Friday, December 28, 2012 Game Recap & Discussion [SPOILERS]

[dhkendall]

I have caught some criticism for saying this on here (or on the old Sony board) before, but in certain situations I would consider betting $1 on FJ if I had a lock tie. The scenario where I would be most likely to do it is if I had already won 5 or more games, liked the FJ category, and was facing a contestant who I thought was very good and I was lucky to have a huge lead on them. The only time that I would absolutely not do is if it was my 5th game, and I would probably not do it in my 4th game, either. With that said, Paula was not a contestant who I would be afraid to play again, so I would have made the $0 wager here.

I fail to understand, though, why $1 is any different from an all-in bet (or, anything more than double third-place, but for the purposes of this discussion, we'll assume a two-player FJ), since a wrong answer results in the same outcome ($2K and second place) and a right answer means you haven't maximized your winnings (and, of course, in both situations, you didn't go for the guaranteed win). (And, yes, I was probably one of those who gave you the same criticism on the Sony board, mainly because wagering strategy is something I still don't fully understand (but at least I recognized today's situation cold, especially with Alex's help))

[bpmod]

I have caught some criticism for saying this on here (or on the old Sony board) before, but in certain situations I would consider betting $1 on FJ if I had a lock tie. The scenario where I would be most likely to do it is if I had already won 5 or more games, liked the FJ category, and was facing a contestant who I thought was very good and I was lucky to have a huge lead on them. The only time that I would absolutely not do is if it was my 5th game, and I would probably not do it in my 4th game, either. With that said, Paula was not a contestant who I would be afraid to play again, so I would have made the $0 wager here.

I fail to understand, though, why $1 is any different from an all-in bet (or, anything more than double third-place, but for the purposes of this discussion, we'll assume a two-player FJ), since a wrong answer results in the same outcome ($2K and second place) and a right answer means you haven't maximized your winnings (and, of course, in both situations, you didn't go for the guaranteed win). (And, yes, I was probably one of those who gave you the same criticism on the Sony board, mainly because wagering strategy is something I still don't fully understand (but at least I recognized today's situation cold, especially with Alex's help))

Me too.

And, David, congratulations on having exactly the same number of open parentheses as close parentheses.

Brian

[alietr]

Let's see if I can explain it ... You have $10,000 and your other opponent has $5,000 (let's simplify things and say the third contestant is out). Furthermore, both of your chances of getting FJ right is 60% (but ignoring homoscedasticity).

If you bet $0, under any circumstances you go home with $10,000 for that game, so your expected value is $10,000.

If you bet $1:
You're right (60% of the time), you get $10,001 (doesn't matter what your opponent does).
You're wrong (40% of the time) and your opponent is wrong (40% of the time, or 16% overall), you get $9,999.
You're wrong (40% of the time) and your opponent is right (60% of the time, or 24% overall), you get $0.

Your expected value is therefore .6 x $10,001 + .16 x $9,999 + .24 x $0 = $7,600.

And even if you raise your likelihood of getting it right to 95%, your expected value is still lower than $10,000. Why on earth would you ever wager anything small?

Now I will agree that if you raise your likelihood of being right to 95% *and* wager all $10,000, your expected value is $19,000. So you have to weigh the two factors.

[Vanya]

You're assuming second place will bet rationally. If I'm not mistaken there have been many times they didn't.

[bpmod]

You're assuming second place will bet rationally. If I'm not mistaken there have been many times they didn't.

QFT.

Betting rationally from first place is far more likely than betting rationally from second place.

Brian

[Austin Powers]

Let's see if I can explain it ... You have $10,000 and your other opponent has $5,000 (let's simplify things and say the third contestant is out). Furthermore, both of your chances of getting FJ right is 60% (but ignoring homoscedasticity).

If you bet $0, under any circumstances you go home with $10,000 for that game, so your expected value is $10,000.

If you bet $1:
You're right (60% of the time), you get $10,001 (doesn't matter what your opponent does).
You're wrong (40% of the time) and your opponent is wrong (40% of the time, or 16% overall), you get $9,999.
You're wrong (40% of the time) and your opponent is right (60% of the time, or 24% overall), you get $0.

Your expected value is therefore .6 x $10,001 + .16 x $9,999 + .24 x $0 = $7,600.

And even if you raise your likelihood of getting it right to 95%, your expected value is still lower than $10,000. Why on earth would you ever wager anything small?

Now I will agree that if you raise your likelihood of being right to 95% *and* wager all $10,000, your expected value is $19,000. So you have to weigh the two factors.

Now, you are completely correct here, but this is why I can't dismiss StevenH's argument out of hand - it's also the chance of winning TOMORROW that has to come into play. His point is that if you are worried that this person is just better, and you were lucky to be ahead, maybe it's not bad risk to look at the FJ, in isolate, think what your chances are, with the idea being that the payoff would be a win today, plus a great chance of winning tomorrow.

Now, if I am in that situation, I take the automatic win, because I'd assume I could hold my own against most anyone. But if you were playing in the UTOC against Brad Rutter and this situation came up... would you chance facing Rutter again?

