Weight of SOS on projections....possible example. Topic

Yesterday in DIII Rupp my team was listed 2nd on the post season projection report, with 24-0 record, rpi of 6 and sos of 32. After today's round of games, an off day for my squad which moved our rpi to 5 and sos to 31, we got bumped to 3rd by a team that lost their game.

Although I can't recall with 100% accuracy the rpi, sos, and projection of the team that surpassed us, I'm pretty confident they had previously been projected at 3rd. After lossing at home to an opponents that now has a 9 rpi, 14 sos, and 22-2 record, the new team projected 2nd has an rpi of 1, sos of 1, and 20-4 record.

my team:

http://www.whatifsports.com/hd/TeamProfile/Ratings.aspx?tid=4773

team projected at 2nd today:

http://www.whatifsports.com/hd/TeamProfile/Ratings.aspx?tid=4199

I'm not making any judgments or claiming this is direct evidence! However, there has been speculation that losing to a team with a high rpi may have more positive weght in these matters than they possibly should. This is just another piece to the puzzle, do with it what you will.

1/27/2012 5:05 AM
Yea I think SOS is weighted much heavier but it also could have been a number of things like maybe a good RPI team on your sched. lost to a low RPI team and a good RPI team on his sched. won against a high RPI team.  I think what it comes down to, as a lot of people have metioned is RPI 1-50 record and key wins which he has had more, and as you said his higher SOS. 
1/27/2012 3:43 PM
You also might have been ahead only a miniscule amount, so the effect might not be large.
1/27/2012 6:48 PM
Posted by m4284850 on 1/27/2012 3:43:00 PM (view original):
Yea I think SOS is weighted much heavier but it also could have been a number of things like maybe a good RPI team on your sched. lost to a low RPI team and a good RPI team on his sched. won against a high RPI team.  I think what it comes down to, as a lot of people have metioned is RPI 1-50 record and key wins which he has had more, and as you said his higher SOS. 
This can be huge. If a team you played was 49 and drops to 51 you will see a big drop in your projection, especially if it's a conference mate you beat twice
1/28/2012 11:03 AM
The projection report doesn't "factor in" SOS for a formula or anything, the big thing with the new seeding is that each game is "graded" as in you get points for how well you play.  
1/28/2012 12:06 PM
Posted by tkimble on 1/28/2012 12:06:00 PM (view original):
The projection report doesn't "factor in" SOS for a formula or anything, the big thing with the new seeding is that each game is "graded" as in you get points for how well you play.  

If this is true then why did a losing team pass me? Seems like you wouldn't get that many "points" in a losing effort. Especially since the computer considered them a 1 point favorite (whether or not that has anything to do with it).
 

1/29/2012 5:56 PM
Because of other losses by other teams, other games scores changed.  perhaps the teams on your schedule lost more games than the teams on his schedule that week.
1/29/2012 7:13 PM
Posted by arssanguinus on 1/29/2012 7:13:00 PM (view original):
Because of other losses by other teams, other games scores changed.  perhaps the teams on your schedule lost more games than the teams on his schedule that week.
Ya, that is totally plausible, I considered that myself and others have pointed it out as well. I just don't think the "grading system" would be able to account for this scenario (nevermind this post if you were just making a statement, rather than responding to my last comment).
1/29/2012 8:01 PM
I think it does.  The "grading" is a rolling system, so each game's "score" changes every cycle, at least as I understand the system.
1/29/2012 8:45 PM
Posted by nachopuzzle on 1/29/2012 8:01:00 PM (view original):
Posted by arssanguinus on 1/29/2012 7:13:00 PM (view original):
Because of other losses by other teams, other games scores changed.  perhaps the teams on your schedule lost more games than the teams on his schedule that week.
Ya, that is totally plausible, I considered that myself and others have pointed it out as well. I just don't think the "grading system" would be able to account for this scenario (nevermind this post if you were just making a statement, rather than responding to my last comment).
But it could. If each game is graded, and part of the grading of each game has to do with the relative position of each team and how well it did, and how well you did against them, then by nature, if your guys had a bad week and his had a good week, it woudl effect your game scores, possibly considerably. For example; did a few of your division mates loose - perhaps ones that you has played two games against?  Could a couple of them have suffered upsets?  The entire rest of the schedules changing COULD possibly outmatch the results of a single week - even if you won and they lost.  Why woudl you think it couldn't account for this?
1/29/2012 8:48 PM
Given that we don't know exactly how it works we have to ask ourselves what is the more probable cause for this event. So, which is more likely: (a) many of those things happening in one day, creating enough enough points to surpass the higher projected team despite the negative impact of that days loss, or (b) the system overvalues losing to teams with a high rpi?

You are aware this situation happened over the course of one round of game? I just want to make sure because you keep referring to weeks. Plus, I remember that day and only like one of my opponents moved out of the top 100 rpi. I can't say your wrong, but I'm doubting whether those factors could make up enough ground in one day despite having to compensate for the corresponding lose.
1/29/2012 10:07 PM
You said that they were ONE slot behind you. . and you don't know how far behind.  They could have been barely even behind without much ground to make up.  All you have is a rank.  You don't have the number to go with it.  ANd what do you mean "Weeks"  In fact. . I didn't use a plural for week in the entire post.

"Your guys had a bad WEEK", his had a good WEEK', "The results of a single week"
 

 

Please, show me where I keep referring to 'weeks'?

 

 


1/29/2012 10:50 PM
(A Note:  When I say "Week" I am referring to one round of games, if not clear from context)
1/29/2012 10:57 PM
nacho, a couple nights later also in Rupp (DI), a team in my conference was at 56 in the projection report and lost on the road to the #1 RPI team, and promptly slid to #62. If a "good" loss (particularly on the road) very clearly moved you to a better place, this wouldn't have happened.
1/29/2012 11:01 PM
"Given that we don't know exactly how it works we have to ask ourselves what is the more probable cause for this event. So, which is more likely: (a) many of those things happening in one day, creating enough enough points to surpass the higher projected team despite the negative impact of that days loss, or (b) the system overvalues losing to teams with a high rpi?"

Honestly, I think you're completely undervaluing how much can change with 10+ teams from your schedule playing other teams.  If your non-con schedule goes 2-8 in a given night, that can significantly affect a lot of things.  I don't know what seble's games scores are or how they work, but I know that one weird night can change a lot, and I
 certainly wouldn't jump to conclusions and attribute an odd move on the projection report to overvaluing losing to good teams.
1/29/2012 11:18 PM
12 Next ▸
Weight of SOS on projections....possible example. Topic

Search Criteria

Terms of Use Customer Support Privacy Statement

© 1999-2026 WhatIfSports.com, Inc. All rights reserved. WhatIfSports is a trademark of WhatIfSports.com, Inc. SimLeague, SimMatchup and iSimNow are trademarks or registered trademarks of Electronic Arts, Inc. Used under license. The names of actual companies and products mentioned herein may be the trademarks of their respective owners.