1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Wendy's dishes up fast food faster, says study

Discussion in 'The OT' started by fluffybear, Oct 9, 2012.

  1. sigma1914

    sigma1914 Well-Known Member DBSTalk Club

    14,582
    369
    Sep 5, 2006
    Allen, TX
    Exactly. ;) I wasn't disagreeing with your position because it's right. I was just trying to say what you just did in a much better way. Stinking Yankees have me frustrated.
     
  2. James Long

    James Long Ready for Uplink! Staff Member Super Moderator DBSTalk Club

    45,307
    914
    Apr 17, 2003
    Michiana
    And here I thought that the topic was speed of service at "fast" food restaurants ... where the best and worst were separated by 78 seconds.

    And that is where knowing the facts behind the figures helps. How was the survey conducted? Was the sample diverse enough not to oversample a particular element? Was the sample taken across the entire geographic area that the outcome is claimed to cover?

    Making a claim that ALL Burger Kings nationwide were 78 seconds slower on average than all Wendy's based only on measurements at a few restaurants in a few states would not be a good claim.

    (And the funny thing is, we are comparing Burger King vs Wendy's. These are not euphanisms for political parties.)
     
  3. sigma1914

    sigma1914 Well-Known Member DBSTalk Club

    14,582
    369
    Sep 5, 2006
    Allen, TX
    Sorry for the political reference, admins, it was all I could think of spur of the moment.
     
  4. spartanstew

    spartanstew Dry as a bone

    12,561
    61
    Nov 16, 2005
    Wylie, Texas
  5. Stewart Vernon

    Stewart Vernon Roving Reporter Staff Member Super Moderator DBSTalk Club

    21,575
    373
    Jan 7, 2005
    Kittrell, NC
    You know... the odds of flipping a coin and getting "heads" are 50/50 each time you flip... but you might flip 100 times in a row the same result...

    Even at their best, statistics are only good at describing PAST behavior... they are NOT meant to be a predictor. You're using them wrong if you're using them as a predictor. That's just one problem with people and statistics.

    Statistics might be very good at documenting things that already happened... but are often virtually useless at predicting what will happen next... that's the nature of the math.

    Your birthday example is less about statistics and more about human behavior. With the average gestation around 9 months... and people's tendencies to procreate more at certain times of the year than others... that's why you find more common birthdays with a small sampling... because the behavior of people is kind of clustered... not because of statistical magic.

    So... I'm not even saying your statistics are wrong... They are correct (probably) for the sampling size you took... but that doesn't necessarily extrapolate to a larger group.

    Also... and I haven't even brought this into things yet... but people lie. People lie all the time in polls, especially if the poll is for something they don't feel comfortable about admitting... How many times have we seen exit polls skew towards one candidate while the actual voting tally skews the other way? People don't always tell the truth, and your random sampling might inadvertently be random enough that you get more of one side than the other.

    I'll reiterate because it is important... At their best, statistics describe past behavior and occurrence. They were never meant to be used to accurately predict future behaviors... but people sure try to use them that way.
     
  6. Lord Vader

    Lord Vader Supreme Member

    8,687
    38
    Sep 20, 2004
    Galactic Empire
    According to something I once read, that is not true. The odds are not exactly 50%.
     
  7. Stewart Vernon

    Stewart Vernon Roving Reporter Staff Member Super Moderator DBSTalk Club

    21,575
    373
    Jan 7, 2005
    Kittrell, NC
    Technically speaking... you're correct. Also, something people seldom take into consideration is... your flip isn't always the same AND the greater environment plays a part in the outcome... and none of those things are a constant. The statistics regarding coin-flip outcomes assume that all other things are held constant, and we know this to not be true in reality... thus, even IF the statistics are valid, we aren't living in a world that conforms to the assumptions made.

    On a larger note... here's a nice article I found earlier today:

    Odds are, It's wrong

    The article itself deals mostly with statistics of things that might actually be predictable... but goes into details about how and why many scientists are disturbed over how statistics are used. They are meant to be a tool to help shortcut some situations, but not to be relied on as proofs... and yet, all too often they are used instead of proof, which leads to faulty assumptions and theories.

    Basically... smarter people than me are not on the statistics-mean-everything train. :)
     
  8. spartanstew

    spartanstew Dry as a bone

    12,561
    61
    Nov 16, 2005
    Wylie, Texas
    Stewart, I think you should just give up, because you're digging your self deeper with your lack of understanding.

    They are meant to be a predictor, NOT describe past behaviour. Using your coin flip example, yes, you might flip 100 times with the same result (heads). What are the odds that it's heads again? 50/50. It has nothing to do with the past.

    Incorrect again. It has nothing to do with anything you've written. It's a simple mathematical formula.

    http://en.wikipedia.org/wiki/Birthday_problem
     
  9. Lord Vader

    Lord Vader Supreme Member

    8,687
    38
    Sep 20, 2004
    Galactic Empire
    This is not correct.
     
  10. Stewart Vernon

    Stewart Vernon Roving Reporter Staff Member Super Moderator DBSTalk Club

    21,575
    373
    Jan 7, 2005
    Kittrell, NC
    You really need to research some statistical facts yourself here, because you are actually the one out on a limb here.

    Even scientists agree... statistics are about studying collected data... they are absolutely not meant to be a reliable way to predict the future.

