Re: What's up with NAQT ICT Invitations

Samer Ismail <*samer_ismail_at_samer_ismail.yahoo.invalid*> · Tue, 21 Feb 2006 20:42:56 -0000

--- In quizbowl_at_yahoogroups.com, "Susan Ferrari" <srferrar_at_...> wrote:
>
> Hi,
> 
> I just wanted to restate something I must not have stated clearly in
> an earlier post.  The autobids for which I'd like to see some sort 
of
> minimum S-value (or whatever--some measure of performance compared 
to
> the total field) imposed to guarantee a bid were not the host
> autobids, just the D1 winner, undergrad winner, and D2 winner 
autobids.  
> 
> I completely understand the purpose of the host bid as an incentive
> for hosting, and I think it would be difficult for NAQT to find 
enough
> competent SCT hosts without offering it.  When I mentioned SFU in an
> earlier post, it was because I had missed the fact that they 
qualified
> with their host bid and had thought that their D2 team's performance
> qualified them.  
> 
> Also, with regard to Stan's post, I understand your frustration, but
> this seems to happen almost every year (see Princeton and Maryland 
at
> the 2001 SCT).  The last time I recall the balance of components in
> the S-value being explained (in this post:
> http://groups.yahoo.com/group/quizbowl/message/3920), it was made
> clear that the weight of a team's win-loss record was low compared 
to
> ppth.  I do recall hearing that NAQT ensures that teams from lower
> brackets don't receive bids ahead of teams in the same sectional 
from
> a higher bracket, but within a bracket, win-loss record probably
> counts for fairly little.  

One minor point: that post actually refers to the older pre-S-value 
formula (i.e., the one that created the Princeton/Maryland issue).

I can say this much about S-values, though:

* The actual S-values themselves do not take into account win-loss 
records.
* The win-loss records are actually used in the seeding process, 
though: for one team to advance ahead of a team with a better record 
*from the same sectional*, that team had to have a significantly 
higher S-value; the greater the difference in record, the greater the 
difference needed.