BTW, I have never advocated trading down from no. 8. I think it would be stupid to do so given the talent that would be available at 8 in all positions, Floyd, Reiff possibly, Ingram or Coples.
ChyrenB Wrote:
-------------------------------------------------------
> BTW, I have never advocated trading down from no.
> 8. I think it would be stupid to do so given the
> talent that would be available at 8 in all
> positions, Floyd, Reiff possibly, Ingram or
> Coples.
But, with the 15 or so "studs" (Crowder's term) available if Miami traded down no further than 15, Miami could still get a good player, along with an additional 2nd round pick.
Crowder52 Wrote:
-------------------------------------------------------
> Yes he is defintely reading this site, I think
> alot of them do, to get a feel for ideas for
> stories... Alot of stories I see in local papers
> fall a couple days behind a good topic here...
> that being said I thought this commenters argument
> as to the validity was pretty strong as well, so I
> though i would share it
>
> THe bigger myth is that statistics outside of
> controlled experiments mean much of anything -
> except revealing spurious correlations. Take your
> 2nd round QB numbers. Even your adapted numbers
> mean nothing. Any accuracy would demand far more
> data. You cannot compare success rates unless you
> also knew (for example): how strong was the team
> they went to, what was their opponents records,
> who were the starting QBs in front of them and how
> long did they block the draft choice from playing,
> how good were the coaches, did they play in
> cold/warm weather, etc. You can certainly run
> analysis that takes these kinds of things into
> account (e.g. multiple regression), but you NEVER
> see that in sports discussions. The other problem
> is that even though everyone knows there is
> missing data, they tend to discount how much it
> distorts the accuracy of their interpreation.
> Without the data from a single key variable the
> conclusions are not just 'a little bit'
> inaccurate, they often are totally meaningless.
> The tendency when we cannot get the 'right' data
> is to pretend the almost right data is nearly just
> as good. It is not. It is nearly impossible to
> know if it means anything. Having said that,
> thanks for the interesting piece
Thanks for sharing Crowder. That is a great comment. It is impossible to conclude a direct correlation with so many uncontrolled variables.
That's why they say: There are lies, damn lies, and statistics.
Crowder52 Wrote:
-------------------------------------------------------
> Yes he is defintely reading this site, I think
> alot of them do, to get a feel for ideas for
> stories... Alot of stories I see in local papers
> fall a couple days behind a good topic here...
> that being said I thought this commenters argument
> as to the validity was pretty strong as well, so I
> though i would share it
>
> THe bigger myth is that statistics outside of
> controlled experiments mean much of anything -
> except revealing spurious correlations. Take your
> 2nd round QB numbers. Even your adapted numbers
> mean nothing. Any accuracy would demand far more
> data. You cannot compare success rates unless you
> also knew (for example): how strong was the team
> they went to, what was their opponents records,
> who were the starting QBs in front of them and how
> long did they block the draft choice from playing,
> how good were the coaches, did they play in
> cold/warm weather, etc. You can certainly run
> analysis that takes these kinds of things into
> account (e.g. multiple regression), but you NEVER
> see that in sports discussions. The other problem
> is that even though everyone knows there is
> missing data, they tend to discount how much it
> distorts the accuracy of their interpreation.
> Without the data from a single key variable the
> conclusions are not just 'a little bit'
> inaccurate, they often are totally meaningless.
> The tendency when we cannot get the 'right' data
> is to pretend the almost right data is nearly just
> as good. It is not. It is nearly impossible to
> know if it means anything. Having said that,
> thanks for the interesting piece