    Again with the coin-flip... it technically isn't 50/50, but assume it is because that's an easier conversation...

    So... I can flip 100 times in a row, and each time the odds are the same 50/50... So, the statistics provide ZERO in terms of reliability in predicting the outcome of the next coin flip. All the statistics say is that theoretically, if you did it infinitely, the coin flips would average out and approach the middle... but they can't predict any individual future coin flip with any accuracy at all.

    That's what you (and a lot of people) are missing. Check out the link I provided as well... the good scientists and mathematicians know statistics are a tool to be used... not in the way the media or politicians or corporations tend to use them, however.

    edit: Forgot to reply to the birthday example... I'm not questioning the math of probability, I'm questioning the blind application to reality. While the statistical analysis might say that the odds show in a room of 30 people it is more likely than not for 2 people to share the same birthday... it absolutely does NOT guarantee it. In fact, according to the math... only having 366 (or 367 really to account for leap years) people in the room would guarantee 2 people with the same birthday. Anything less than that, and you're just talking about probability. Probability is right there in the definition or the word... IF it was a guarantee, then they would have named it Guaranteeability or something like that :)

    Again, the key... trying to take statistics from collected data (or mathematical probabilities) and guarantee a future outcome is not something you can rely on.
     
  11. Nick

    Nick Retired, part-time PITA DBSTalk Club

    21,838
    186
    Apr 23, 2002
    The...
    A leap-year is 366 days. For non-scientific purposes, the average Earth year is 365.25 days.
     
  12. AntAltMike

    AntAltMike Hall Of Fame

    3,783
    107
    Nov 20, 2004
    College...
    I saw the episode of Get Smart where the Chief was going to make a selection based on a coin flip, but the coin landed on its edge.
     
  13. AntAltMike

    AntAltMike Hall Of Fame

    3,783
    107
    Nov 20, 2004
    College...
    http://books.google.com/books?id=YR...cruiting+on+campus"&ie=ISO-8859-1&output=html


    From: Statistics as Principled Argument, by Robert P. Abelson, 1995
     
  14. 4HiMarks

    4HiMarks Hall Of Fame

    1,560
    33
    Jan 21, 2004
    Laurel, MD
    That's right. No one ever said statistics was a guarantee. It does give an accurate probability of outcome, though. With proper methodology, it can be VERY accurate. I used to know a statistician who had a quote on his office door that responded to the famous Mark Twain quote with the following: It is indeed easy to lie with statistics, but it is impossible to tell the truth without them."

    Nothing is absolutely certain. There is always a (very) slight chance a "mortal lock" will go wrong, like a coin landing on its edge, or the hare dropping dead of a heart attack just before crossing the finish line of his race with the tortoise. But statistics gives us a scientifically proven way of assigning probabilities to future outcomes. The entire insurance industry is built upon this. They can't tell you when you will die, but they can tell with near certainty how many people in your demographic will die in a given year.
     
  15. TXD16

    TXD16 Icon

    955
    30
    Oct 29, 2008
    No, "scientists" do not agree, nor do mathematicians, nor do statisticians.

    As has been pointed out to you numerous times already (spartanstew even provided a couple of excellent links, which you either did not read or did not comprehend), properly modeled and properly conducted statistical surveys are superbly accurate at, as you put it, "predict[ing] the future," often with a greater than 95% certainly. It's not a matter of opinion, it's a matter of fact, yet, for some reason, you continue to argue your opinion against that established fact.

    As has been suggested, you are obviously so far out of your element that you should just stop.
     
  16. djlong

    djlong Hall Of Fame

    4,343
    57
    Jul 8, 2002
    New Hampshire
    My personal favorite line about statistics...

    He uses statistics the way a drunk uses a streetllamp-post - for support rather than illumination.
     
  17. spartanstew

    spartanstew Dry as a bone

    12,561
    61
    Nov 16, 2005
    Wylie, Texas
    Yes, it is. Do you think coins have memories?
     
  18. Lord Vader

    Lord Vader Supreme Member

    8,687
    38
    Sep 20, 2004
    Galactic Empire
    It's not correct. There's an interesting factoid about why a coin flip isn't actually a 50/50 scenario.
     
  19. spartanstew

    spartanstew Dry as a bone

    12,561
    61
    Nov 16, 2005
    Wylie, Texas
    Nice loophole.

    I'm not going to debate with you whether or not a coin flip is exactly 50/50 or if it's 49.8/50.2 (but, I'm a believer that it is 50/50, especially when played out to infinity - or as close as you can get)

    The point I was making (which I'm sure you're aware of), is that the previous flips have nothing to do with the probability of the next flip.
     
  20. Stewart Vernon

    Stewart Vernon Roving Reporter Staff Member Super Moderator DBSTalk Club

    21,575
    373
    Jan 7, 2005
    Kittrell, NC
    Yes... which is why I said 367 people would be required to guarantee duplication. While unlikely, IF you only had the same number of people as possible birthdays, it would be possible (if unlikely) to have each person with a unique birthday.

    Once you exceed the number of different days, 367, then you would be guaranteed duplication.

    Of course the duplication probably happens sooner... but there's no way to guarantee it without having the larger sample size.

    Oh wait, did I just say that? That while the probability of an outcome becomes greater than 50% at a much lower sample size, the only way to be sure of an outcome is with a larger sample size?

    Where have I heard that before? :)
     

Share This Page