realtimerpi.com missing at least one game

#1
ukvols

ukvols

Senior Member
Joined
Feb 16, 2005
Messages
13,892
Likes
1,247
#1
Drexel's win at Cleveland State on 2/18 is missing. I am data checking my own rankings before posting, and I found this one missing from realtimerpi.com. I don't know yet if it's just this game. The fact that it was a BracketBuster game and thus not on the schedule at the beginning of the year probably caused them to miss it. I'll give them credit for the information they provide; I did go there to compare data.

This isn't a HaHa, look what I found post. It's more to point out that the data in these systems isn't always going to be complete.

I recall a story from a recent football season in which a formula used by the BCS either missed a game or reported it wrong.

Edit: Butler's win over Indiana State on 2/18 is missing.
UC Riverside's win over Southern Utah on 2/18 is missing.
UC Irvine's win over Eastern Washington on 2/18 is missing.
Wagner's win over Fairleigh Dickinson on 2/18 is missing.
Robert Morris' win over Bryant on 2/16 is listed twice for RMU and counted in the RPI as 2 wins for RMU but just the 1 loss for Bryant.
Mt. St. Mary's' win over St. Francis (PA) on 2/2 is listed twice for both teams, as the same game is re-listed on 2/4. It's counted as 2 wins for MSM and 2 losses for St.F.

There are also a few from this past weekend that are missing, but I'll chalk those up to the site simply not being updated yet, even though it says it was last updated on Monday.
 
Last edited:
#5
#5
Volman09 said:
Who to believe, you or the NCAA? hmm, tough decision. :p
Click to expand...

Haha. I figured I'd get that reaction when we didn't climb as high as in the RPI. I obviously don't use the same formula as the RPI.

My formula adds some weight to the winning % when the final number is calculated to keep teams from staying ranked too high even when they keep losing, just because their SOS grades out high. SOS is important, but I think the RPI gives it a little too much weight, and it leads to a battle of whose cupcakes beat whose. Unfortunately, in a season where we lose 13 games, that part of the formula isn't fun.

I do have us as the highest-ranked 13-loss team, just ahead of Pitt and their 14 losses at 109.
 
Last edited:
#6
#6
ukvols said:
Haha. I figured I'd get that reaction when we didn't climb as high as in the RPI. I obviously don't use the same formula as the RPI.

My formula adds some weight to the winning % when the final number is calculated to keep teams from staying ranked too high even when they keep losing, just because their SOS grades out high. SOS is important, but I think the RPI gives it a little too much weight, and it leads to a battle of whose cupcakes beat whose. Unfortunately, in a season where we lose 13 games, that part of the formula isn't fun.

I do have us as the highest-ranked 13-loss team, just ahead of Pitt and their 14 losses at 109.
Click to expand...

Lol. I really did chuckle at being the highest ranked 13 loss team on your board.
 
#7
#7
As a test, I just changed the 4-pt loss to APSU to a 4-pt win, and we jumped from 107 to 92 on my board.

I'll try to update my website tomorrow and post the link. I've been doing the rankings since '06, but the website is still fairly crude, with the ad placement a little goofy.

Website is experimental for now. I know a lot more about ranking teams than web design.
 
Last edited:
#8
#8
ukvols said:
As a test, I just changed the 4-pt loss to APSU to a 4-pt win, and we jumped from 107 to 92 on my board.

I'll try to update my website tomorrow and post the link. I've been doing the rankings since '06, but the website is still fairly crude, with the ad placement a little goofy.

Website is experimental for now. I know a lot more about ranking teams than web design.
Click to expand...

Volnation will decide that :)
 
#10
#10
"The other 25" are using the same formula.
I'm not going to argue that though; I prefer to think we're 92 as well.
 
#11
#11
ukvols said:
"The other 25" are using the same formula.
I'm not going to argue that though; I prefer to think we're 92 as well.
Click to expand...

No disrespect intended.
We need to see all the variations using different algorithms.
Anticipating your link.
 
#12
#12
None taken. I'll post tonight after work.
I cringed a little when we didn't climb into the Top 100 as I finished getting all of the scores in, because I knew it would make anything I posted less popular here. Unlike the RPI, I don't have UK below Duke and their four losses. I keep UK from being underrated by placing a greater value on W/L, but the side effect is 15-13 teams having a hard time moving up.
 
#14
#14
OrangeBalls said:
Pomeroy has us at 65.
Click to expand...

That sounds great, but I don't see how anyone can honestly say we deserve to be 65. I don't see how any computer could have us that high, unless it's subjectively applying less weight to every game before Stokes started playing. Even odder is Lunardi's version of the RPI having Tennessee at 37 last week. LOL, I wish we were 37.
 
#15
#15
The only difference between Lunardi and the rest of the bracketologists is the ESPN next to his name. All bracketology is is a guessing game and he hasn't been the best at it.
 
#16
#16
ukvols said:
That sounds great, but I don't see how anyone can honestly say we deserve to be 65. I don't see how any computer could have us that high, unless it's subjectively applying less weight to every game before Stokes started playing. Even odder is Lunardi's version of the RPI having Tennessee at 37 last week. LOL, I wish we were 37.
Click to expand...

Don't feel bad. I have the Vols at 115 (Crotistics), and I would much prefer them to be higher as well. The real issue obviously is that overall the Vols are only 2 games over .500 against D-1 competition. Before playing Ole Miss and South Carolina, the RPI number was around 110. Just hoping something happens over the next week that can bump the Vols to the 3 seed for the SEC tournament.
 
#17
#17
Cro said:
Don't feel bad. I have the Vols at 115 (Crotistics), and I would much prefer them to be higher as well. The real issue obviously is that overall the Vols are only 2 games over .500 against D-1 competition. Before playing Ole Miss and South Carolina, the RPI number was around 110. Just hoping something happens over the next week that can bump the Vols to the 3 seed for the SEC tournament.
Click to expand...

Just win and let nature take place. No way Florida beats Vandy at memorial and Kentucky will likely win. Pretty sure that wraps it up for us.
 
#18
#18
ukvols said:
That sounds great, but I don't see how anyone can honestly say we deserve to be 65. I don't see how any computer could have us that high, unless it's subjectively applying less weight to every game before Stokes started playing. Even odder is Lunardi's version of the RPI having Tennessee at 37 last week. LOL, I wish we were 37.
Click to expand...

Pomeroy uses different weights and a lot of people think his is more accurate than the rpi. Who knows what the committee all uses, but I know they use rpi and look at who has been on the team and what they did down the stretch. We're 8-5 since Stokes started playing.

I'd like to know what the hell Hall did to get such a lengthy suspension.
 
#19
#19
OrangeBalls said:
Pomeroy uses different weights and a lot of people think his is more accurate than the rpi. Who knows what the committee all uses, but I know they use rpi and look at who has been on the team and what they did down the stretch. We're 8-5 since Stokes started playing.

I'd like to know what the hell Hall did to get such a lengthy suspension.
Click to expand...

Did you ever see Cheech and Chong?
 
#23
#23
OrangeBalls said:
Pomeroy uses different weights and a lot of people think his is more accurate than the rpi. Who knows what the committee all uses, but I know they use rpi and look at who has been on the team and what they did down the stretch. We're 8-5 since Stokes started playing.

I'd like to know what the hell Hall did to get such a lengthy suspension.
Click to expand...

I don't disagree with the fact that we're better now with Stokes. I just don't see how you can objectively account for that in a computer poll. For example, by the same token, it seems like our loss to Bama would be adjusted to be an even worse loss for us, since they were missing some guys on suspension, and yet we still lost. I also question how someone can keep up with every injury, suspension, and other absence over all 344 teams. Perhaps they don't consider suspensions, and it's just injuries and Stokes' situation; I don't know. But, still, what happens to a team that loses to a team that's missing a guy? I just don't see how you can devalue games for missing players without injecting a large amount of subjectivity into the formula.
 
#24
#24
I don't know that any of the computer formulas actually take this into account, my understanding was its the committee themselves. When they go in and have their list of teams for at large bids, my understanding is that they then take into consideration losses/additions to each team.

With that said, they DO NOT take into consideration suspensions what do ever, and I don't think they put much weight on one or two game injuries. I believe it has to be a lengthy injury or in our case a lengthy addition of a new player, for them to really put much weight on it.

TIFWIW
 
New posts
You must log in or register to reply here.

VN Store



Back
Top