LapDawg said:
^but see how no matter what you do, the stats still match up with my original percentage claim.
It's not like what I was saying was stretching as far as you make it out to be. You could use the 41% and when you match it with the 60% for 40-49, my opinion of the 50-yard barrier being a better indicator for <50% conversion holds perfectly fine. How does it not?
And if we're looking for the "average kicker", I'm not sure why you don't accept my usage of kickers who are percieved as the "average" through their FG %, in a more focused group that matches the supposed query of what we're discussing. Instead, you want to eliminate the best kickers for stats you assume are far worse when the stats simply aren't available.
Again...I'm talking about 50+ yard field goal attempts in regards to taking the bottom 7 out. Not all of the kickers even tried that range, so you can only go with who DID. And from that alone, the 50+ range is more accurate than what you keep claiming since you refuse to acknowledge your manipulation of that stat by eliminating the best kickers who attempted that range without equally balancing the usage of that stat by taking out the bottom kickers regarding that stat alone.
I see no reason from kickers 80-100 to believe there's a significant drop-off in the 45-49 range for kicker 101-119 when compared to them. Seems like you're reaching for straws when you keep having to claim the lack of those stats being such a significant missing piece.
Okay, we'll agree to disagree, but since you did ask, I'll respond one last time.
As far as eliminating an additional bottom 7 50+ stats, okay, now I see your point...I thought you were talking about an additional bottom 7 from the 40-49 as well, which I think we both agree would be unfair from my above analysis. Eliminating an additional bottom 7 from the 50+ calculus would only eliminate an additional 3-4 attempts with no conversions though, and the resulting conversion rate would only rise an approximate 1.2% from 37.5% to 38.7%. Not enough of a difference to quibble over, frankly.
Now the selected 50-70 sample you used IS worth quibbling over, though. You
know that the smaller a sample, the greater the statistical margin of error, so why would you choose a smaller sample when we have a larger sample already available? E.g., look at your MOE in the 50+ category in your sample...it renders a conversion rate of 56%, when we already KNOW from the entire sample, even assuming
none of the top kickers are eliminated, that probably
at best the 50+ conversion rate is 47% (do you
really believe that those unlisted 19 worst kickers are going to
improve that percentage?). Just a crude subtraction tells you that's probably somewhere around a 9% MOE, dude...how is that acceptable to you, when you KNOW that you can lower the MOE by simply employing the larger sample? Are you telling me you're perfectly happy being more wrong than you would otherwise be by using the larger sample, as long as you don't have to adjust your assertion?
The reason that some adjustment HAS to be made at the top is that we
know that the list is missing data at the bottom, so the resulting percentages are biased to some degree toward the top...that's virtually undeniable. E.g., it's like how they remove the highest and lowest judging scores from consideration in Olympic diving contests to attempt to remove the inherent bias from any judging subjectivity. It's an entirely fair procedure, and the only question is what exclusion number best offsets that bias.
And mind you, by excluding X number of top kickers we're simply treating them as if we know nothing about their stats, just as we know nothing about the worst 19's stats. I would venture that the reason it sticks in your craw so much is only because you already know the data from those top 7 kickers and how strongly it supports your assertion, particularly in the 50+ category...if you had never known the data in the first place, I'm guessing you wouldn't object so strenuously.
Now granted, due to the likelihood of the better kickers having more attempts from those ranges than the worst, it's not a perfect 1:1 correlation or we could simply strike the 19 top kickers to account for the omission of the 19 worst. So that's why I tried to figure out the number to be excluded as I did. And maybe I didn't come up with a perfect means of figuring how many top kickers to strike, but I did my best to be completely impartial about it, and if anything I think I weighed it a little more in your favor. Unless you had a better proposal for addressing the top-heavy bias of the entire sample, I don't see how you can fault mine so severely.
So in summary, I'll go along with your exclusion of 7 more from the 50+ category, because I see your point now. It only makes a difference of slightly more than 1%, though.
Now that I've addressed your final concerns, we can agree to disagree.
