Re: Minnesota Golden Gopher Season 2018-2019: Fire Motzko
Another contributor is mentioned in the article that I don't think many people talk about is the apparent increase in talent pool. There seems to be more talented kids to go around. Not just due to the rise in other states. I was talking to a friend about some HS players who are destined for D1 teams didn't seem to stand out that much. And we both seemed to recollect that kids who went on to play D1 20 and 30 years ago seemed to stand out from the rest of the players on the ice pretty quickly, whereas now it seems like it is more like a shift here or there you notice them. Maybe it is poor memory, but it seems kids today are better trained and there is more depth, at that level, which would translate into colleges who get the first choice of kids would have less of a talent gap than the rest. And I would argue that the fact that about a third of NHL players come from NCAA now would support that. (Although this could also be due to more talented kids just choosing the NCAA route too, but there seems to be more Minnesotans in the NHL as well.) Add in the fact that the kids who ARE clearly more talented than the average will leave early, and one can see why this would contribute to parity.
As for the one and done influencing stats, if all you are doing is looking at win percentages in all games, how does that influence the stats? And even if you want to look at number of championships by seeding, or mean seed of Champion from a period pre CBA to post, while the sample would be small, even with the confounding effect I would expect one would see a trend towards an increase in low seeds winning. (Keep in mind, it was one and done pre and post CBA.)
I have moved into a different position, but in my previous position did some stats on biological data. While you can talk about ideal sample sizes and having perfect design, you can never create a perfect statistical sampling method, as you are sampling highly variable ever changing systems. If you held to rigid rules of statistical agreement, you would never conclude anything. The bar for statistical significance was lower than in other fields. What you typically were looking at were trends, not conclusive answers. And even though some people might argue the statistical significance of a conclusion, it was the best you could do in many situations. That did not mean the conclusions were wrong and the decisions we made were incorrect. SO if there are trends in number of 4 seeds winning games against #1 and trends in increasing low seeds winning it all, I will believe there is a good chance they represent reality. So go ahead and do stats on a better data set if you like. I don't need to. I think most people sense the shift that has occurred. Some of that is based on the observations I have mentioned. So why do some people feel the need to criticize such an approach without offering a better analysis? (And most people don't care about the details of stats and just want to discuss what they are seeing.) Argue if you think that there is not more parity in the league and offer your supporting observation. Buy why agree with my point but argue about the way I chose to show it? Does anyone think there isn't more parity now?
Another contributor is mentioned in the article that I don't think many people talk about is the apparent increase in talent pool. There seems to be more talented kids to go around. Not just due to the rise in other states. I was talking to a friend about some HS players who are destined for D1 teams didn't seem to stand out that much. And we both seemed to recollect that kids who went on to play D1 20 and 30 years ago seemed to stand out from the rest of the players on the ice pretty quickly, whereas now it seems like it is more like a shift here or there you notice them. Maybe it is poor memory, but it seems kids today are better trained and there is more depth, at that level, which would translate into colleges who get the first choice of kids would have less of a talent gap than the rest. And I would argue that the fact that about a third of NHL players come from NCAA now would support that. (Although this could also be due to more talented kids just choosing the NCAA route too, but there seems to be more Minnesotans in the NHL as well.) Add in the fact that the kids who ARE clearly more talented than the average will leave early, and one can see why this would contribute to parity.
As for the one and done influencing stats, if all you are doing is looking at win percentages in all games, how does that influence the stats? And even if you want to look at number of championships by seeding, or mean seed of Champion from a period pre CBA to post, while the sample would be small, even with the confounding effect I would expect one would see a trend towards an increase in low seeds winning. (Keep in mind, it was one and done pre and post CBA.)
I have moved into a different position, but in my previous position did some stats on biological data. While you can talk about ideal sample sizes and having perfect design, you can never create a perfect statistical sampling method, as you are sampling highly variable ever changing systems. If you held to rigid rules of statistical agreement, you would never conclude anything. The bar for statistical significance was lower than in other fields. What you typically were looking at were trends, not conclusive answers. And even though some people might argue the statistical significance of a conclusion, it was the best you could do in many situations. That did not mean the conclusions were wrong and the decisions we made were incorrect. SO if there are trends in number of 4 seeds winning games against #1 and trends in increasing low seeds winning it all, I will believe there is a good chance they represent reality. So go ahead and do stats on a better data set if you like. I don't need to. I think most people sense the shift that has occurred. Some of that is based on the observations I have mentioned. So why do some people feel the need to criticize such an approach without offering a better analysis? (And most people don't care about the details of stats and just want to discuss what they are seeing.) Argue if you think that there is not more parity in the league and offer your supporting observation. Buy why agree with my point but argue about the way I chose to show it? Does anyone think there isn't more parity now?
