Re: [hockeydisk] Disk Concers
- A 192 seasons....I have a hard enough time running one....
If it takes longer Dave and Jeff, that is just fine. People have to realize
that not every player will perform exactly 100% to the NHL in their leagues
because the lines are not 100% of the same. Even in a perfect simmed year,
as there are trades, injuries etc.
The 10% high and low range, sounds more than acceptable, especially if it
sometimes it is high and sometimes it is low.
Much appreciation is given from our league again.
>From: "D. Atkinson" <bigdog@...>_________________________________________________________________
>Subject: Re: [hockeydisk] Disk Concers
>Date: Sun, 02 Sep 2001 14:19:05 -0600
>Greg Andrade wrote:
> > I was just wondering if there is any news on the data disk. The last
>message someone sent about
> > this went without a response and seemed to disappear from the mailing
>The data disk is about done. I'm on the 24th set of test seasons and
>tweaking, and everything is
>just about in perfectly. The Isles always win too many games, and Adam
>Oates always takes
>too many shots, but other than that things are looking good.
>By the way, the last message someone sent out on this got answered. I just
>did it privately. So,
>it didn't go without a response as stated.
> > I've had several people contact me with concers about the disk,
>especially since the OTHL site
> > which normally hosts the data disk is dead right now.
>That's really nice that posted a "several people contacted me with
>concerns" email again this year.
>By now, those "people" should really learn that they'd get their concerns
>answered if they
>emailed Jeff or I. You're obviously not helping them too much. Jeff will
>have the download
>site up and ready when the disk is finished.
>I'm pretty happy with the disk this year again. It is as good as last
>year's...but I am learning
>the variability of the APBA engine makes it impossible to get everyone
>perfect. Some players
>will be 15% low on shots on one test set and then be 10% high on the next.
>I set my mark
>to be within 10% on things, and we came really close, except for the
>players with <20 games
>who vary even more. This disk is a perfect representation of the NHL given
>the contraints of
>the APBA engine.
>Just so everyone knows, I ran 24 test sets of 8 seasons per set. I adjusted
>hiddens based on the results, and reran to check my tweaks. Some players
>tweaks, others fell in quickly. This process eventually gets out any
>ratings anomalies and
>leaves a accurate representation of the past NHL season. Again, it was more
>I remember, but it was worth it if the users find the final product to
>I'll get Jeff my final data set in the next 24 hours. He will adjust the
>ratings for the league disk
>(like melding multiple team players with a games played weighted rating
>set), as I
>always test on the split disk for accuracy's sake.
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp