diagram has cycles in it. I resolved the cycles by hand, and then I used a nice little vector-based drawing program (Geometer's Sketchpad, if anybody is wondering) to quickly, but implicitly (rather than explicitly... this is why I can't really post my results right now) solve for the probability of winning a game from 40-0, or 0-0 (or any other state of the game) then a tiebreaker from 6-0, and, using the probabilities for winning from 0-0, a set from 5-1 (actually, I ignored the game-deuce situation to simplify things a bit, but this underestimates your chances in my suggested situation, so it doesn't change things substantially). Ultimately the result is that if x > .45 (about, I'm eyeballing the intersection point) then my situation is slightly better than Triple-M's. Now, one might argue that if you have a 45% chance of taking a point off of Federer, you'd be a pretty good tennis player, and you'd probably be correct. However, by my calculations, the chances of taking a single game off Federer when your point-probability is .45 is less than 1/3. And the probability of taking a match off Federer is substantially less than 1%. Now, while Federer's winning percentage is very very good, its not THAT good (even throwing out his losses to Nadal on clay), implying that, on average, a professional tennis player has a better than 45% chance of winning a single point against Federer.

____________________________
Progress Quest Progress

[Last edited by stigant at 03-31-2008 11:23 PM]
03-31-2008 at 11:20 PM
View Profile Send Private Message to User Show all user's posts Quote Reply
RoboBob3000
Level: Smitemaster
Avatar
Rank Points: 1982
Registered: 10-23-2003
IP: Logged
icon Re: Puzzle "tag" (0)  
stigant wrote:
The chances of not losing 6 straight points far outweigh the chances of being able to win an entire game in the rest of the set
Actually, with quite a bit of analysis, (which I'm happy to post but it will take some time to type it up into a form that will make sense to anybody else), this is not quite true.
You know, I thought this might be the case, but was too terrified of the work involved to get to the answer. Well played! :thumbsup

____________________________
http://beepsandbloops.wordpress.com/

[Last edited by RoboBob3000 at 04-01-2008 01:14 AM]
04-01-2008 at 01:13 AM
View Profile Send Private Message to User Show all user's posts This architect's holds Quote Reply
TripleM
Level: Smitemaster
Rank Points: 1379
Registered: 02-05-2003
IP: Logged
icon Re: Puzzle "tag" (0)  
Ah, I was planning to do something similar to that at one point to find that exact value.

I guess my original question wasn't entirely clear; when I said 'you' I meant you :P The chances of a person who has a 45% chance of taking a point off Federer reading this puzzle were around 0, in my opinion ;)

So, yes, professional players would indeed prefer to be up 5-0 in the third set. Anyone else should take the tiebreak option.

(I think you've earned the right to post a puzzle now :))

[Last edited by TripleM at 04-01-2008 01:15 AM]
04-01-2008 at 01:15 AM
View Profile Send Private Message to User Show all user's posts Quote Reply
TripleM
Level: Smitemaster
Rank Points: 1379
Registered: 02-05-2003
IP: Logged

File: tennis.cpp (3.3 KB)
Downloaded 56 times.
License: Public Domain
icon Re: Puzzle "tag" (0)  
I've run my own calculations and get some different results. I've included deuce and all the normal rules of tennis.

My cutoff probability is 0.414. I'm not quite sure how you worked out the chance of winning a game/set/match for a certain probability Stigant, but my results for that part are vastly different - with a probability of 0.414 of winning a point, I make that 0.294 chance of winning a game, 0.252 chance of winning a set, and 0.106 chance of winning a match. There are a few players who can win 11% of matches against Federer, but not a huge deal. This seems much higher than 'much less than 1%' you came up with.

Commented C++ code attached, I don't think there's any mistakes in there.
04-01-2008 at 07:14 AM
View Profile Send Private Message to User Show all user's posts Quote Reply
stigant
Level: Smitemaster
Avatar
Rank Points: 1182
Registered: 08-19-2004
IP: Logged
icon Re: Puzzle "tag" (0)  
ok, you're getting a (slightly... 4% isn't much) lower cutoff probability because when I ran the numbers on my answer, I assumed that if Federer managed to scrape back to 5-5 in the third set, that he was a lock to win that set. This is not actually the case, so I underestimated the overall probability of winning the match in that scenario (as I mentioned in the writeup). My ultimate goal was to see if there was a cuttoff probability at all, not so much to calculate the exact point. Any probability below 50% would have justified my original comment that you can't know which situation is better without looking at the numbers. (cutoff probabilities above 50% would be meaningless since the tacit assumption here is that Federer is the best player in the world)

With a point probability of .414, I also calculate a game probability of .294, so I think we're essentially doing the same work there. I haven't actually run the set probability (for reasons explained above). The 1% was an estimate based on previous work that I've seen done in this vein. I expect though, that .252 is extremely high for this number considering there's only a .294 chance of winning each game. I'll work on the analysis in a minute.

EDIT: Ok, with a probability of .414 of winning a point, I get a probability of .0457 of winning the set. I'm pretty sure that will result in less than a 1% chance of winning the match (there are no cycles in the state diagram for sets, so you're pretty much doing straight multiply and add probability from here)

EDIT2: With a .414 probabilty of winning a GAME, I calculate a .24 probability of winning a set.

EDIT3: I just re-ran the cutoff numbers with my improved set-probability estimate and I get essentially the same answer I got before for the cutoff probability.

What all this shows is that to have even a remote chance to win a match against Federer, or for that matter anybody who is better than you, you must be extremely close in per-point ability.

I don't have another puzzle ready to go out, so if anybody else wants to post one, please do.

____________________________
Progress Quest Progress

[Last edited by stigant at 04-01-2008 02:22 PM]
04-01-2008 at 01:33 PM
View Profile Send Private Message to User Show all user's posts Quote Reply
stigant
Level: Smitemaster
Avatar
Rank Points: 1182
Registered: 08-19-2004
IP: Logged

File: tennis.scm (8 KB)
Downloaded 56 times.
License: Public Domain
icon Re: Puzzle "tag" (0)  
Here's a scheme program I wrote to check my GSP results (which were a bit off, it seems). It takes into account everything I know about tennis scoring (tie breakers in the 1st 4 but not the 5th set etc)

You'll need to download Dr Scheme to run this. Make sure your language level is set to Pretty Big (this can be set in the language menu)

Executive summary:
Cutoff probability: 0.414013671875 (confirming TripleM's result)
Prob of winning a set with tiebreaker at this pt-prob: 0.06163409134669725 (contradicting TripleM's result for a single set)
Prob of winning 5th set: 0.05764230279357295 (also contradicting TripleM's result for a single set)
Prob of winning match: 0.002050096243150496 (confirming my estimate of less than 1% chance of beating Federer straight up)
Pro cutoff: 0.46202392578125 (this is the per-pt percentage you'd have to win agains Federer in order to have a 10% chance of beating him)

____________________________
Progress Quest Progress
04-01-2008 at 05:28 PM
View Profile Send Private Message to User Show all user's posts Quote Reply
stigant
Level: Smitemaster
Avatar
Rank Points: 1182
Registered: 08-19-2004
IP: Logged
icon Re: Puzzle "tag" (0)  
Ok, here's a puzzle from another site that I solved last night. Its a bit mathy, but I'm sure you'll enjoy the whimsy of it:

Alice has a drawer full of identical black socks and identical white socks. She chooses two socks at random from the drawer. The probability that they match is 1/2. How many socks of each color are in the drawer? (There is more than one correct answer.)

(a complete solution should A) characterize all the solutions and B) demonstrate/prove that the answer is correct. I'll accept A, but the more mathy out there will want to try for B as well).



____________________________
Progress Quest Progress
04-01-2008 at 07:46 PM
View Profile Send Private Message to User Show all user's posts Quote Reply
Wolfcastle
Level: Master Delver
Rank Points: 117
Registered: 02-27-2006
IP: Logged
icon Re: Puzzle "tag" (+1)  
x:=number of black socks, y:=number of white socks
x/(x+y) chance of choosing a black sock (first draw)
(x-1)/(x+y-1) chance of choosing a black sock (second draw) if first sock was black as well
same for y and white socks, that gives;
x/(x+y)*(x-1)/(x+y-1)+y/(x+y)*(y-1)/(x+y-1)=0.5 or
(x^2-x+y(y-1))/((x+y)*(x+y-1))=0.5 that is approx
y=1.41421*(\sqrt(x+.125)+.707107*(x+.5))

for example (1 and 3) or (3 and 6) or (6 and 10) should work

Edit: looking at how the soultions develope
f(i)=sum_(n=0)^(i) n
f(i) and f(i+1) should work (for i>0), proof needed

[Last edited by Wolfcastle at 04-01-2008 08:46 PM]
04-01-2008 at 08:39 PM
View Profile Send Private Message to User Show all user's posts Quote Reply
Jutt
Level: Smitemaster
Rank Points: 863
Registered: 06-29-2007
IP: Logged
icon Re: Puzzle "tag" (+2)  
Here's a different approach:

Let b := # black socks, w := # white socks.

# possibilities to draw 2 black socks: b choose 2 = ½b² – ½b
# possibilities to draw 2 white socks: b choose 2 = ½w² – ½w
# possibilities to draw black/white pair: bw

That means we need to solve: ½b² – ½b + ½w² – ½w = bw.
Some algebraic manipulations:

½w² – (½+b)w + ½b² – ½b = 0
w² – (2b+1)w + b² – b = 0
With the abc-formula we find:
w = ( 2b+1 ± √((2b+1)² + 4b² + 4b) ) / 2
w = ( 2b+1 ± √(8b+1) ) / 2
w = b + ½ ± ½√(8b+1)

Now to get an integer solution for w, 8b+1 must be a square number. Let's call its root x.
Since 8b+1 is odd, x must be odd as well. The other way around, if x is odd, x² will be always 1 mod 8, so any odd x will suffice.
Finally we can write out b and w in terms of x.
From x = √(8b+1) we obtain that
b = ⅛(x²–1) and
w = ⅛(x²–1) + ½ ± ½x,
for any odd x > 1 (x=1 gives the unwanted solutions b=0, w=0 and b=0, w=1).

Note: substituting x with 2n+1 should lead to the formula which Wolfcastle decribed, but had no proof for yet.

____________________________
Holds: An Architects Audition, Artful Architecture, Salamander, Elusive Exhibitions, Leftover Levels, Six Times Six
Collaborative: Way Forward, Advanced Concepts 2
Styles/Mods: Basalt, Sandstone, Garden, Clock using game elements
04-01-2008 at 09:59 PM
View Profile Send Private Message to User Send Email to User Visit Homepage Show all user's posts High Scores This architect's holds Quote Reply
stigant
Level: Smitemaster
Avatar
Rank Points: 1182
Registered: 08-19-2004
IP: Logged
icon Re: Puzzle "tag" (0)  
That's correct! Well done

____________________________
Progress Quest Progress
04-01-2008 at 10:52 PM
View Profile Send Private Message to User Show all user's posts Quote Reply
TripleM
Level: Smitemaster
Rank Points: 1379
Registered: 02-05-2003
IP: Logged
icon Re: Puzzle "tag" (0)  
stigant wrote:
Here's a scheme program I wrote to check my GSP results (which were a bit off, it seems). It takes into account everything I know about tennis scoring (tie breakers in the 1st 4 but not the 5th set etc)

You'll need to download Dr Scheme to run this. Make sure your language level is set to Pretty Big (this can be set in the language menu)

Executive summary:
Cutoff probability: 0.414013671875 (confirming TripleM's result)
Prob of winning a set with tiebreaker at this pt-prob: 0.06163409134669725 (contradicting TripleM's result for a single set)
Prob of winning 5th set: 0.05764230279357295 (also contradicting TripleM's result for a single set)
Prob of winning match: 0.002050096243150496 (confirming my estimate of less than 1% chance of beating Federer straight up)
Pro cutoff: 0.46202392578125 (this is the per-pt percentage you'd have to win agains Federer in order to have a 10% chance of beating him)

Yup, found one little mistake in my code. I come up with the same results.
04-02-2008 at 01:18 AM
View Profile Send Private Message to User Show all user's posts Quote Reply
marphlets
Level: Goblin
Avatar
Rank Points: 17
Registered: 03-31-2008
IP: Logged
icon Re: Puzzle "tag" (0)  
WOW! We certainly have some "math-y" people around here. :lol Is there a puzzle forum for people who made B's in College Algebra?

Click here to view the secret text


____________________________
I eat ogres for breakfast. Really large tooth picks.

[Last edited by marphlets at 04-04-2008 07:59 AM]
04-04-2008 at 07:59 AM
View Profile Send Private Message to User Show all user's posts Quote Reply
Blondbeard
Level: Smitemaster
Avatar
Rank Points: 1486
Registered: 03-31-2005
IP: Logged
icon Re: Puzzle "tag" (+1)  
stigant wrote:
That's correct! Well done

I think an even easier way of putting the answer is that
b or w = ½(x²–x)
w or b = ½(x²+x)

Where x is any positive integer except 1. It's easy to see that this solution is correct because bw/((b+w)*(b+w-1))=0.25

[Last edited by Blondbeard at 04-06-2008 01:03 PM]
04-06-2008 at 01:02 PM
View Profile Send Private Message to User Show all user's posts This architect's holds Quote Reply
stigant
Level: Smitemaster
Avatar
Rank Points: 1182
Registered: 08-19-2004
IP: Logged
icon Re: Puzzle "tag" (0)  
Yeah, that was the way I phrased it in my solution.

____________________________
Progress Quest Progress
04-07-2008 at 01:23 PM
View Profile Send Private Message to User Show all user's posts Quote Reply
<<64656667
Page 68 of 88
69707172>>
New Topic New Poll Post Reply
Caravel Forum : Other Boards : Forum Games : Puzzle "tag" (Don't forget to read the rules in the first post!)
Surf To:


Forum Rules:
Can I post a new topic? No
Can I reply? No
Can I read? Yes
HTML Enabled? No
UBBC Enabled? Yes
Words Filter Enable? No

Contact Us | CaravelGames.com

Powered by: tForum tForumHacks Edition b0.98.9
Originally created by Toan Huynh (Copyright © 2000)
Enhanced by the tForumHacks team and the Caravel team.