Subject: | |
From: | |
Reply To: | Wayne T. Smith |
Date: | Wed, 18 Dec 2002 02:37:28 -0500 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
Ken Butler wrote, in part:
> There seem to be about five rating systems going around at the moment,
> so I thought it would be interesting to see how (and whether) they agree
> in their rankings of the teams.
>
> I've looked at:
>
> KRACH (calculated by me, or go digging at http://slack.net/hockey)
> CHODR (http://it.stlawu.edu/~chodr/current.html)
> RPI (http://slack.net/~whelan/tbrw/2003/rpi.shtml)
> HEAL (http://maine.edu/heal/heal.latest)
> RHEAL ("iterated HEAL"; http://maine.edu/heal/rheal.latest)
...
> Where are the big discrepancies? One idea is to see which rankings
> differ wildly from their median. Veterans of ratings discussions on
> Hockey-L will not be surprised by the team on the top of the next list
> (though at least it's not RPI to blame this time :-) ):
>
>
> Team System Rank Median Diff
> Quinnipiac HEAL 20 41 21
...
>
> It's mainly HEAL and CHODR that differ greatly from the consensus; it
> looks as if HEAL likes the MAAC and dislikes the Western conference,
> while CHODR is just different... interestingly, RPI agrees pretty well
> with the consensus.
May sound crazy, but who really cares about the lower half of the
ratings, (besides those teams)? I'll just note that the HEAL outliers
are not in the top-16.
...
> Observed results "predicted" by system and conference
>
> n KRACH CHODR RPI HEAL RHEAL
> cc 66 51 53 51 53 53
> ch 13 9 6 7 8 8
> ec 43 35 33 34 32 30
> he 34 28 24 27 28 28
> ma 46 37 30 37 31 33
> nc 166 140 130 134 127 136
> wc 43 36 33 36 36 36
> all 411 336 309 326 315 324
>
>
> Any half-decent rating system should do a good job within a conference,
> because there are so many conference games (perhaps not yet, but by the
> end of the season). The real test is with the non-conference games. RPI
> is actually doing quite well here; also, RHEAL's improvement over HEAL
> is entirely down to the non-conference games.
Good stuff! I'll guess that HEAL does poorly with nc games, relative to
RPI, because RPI "reaches" further due to its 2nd order/(opponents -
opponents) calculation ... missing in HEAL.
I'm not surprised KRACH's "predictions" are better .. just that the
better part varies so much from conference to conference.
I'm also quite surprised that all of the "prediction" rates are so high,
since we're probably not yet to mid-season.
I'm quite surprised that RPI is doing so well!
> (It's perhaps unfair to include CHODR here because this is not what it
> does. CHODR's aim is to predict goalscoring, which it does as accurately
> as possible.)
Agreed!
> Anyhow, a little late-night food for thought.
Thanks, Ken!
cheers, wayne
|
|
|