[alietr]

Personally, I'd take the money with the win today (over someone whose score I had doubled) against the unknown opponents tomorrow. I think it's ridiculous to chance it.

[Austin Powers]

Personally, I'd take the money with the win today (over someone whose score I had doubled) against the unknown opponents tomorrow. I think it's ridiculous to chance it.

And I think in most situations, yes. The range of contestants is such that, on a day-to-day basis, I don't think any one opponent will be better than the next.

But if I am the challenger, and I'm in that situation against 7 day champ Roger Craig, I might just decide that a 60% chance of winning today plus a decent chance tomorrow (and beyond!) beats a 100% chance of tying today and likely losing tomorrow.

Now, I wouldn't do that, but it's not CRAZY.

[harrumph]

You'd also be playing against someone who has just gotten some buzzer experience in a real game versus someone coming in cold. We've seen plenty of times where a new contestant is outplayed by the defending champ, but wins when the defending champ stumbles in FJ, and in the next game the new champ suddenly has buzzer skills.

[Volante]

"Definition of a classic — something that everybody wants to have read and nobody wants to read." --Mark Twain
Knowing that bon mot pretty much assured me it was Twain who said the FJ quote. No one reads classics, everyone reads Twain, ergo Twain's works aren't classics. Can't get any more Twainian than that!

[StevenH]

You'd also be playing against someone who has just gotten some buzzer experience in a real game versus someone coming in cold. We've seen plenty of times where a new contestant is outplayed by the defending champ, but wins when the defending champ stumbles in FJ, and in the next game the new champ suddenly has buzzer skills.

Yep. I would point to David Madden's first game as a good example. Madden was never a buzzer beast, but he didn't seem to struggle on it in his last 19 games like he did in his first one.

[goforthetie]

And I think in most situations, yes. The range of contestants is such that, on a day-to-day basis, I don't think any one opponent will be better than the next.

But if I am the challenger, and I'm in that situation against 7 day champ Roger Craig, I might just decide that a 60% chance of winning today plus a decent chance tomorrow (and beyond!) beats a 100% chance of tying today and likely losing tomorrow.

Now, I wouldn't do that, but it's not CRAZY.

This might be the one situation in which I wouldn't bet for the tie. Let's say by some miracle (he missed two DDs and I hit a DD in a wheelhouse and made a big bet, along with a few other lucky bounces)I have doubled up on Ken Jennings.

If I bet zero:
~75% chance that we tie. I faced KJ again tomorrow and probably have at most a 10% chance of winning another game, after which I might face 2 new contestants.
~25% chance that he misses FJ. I face 2 new opponents.

Expected # of games won (X = expected number of additional games a returning champ is expected to win if he does not have KJ to deal with):
.75 (1 + .1 + .1 * X) + .25 (1 + X) = 1.075 + .325X

If I bet a non-zero amount:
~75% chance that I win. I face 2 new opponents.
~25% chance that I lose. I become sad.

Expected # of games won:
.75 (1 + X) + .25 (0) = .75 + .75X.

If X = 1 (reasonable assumption) then you win more games on average with the non-tie strategy.

Regular opponent, though? I'm betting zero. The "buzzer warm-up" theory is merely anecdotal.

[cheezguyty]

As for non-zero wagers from lock-tie leads, I know Billy Baxter did it in his College Championship semifinal (which I can't blame him for, since there would have had to be a tiebreaker round otherwise). There's also this game.

There's also this game from 2006 where the leader wagered for the win and second was more than happy to let him have it.

[bpmod]

As for non-zero wagers from lock-tie leads, I know Billy Baxter did it in his College Championship semifinal (which I can't blame him for, since there would have had to be a tiebreaker round otherwise). There's also this game.

There's also this game from 2006 where the leader wagered for the win and second was more than happy to let him have it.

Uh, no. That leader wagered to risk the win. Not for the win, which he already had in hand.

Brian

[cheezguyty]

There's also this game from 2006 where the leader wagered for the win and second was more than happy to let him have it.

Uh, no. That leader wagered to risk the win. Not for the win, which he already had in hand.

Good catch. I meant to say that he wagered for the outright win.

[Bamaman]

At least he didn't spell the title the way I pronounce it.

My thought process on FJ........

How about Thoreau? No, they just used him not too long ago. Let's start over. How about Twain? He was around in that year. His books were aimed at the common man more than being considered "classic" literature. He was also known for saying witty things. I'll go with him. Wait, I'm not sure, maybe I'll just leave it blank. (Just kidding).

At first, I thought her DD bet was too small. But it wasn't that bad. A huge bet means she has to get it right or the game is over. The small bet helped her situation a little bit, but gave her a chance to still break the lock if she misses it.

As for the FJ bets.......

first, I hate it when Alex says the game is a lock tie or one just shy of being a lock. If you can't figure out the math, too bad.

If I had a lock tie lead and someone put a gun to my head and said I could not bet zero, I'd bet a dollar. I understand the logic of betting all but a dollar or putting yourself a dollar ahead of second places pre-FJ score, but the $1 bet gives you a safety point in case the other player decides to hold back a dollar for some reason.

