Archive

Cool Look at Seeds and Advancement to the Final Four

  • Laley23
    Here is a really cool graph showing the ACTUAL percentage of times seeds have advanced to the Final Four versus the PROBABILITY it will happen. Maybe the committee does know what they are doing. These numbers are really close.

  • Al Bundy
    Laley23 wrote: Here is a really cool graph showing the ACTUAL percentage of times seeds have advanced to the Final Four versus the PROBABILITY it will happen. Maybe the committee does know what they are doing. These numbers are really close.

    What is the probability based upon?
  • hoops23
    Seeding.
  • Laley23
    http://www.fivethirtyeight.com/2010/03/incredible-izzo-again-defies-odds.html?utm_source=twitterfeed&utm_medium=twitter

    No clue, but here is the article. Its about how good Izzo has been versus the odds. Im not sure how they calculate it.

    http://www.fivethirtyeight.com/2010/03/incredible-izzo-again-defies-odds.html?utm_source=twitterfeed&utm_medium=twitter
  • hoops23
    The committee does know what they're doing. The only thing they really mess up on is location, though it's difficult to get everything perfect.

    I know it's a tough job, given the circumstances and time frame.
  • SQ_Crazies
    ccrunner609 wrote: that is what it should look like.
    Yup.
    LTrain23 wrote: The committee does know what they're doing.
    ^That's why it looks the way it's supposed to.

    The committee gets it right. People just find ways to bitch about or find fault with anything. There are probably dudes who don't think Marisa Miller is hot--I rest my case...
  • hoops23
    CCrunner thinks Marisa Miller looks like a dude. (which is what he likes) :D



    I kid the guy.
  • Al Bundy
    Without seeing the how they determined the probabilities, this is just self-fulling graph.
  • SQ_Crazies
    I don't understand exactly how you can be confused with this graph...
  • Cleveland Buck
    If the probabilities are based on the past performances of the various seeds, then the graph of the probabilities will be exactly the same as the graph of past performance, because they are using the same data.
  • SQ_Crazies
    It's a theory dude.

    The green line is all that you could have a problem is, and it's just a theory--that's why there are seeds in the first place. Theoretically the line makes total sense--theoretically that's why they seed teams where they do. If you don't think the theoretical line makes sense then you clearly must have another theory than most people. I mean there is no way to prove that the theory on how the seeds will do is true--it's pure speculation, assuming that the best teams are usually one seeds so they win more often and so on. That would have to be assumed true, otherwise the seeds shouldn't be based on what the team did over the course of the year--they might as well be random.
  • Al Bundy
    SQ_Crazies wrote: I don't understand exactly how you can be confused with this graph...
    I'm not confused on the graph, but they aren't saying how the probabilities were calculated. It is very easy to back and make up probabilities the correspond with what has already happened in the past.
  • SQ_Crazies
    Al Bundy wrote:
    SQ_Crazies wrote: I don't understand exactly how you can be confused with this graph...
    I'm not confused on the graph, but they aren't saying how the probabilities were calculated. It is very easy to back and make up probabilities the correspond with what has already happened in the past.
    It's assuming that the committee gets the seeding right based on who the best teams are. You think they do the VAST majority of the time don't you? It's no exact science, it's sports--anything can happen, that's why they play the games. But I'd say almost everybody agrees with that theoretical line. If they didn't then there wouldn't even be a committee to select seeds.
  • Al Bundy
    SQ_Crazies wrote: It's a theory dude.

