TPiR Season 37 changes

[Don Howard]

[quote name=\'Joe Mello\' post=\'192147\' date=\'Jul 26 2008, 11:23 AM\']
(Although, if someone is bananas enough to risk $10K, it will be interesting to watch....hopefully)
[/quote]
I suppose that Dan Avila-in-training who gambled $5000 to win $10000 and advised the scared members of the studio audience to buy a dog might take the risk. As for me, as proof of my cowardice, I pull out the $1000 slip and I'm stoppin'.

[TroubadourNando]

Honestly, to me, Check Game seems to be somewhat tied by inflation---not as badly as Poker Game* was---but still. Once inflation makes Check's typical prize reach/top $10,000, you're offering a very good amount of money in addition to that. It probably could be fixed easily enough though.

The only problem I have with the (overdue) Punch upping is that the distribution is a little odd.

*-and one can legitimately argue that poker in a game show context doesn't work, but I think the issue was more Price-specific...eh.

[dale_grass]

What I think pyldude's point was (God help me) is that the relative error shrinks as the prices get bigger.  To use Chris L.'s example, $1000 breathing room for a $5000 prize is a lot more generous than $1000 breathing room for a $1,000,000 prize.  

\Pedantic and academic?  You betcha.
\\I'll get off my podium now.

[clemon79]

[quote name=\'dale_grass\' post=\'192226\' date=\'Jul 27 2008, 09:18 AM\']
What I think pyldude's point was (God help me) is that the relative error shrinks as the prices get bigger.  To use Chris L.'s example, $1000 breathing room for a $5000 prize is a lot more generous than $1000 breathing room for a $1,000,000 prize.  
[/quote]
Okay, that's a little more fair.

But, that said, would we not all agree that the delta in margin of error between a $4000 prize and a $6000 one would be so small as to be nonexistent?

[dale_grass]

[quote name=\'clemon79\' post=\'192234\' date=\'Jul 27 2008, 12:47 PM\']
[quote name=\'dale_grass\' post=\'192226\' date=\'Jul 27 2008, 09:18 AM\']
What I think pyldude's point was (God help me) is that the relative error shrinks as the prices get bigger.  To use Chris L.'s example, $1000 breathing room for a $5000 prize is a lot more generous than $1000 breathing room for a $1,000,000 prize.  
[/quote]
Okay, that's a little more fair.

But, that said, would we not all agree that the delta in margin of error between a $4000 prize and a $6000 one would be so small as to be nonexistent?
[/quote]

Let's just say you were dealing with prizes circa $4000 and will now deal with prizes circa $6000.  The maximum relative error was 25% and is now 17%.  I wouldn't go so far as to call is nonexistent. Admittedly, at first blush it did seem minute, but there you have it.  The absolute error is the same, but the relative error has shrunk.

Another quick point of math:  the next time they make a $2000 jump in the range ends, it will have less of an effect on the relative error, and even less the next time, and so on.

[PYLdude]

[quote name=\'dale_grass\' post=\'192226\' date=\'Jul 27 2008, 09:18 AM\']
What I think pyldude's point was (God help me) is that the relative error shrinks as the prices get bigger.  To use Chris L.'s example, $1000 breathing room for a $5000 prize is a lot more generous than $1000 breathing room for a $1,000,000 prize.  
[/quote]

Well, dale_grass, you hit it. Because that was exactly my point.

[clemon79]

[quote name=\'dale_grass\' post=\'192240\' date=\'Jul 27 2008, 11:40 AM\']
Let's just say you were dealing with prizes circa $4000 and will now deal with prizes circa $6000.  The maximum relative error was 25% and is now 17%.  I wouldn't go so far as to call is nonexistent. Admittedly, at first blush it did seem minute, but there you have it.  The absolute error is the same, but the relative error has shrunk.[/quote]
I knew you were going to say that. And so, I say this: You're not going to bid $1 on a prize whether it's a $4000 prize or a $6000 one. A lot of what you're counting towards that relative error simply isn't applicable because no idiot is going to bid that low. I still suggest it's not mathematically significant.

[dale_grass]

[quote name=\'clemon79\' post=\'192248\' date=\'Jul 27 2008, 03:00 PM\']
[quote name=\'dale_grass\' post=\'192240\' date=\'Jul 27 2008, 11:40 AM\']
Let's just say you were dealing with prizes circa $4000 and will now deal with prizes circa $6000.  The maximum relative error was 25% and is now 17%.  I wouldn't go so far as to call is nonexistent. Admittedly, at first blush it did seem minute, but there you have it.  The absolute error is the same, but the relative error has shrunk.[/quote]
I knew you were going to say that. And so, I say this: You're not going to bid $1 on a prize whether it's a $4000 prize or a $6000 one. A lot of what you're counting towards that relative error simply isn't applicable because no idiot is going to bid that low. I still suggest it's not mathematically significant.
[/quote]

Why would you bid $1 in Check Game?  The relative error is found by taking the absolute error ($1000) and dividing by the actual value ($4000 or $6000).  It doesn't matter who's playing the game.  The farthest off you can be (and still win) is $1000, which is 25% of $4000 and 17% of $6000.  (Actual percentages vary with actual price, obviously.)  The contestant's breathing room has shrunk.  Mathematically significant?  I'll let the more statistically-inclined to rule.