It would take a lot for me to bet anything in that spot short of the gun to my head. Maybe if I have a lock tie lead over a superchamp, but even then I'd want to bank the money.

[jeff6286]

As for non-zero wagers from lock-tie leads, I know Billy Baxter did it in his College Championship semifinal (which I can't blame him for, since there would have had to be a tiebreaker round otherwise). There's also this game.

There's also this game from 2006 where the leader wagered for the win and second was more than happy to let him have it.

Oh good, I had been waitin' for somebody to post this game.

I'm aware, and very appreciative, of Brian's posts (but they aren't always the best option when spending lots of time with family during the holidays). I've definitely missed the FJ/results postings that have vanished these last couple weeks... did we figure out what happened to those, or if someone is going to step in?

That would be me, and while this sounds a bit like deja vu, I do apologize for my absence. Much like I did over Thanksgiving weekend, I fell behind in watching the show on the pre-Christmas weekend, and with the combination of family stuff, a busy post-Christmas work schedule and a blizzard in Indianapolis this week, it took me until now to finally get up to date. Sorry to leave hanging those who come looking for the FJ clue in the daily thread. I told myself last time that I would put out a request for someone to fill in for me if I was going to be unable to fulfill my obligations, but once again, I didn't quite see it coming and it just kept snowballing on me. (quite literally, in fact)

I should be back on schedule Monday, and my local station shouldn't have any conflicts with New Year's Eve or Day programming, so barring severe weather bulletins, I'll hope to get back to posting each day's clue in a timely manner. If anyone would like to volunteer to fill my shoes in case of another prolonged absence, please feel free to let me know either here, or by PM. If I know of someone willing to fill in for a bit, it might make me more likely to let that person know in advance if I am going to be unable to make my usual posts.

I'll likely be going on vacation the first week of April, but hopefully shouldn't miss any more shows until then.

[lisa0012]

I'm aware, and very appreciative, of Brian's posts (but they aren't always the best option when spending lots of time with family during the holidays). I've definitely missed the FJ/results postings that have vanished these last couple weeks... did we figure out what happened to those, or if someone is going to step in?

That would be me, and while this sounds a bit like deja vu, I do apologize for my absence. Much like I did over Thanksgiving weekend, I fell behind in watching the show on the pre-Christmas weekend, and with the combination of family stuff, a busy post-Christmas work schedule and a blizzard in Indianapolis this week, it took me until now to finally get up to date. Sorry to leave hanging those who come looking for the FJ clue in the daily thread. I told myself last time that I would put out a request for someone to fill in for me if I was going to be unable to fulfill my obligations, but once again, I didn't quite see it coming and it just kept snowballing on me. (quite literally, in fact)

I should be back on schedule Monday, and my local station shouldn't have any conflicts with New Year's Eve or Day programming, so barring severe weather bulletins, I'll hope to get back to posting each day's clue in a timely manner. If anyone would like to volunteer to fill my shoes in case of another prolonged absence, please feel free to let me know either here, or by PM. If I know of someone willing to fill in for a bit, it might make me more likely to let that person know in advance if I am going to be unable to make my usual posts.

I'll likely be going on vacation the first week of April, but hopefully shouldn't miss any more shows until then.

No problem, glad to see you are still around and it was (mostly) good things keeping you away. If you ever need a sub let me know- we get the show early in Dallas, so assuming I'm in town (which I am obviously not this week, but ya know) I watch my DVR when I get home from work. Your services are a valuable contribution to the board!

[davey]

I thought this board was very hard. I am glad that I was not on the show to play this one.

I got FJ mainly from the year.

I have caught some criticism for saying this on here (or on the old Sony board) before, but in certain situations I would consider betting $1 on FJ if I had a lock tie. The scenario where I would be most likely to do it is if I had already won 5 or more games, liked the FJ category, and was facing a contestant who I thought was very good and I was lucky to have a huge lead on them. The only time that I would absolutely not do is if it was my 5th game, and I would probably not do it in my 4th game, either. With that said, Paula was not a contestant who I would be afraid to play again, so I would have made the $0 wager here.

I don't understand why you'd be afraid of someone you had a huge - lock-tie - lead on. Talk about a low confidence level!

[jeff6286]

I thought this board was very hard. I am glad that I was not on the show to play this one.

I got FJ mainly from the year.

I have caught some criticism for saying this on here (or on the old Sony board) before, but in certain situations I would consider betting $1 on FJ if I had a lock tie. The scenario where I would be most likely to do it is if I had already won 5 or more games, liked the FJ category, and was facing a contestant who I thought was very good and I was lucky to have a huge lead on them. The only time that I would absolutely not do is if it was my 5th game, and I would probably not do it in my 4th game, either. With that said, Paula was not a contestant who I would be afraid to play again, so I would have made the $0 wager here.

I don't understand why you'd be afraid of someone you had a huge - lock-tie - lead on. Talk about a low confidence level!

Just because it's a lock tie doesn't mean you have a huge lead! Maybe you're playing against Joon Pahk and you lead $1600 to $800, after Joon bet $20,000 and missed on a late Daily Double in a basic arithmetic category. Hey, it could happen!