    The green line is all that you could have a problem is, and it's just a theory--that's why there are seeds in the first place. Theoretically the line makes total sense--theoretically that's why they seed teams where they do. If you don't think the theoretical line makes sense then you clearly must have another theory than most people. I mean there is no way to prove that the theory on how the seeds will do is true--it's pure speculation, assuming that the best teams are usually one seeds so they win more often and so on. That would have to be assumed true, otherwise the seeds shouldn't be based on what the team did over the course of the year--they might as well be random.
    The don't give the formula that they used for the theory. Have you ever taken an elementary stats class? Without the formula used, any graph can be created to show whatever results you want to conclude.
  • SQ_Crazies
    Did you read the article? It's simple statistics, it makes perfect sense to me. That's the thing about statistics though--you can't read too far into a lot of models like this, you have to take them for face value. Believe it or don't believe it, it doesn't matter--that's another think about statistics, there are always outliers.
  • Al Bundy
    SQ_Crazies wrote:
    Al Bundy wrote:
    SQ_Crazies wrote: I don't understand exactly how you can be confused with this graph...
    I'm not confused on the graph, but they aren't saying how the probabilities were calculated. It is very easy to back and make up probabilities the correspond with what has already happened in the past.
    It's assuming that the committee gets the seeding right based on who the best teams are. You think they do the VAST majority of the time don't you? It's no exact science, it's sports--anything can happen, that's why they play the games. But I'd say almost everybody agrees with that theoretical line. If they didn't then there wouldn't even be a committee to select seeds.
    I was just asking about the formula that they used to generate their results. I am not debating that the better seeds normally do better. I just question their percents without seeing the actual formula behind them. Maybe they have a legit formula, but if it is just a regression formula based upon past results, it is going to generate a very low margain or error and basically be the same as the actual results.
  • Al Bundy
    SQ_Crazies wrote: Did you read the article? It's simple statistics, it makes perfect sense to me. That's the thing about statistics though--you can't read too far into a lot of models like this, you have to take them for face value. Believe it or don't believe it, it doesn't matter--that's another think about statistics, there are always outliers.
    I did read the article, but it doesn't show the formula that they used.
  • SQ_Crazies
    You're sure you read the article?
    Very, very low. We can do a good job of modelling a team's probability of making the Final Four by using a logistic regression model based on the square root of its seeding
    The words logistic regression model are even a link, and therefore a different color.
  • Al Bundy
    SQ_Crazies wrote: You're sure you read the article?
    Very, very low. We can do a good job of modelling a team's probability of making the Final Four by using a logistic regression model based on the square root of its seeding
    The words logistic regression model are even a link, and therefore a different color.
    Yes, I'm sure I read the article. It gives the model that they used, but it doesn't give the exact values they they substituted into the formula. Just like I told you above, it is very easy to take a regression model and created a formula that basically mirrors past results with a very small error.
  • SQ_Crazies
    And like I said, take it for face value. Even when something is based on past results it guarantees nothing for the future. Like I said, if that line doesn't make sense to you then you must think that seeding is pointless.
  • Al Bundy
    SQ_Crazies wrote: And like I said, take it for face value. Even when something is based on past results it guarantees nothing for the future. Like I said, if that line doesn't make sense to you then you must think that seeding is pointless.
    First of all, it isn't a line, it is a an inverse log curve (a topic you probably had in algebra or algebra II).

    Secondly, I never said that seeding was pointless, I just highly question the validity of their regression equation without seeing it. It is very easy to manipulate a formula to generate the results to try to prove anything.
  • SQ_Crazies
    Jesus Christ you can't just accept the word line for the point of conversation? Technically it is a fucking line, a curve is a line. Sort of like a square is a rectangle--that's probably first grade math you dick. I'm done with this argument, you'd rather try to question my intelligence and try to be a hardass than read, comprehend, understand and try to have an adult level discussion.
  • Al Bundy
    SQ_Crazies wrote: Jesus Christ you can't just accept the word line for the point of conversation? Technically it is a fucking line, a curve is a line. Sort of like a square is a rectangle--that's probably first grade math you dick. I'm done with this argument, you'd rather try to question my intelligence and try to be a hardass than read, comprehend, understand and try to have an adult level discussion.
    A curve is not a line. A line a is a curve with a constant rate of change, but a log curve is completely different from a line. No wonder the flaws in the article were way over your head.
  • SQ_Crazies
    A curve is a line that doesn't have to be straight, a line is a straight curve...whatever, it's the same thing, it's a fucking line. Like I said, you can't just accept the word line for the point of conversation?

    Eat shit you fucking douche, learn how to not talk like a cock sucker and maybe you won't be one...
  • SQ_Crazies
    http://www.merriam-webster.com/dictionary/curve

    The first two words of the definition are "a line".