[clemon79]

[quote name=\'dale_grass\' post=\'192249\' date=\'Jul 27 2008, 01:07 PM\']
Why would you bid $1 in Check Game?[/quote]
By "bid" I mean "guess the price of the prize." I'm removing all of the add-the-value-of-the-check chrome from the equation, because it simply doesn't matter for the purposes of this discussion.

And that's my point. You wouldn't. You're gonna look at that $4000 prize and say "That has to be AT LEAST $2000." And you're going to look at that $6000 prize and say "That's gonna be at least $4000." The increase in prize values is going to cover the relative margin of error.

[tpirfan28]

[quote name=\'dale_grass\' post=\'192249\' date=\'Jul 27 2008, 04:07 PM\']
Why would you bid $1 in Check Game?[/quote]
To have Barker yell from his armchair "HISTORIC MOMENT!", then promptly call Roger at home to see if it really was the first time ever someone wrote a check for $1?

[Kevin Prather]

[quote name=\'tpirfan28\' post=\'192253\' date=\'Jul 27 2008, 01:38 PM\']
[quote name=\'dale_grass\' post=\'192249\' date=\'Jul 27 2008, 04:07 PM\']
Why would you bid $1 in Check Game?[/quote]
To have Barker yell from his armchair "HISTORIC MOMENT!", then promptly call Roger at home to see if it really was the first time ever someone wrote a check for $1?
[/quote]
Well played.

[DrJWJustice]

[quote name=\'dale_grass\' post=\'192249\' date=\'Jul 27 2008, 03:07 PM\']
Why would you bid $1 in Check Game?  The relative error is found by taking the absolute error ($1000) and dividing by the actual value ($4000 or $6000).  It doesn't matter who's playing the game.  The farthest off you can be (and still win) is $1000, which is 25% of $4000 and 17% of $6000.  (Actual percentages vary with actual price, obviously.)  The contestant's breathing room has shrunk.  Mathematically significant?  I'll let the more statistically-inclined to rule.
[/quote]

Tthe only thing that has changed that means anything is the prize value, which means a winning contestant will take home between $6k and $7k in cash and prizes on this game.  The argument has been on margin of error, which did not change.  There remains a $1000 window, so the contestant's breathing room has not shrunk, nor has he gained any.  It remains the same.  

As Chris Lemon said (among others), it is a simple game of third-grade math:  Figure out what the price of the prize is, and then write a check that will raise the prize package value to somewhere between $6,000 and $7,000, inclusive.

I don't see where this "relative error" has any significance at all in this case, but if someone can show that it has, let him or her do it, and we'll all learn something.  The fact is that there was a $1,000 window before, and that value has not change with the increased prize value.  

I wish someone had all-time win/loss statistics on this game so that we could run the numbers, and I do know that this is not the first time that this game has seen a change in the value of its prizes.  That would help to settle this.

Meanwhile, I'm just going to tune into Season 37 and enjoy Check Game ... and all the others ... as I always have.

[DrJWJustice]

[quote name=\'PYLdude\' post=\'192242\' date=\'Jul 27 2008, 01:50 PM\']
[quote name=\'dale_grass\' post=\'192226\' date=\'Jul 27 2008, 09:18 AM\']
What I think pyldude's point was (God help me) is that the relative error shrinks as the prices get bigger.  To use Chris L.'s example, $1000 breathing room for a $5000 prize is a lot more generous than $1000 breathing room for a $1,000,000 prize.  
[/quote]

Well, dale_grass, you hit it. Because that was exactly my point.
[/quote]
No, that was not your point, or at least not as it appeared in words. Your point was that the margin of error changed.  It did not.  The player had $1,000  with which to play before, and that did not change.  I accept that you might have miswritten your intentions, but your term was "margin of error," which you claimed had changed, when it has not changed.  When you deal with stats, you cannot substitute terms like you did.  If you had used "relative error" in the first place, I don't think you would have found yourself in the frying pan as you did.  I will welcome Dale's comments -- although I don't think this particular relative margin is going make a difference -- but I will also stick with my previous postings on this as far as the margin of error itself goes.  

Welcome to the world in which I live ... I cannot go to an academic conference these days where there is not at least one pissing contest over which set of numbers ought to be used to do a particular piece of research.  It never gets settled, and this one probably won't, either.

[clemon79]

[quote name=\'DrJWJustice\' post=\'192261\' date=\'Jul 27 2008, 03:03 PM\']
If you had used "relative error" in the first place, I don't think you would have found yourself in the frying pan as you did.[/quote]
I still absolutely would have called bullshiat on him, because I hold fast that the change is statistically insignificant.

[dale_grass]

Coming this fall to Fox: Semantics.  Watch contestants try to figure out why a container of gasoline is marked "inflammable" while the host steps from behind his lectern and onto his podium.

The absolute error (margin of error, window) involved is still $1000.  The relative error has shrunk.

As for "breathing room," I wasn't using it synonymously with "window." I was using it as "comfort level at which a contestant can stab at the price and still feel hopeful to win."  I'd feel safer having a $1000 window on a $6000 prize than on a $60,000 prize.

Another example of relative error:  Let's say someone asked you to count the number of stars visible to the naked eye and how many pennies were in the shallow give-a-penny-take-a-penny dish at the local convenience store.  If you were off by 5 in both cases, you'd be admired in the first and mocked in the second.

I'm not defending anyone's point, just the math involved.

\So help me if you harm the math!