Posted by zorzii on 10/1/2016 10:45:00 AM (view original):
If the Vh split chances and the lone h is not affected accordingly, it could swing the balance towards H.

seble addressed that in beta - Here's the posts from z0601, elmo and seble:

z0601
Posts: 743 (0)
Veteran
Add Favorite
Block User
So we can get a better gauge on the percentage in a multiple team battle, lets us the data that was posted here: http://test.whatifsports.net/whatifsports/forums/Posts.aspx?TopicID=420740&TopicsTimeframe=30

In this three team battle, the lowest considered team won. Seble, what was the percentage of Bucknell winning this battle? The reason I bring this up or the question I have is, did the two teams at VH split the 80% chance of signing and the High team keep same percentage? So in another words, the two VH percentage being 40% and the High team at 20%?
7/22/2016 10:27 AM


seble
Posts: 2480 (0)
Site Staff
Remove Favorite
La Salle had a 40.5% chance, St. Joe's had a 36.6% chance, and Bucknell 22.8%.
7/22/2016 10:31 AM


seble
Posts: 2480 (0)
Site Staff
Remove Favorite
Bucknell was on the higher end of the High range in that case.
7/22/2016 10:33 AM


elmossle
Posts: 46 (0)
Hall of Famer
Add Favorite
Block User
So using this as an example, are we saying that in a 3 or more team battle, those at High consideration have an increased chance of signing the player than those at VH? If that is indeed the case, then we might look at increasing the odds a little bit. It just seems strange to me for a team like La Salle to have almost a 2 to 1 advantage lose the battle. I am all for increasing battles and having the lower school win a percentage of the time, but this example looks to me that we need to make an adjustment of some kind.
7/22/2016 10:40 AM


seble
Posts: 2480 (0)
Site Staff
Remove Favorite
Not sure what you mean by your first sentence. Having a third team is going to proportionally lower the other teams' odds.
7/22/2016 10:44 AM


elmossle
Posts: 46 (0)
Hall of Famer
Add Favorite
Block User
Sorry for my poor wording, let me try again. In looking at the data you provided, it appears to me that the two VH teams "split" the 80% chance of signing this player if it was a race between 2 schools, 1 at VH and 1 at High. While those two VH schools split the 80%, the third team who was at High, did not split the remaining 20% with any other team thus increasing their chances and probability of signing the recruit. Is this correct or am I misreading the data?
7/22/2016 10:49 AM


seble
Posts: 2480 (0)
Site Staff
Remove Favorite
No, that's not correct. If we removed St. Joe's, it would've been 64/36, so adding in that third team reduces odds of both teams.
7/22/2016 10:51 AM

A low. Very high and a high high might be right next to eachother in recruiting value ...

Posted by arssanguinus on 10/1/2016 11:04:00 AM (view original):
A low. Very high and a high high might be right next to eachother in recruiting value ...

Yes, I think people are using very simplistic terminology to represent what is probably a very intricate "formula". I'm not sure formula is the right term, either.

I should have clarified, when I used the 35-35-30 scheme, I was talking about effort, not signing odds. In a 3 way race, it is likely that the "high" is relatively close to the others in effort, at least closer than many of the 2-way battles.

And we need to keep in mind that the discrepancy in effort is not the same as discrepancy in signing odds.

Good stuff, possum. We just learned too important things there:

1. A 64/36 split is very high vs high (although it is unclear whether this was before seble boosted the odds of the leading team. anyone want to sleuth it out?)
2. The addition of the third team reduces everyone's changes proportionally. The top very high team would've had 77% better odds than the high team in a two team battle, and the top very high team had 77% better odds than the high team in the three-team battle.

Posted by pkoopman on 10/1/2016 11:21:00 AM (view original):

Posted by arssanguinus on 10/1/2016 11:04:00 AM (view original):
A low. Very high and a high high might be right next to eachother in recruiting value ...

Yes, I think people are using very simplistic terminology to represent what is probably a very intricate "formula". I'm not sure formula is the right term, either.

I should have clarified, when I used the 35-35-30 scheme, I was talking about effort, not signing odds. In a 3 way race, it is likely that the "high" is relatively close to the others in effort, at least closer than many of the 2-way battles.

And we need to keep in mind that the discrepancy in effort is not the same as discrepancy in signing odds.

This is the one area in which I felt the most bad for seble. He put in the boost for the leading team specifically to head off this sort of criticism (and was right to do so, IMO), and in doing so set off a firestorm of criticism based on people assuming that somebody who only has half the signing odds is actually massively behind in effort.

Posted by tarvolon on 10/1/2016 11:22:00 AM (view original):
Good stuff, possum. We just learned too important things there:

1. A 64/36 split is very high vs high (although it is unclear whether this was before seble boosted the odds of the leading team. anyone want to sleuth it out?)
2. The addition of the third team reduces everyone's changes proportionally. The top very high team would've had 77% better odds than the high team in a two team battle, and the top very high team had 77% better odds than the high team in the three-team battle.

I believe he made changes to the odds AFTER this discussion (it appears he made changes in early August and the discussion above was in July). It's also possible that cwis and redhawk have made further changes since taking over.

I also came across this little tidbit that seble posted on 8/6 ... thought this was interesting to pass along.

2. If credit is 65/35, the team at 35 wouldn't even be at High interest, so odds would be 100%.

Posted by possumfiend on 10/1/2016 11:55:00 AM (view original):

Posted by tarvolon on 10/1/2016 11:22:00 AM (view original):
Good stuff, possum. We just learned too important things there:

1. A 64/36 split is very high vs high (although it is unclear whether this was before seble boosted the odds of the leading team. anyone want to sleuth it out?)
2. The addition of the third team reduces everyone's changes proportionally. The top very high team would've had 77% better odds than the high team in a two team battle, and the top very high team had 77% better odds than the high team in the three-team battle.

I believe he made changes to the odds AFTER this discussion (it appears he made changes in early August and the discussion above was in July). It's also possible that cwis and redhawk have made further changes since taking over.

I also came across this little tidbit that seble posted on 8/6 ... thought this was interesting to pass along.

2. If credit is 65/35, the team at 35 wouldn't even be at High interest, so odds would be 100%.

Possum so is your interpretation that moderate vs very high is 0/100% odds?

Yeah Moderate has no chance of signing.

Possum- thanks for digging that up. I remember that example with Bucknell. I think the reduction of Bucknell's chances from 36 to 22% with the introduction of another team makes sense.

And thanks for posting about the credit split too. That's helpful.

Posted by oldresorter on 10/1/2016 12:12:00 PM (view original):

Posted by possumfiend on 10/1/2016 11:55:00 AM (view original):

Posted by tarvolon on 10/1/2016 11:22:00 AM (view original):
Good stuff, possum. We just learned too important things there:

1. A 64/36 split is very high vs high (although it is unclear whether this was before seble boosted the odds of the leading team. anyone want to sleuth it out?)
2. The addition of the third team reduces everyone's changes proportionally. The top very high team would've had 77% better odds than the high team in a two team battle, and the top very high team had 77% better odds than the high team in the three-team battle.

I believe he made changes to the odds AFTER this discussion (it appears he made changes in early August and the discussion above was in July). It's also possible that cwis and redhawk have made further changes since taking over.

I also came across this little tidbit that seble posted on 8/6 ... thought this was interesting to pass along.

2. If credit is 65/35, the team at 35 wouldn't even be at High interest, so odds would be 100%.

Possum so is your interpretation that moderate vs very high is 0/100% odds?

Yes

Bottom line: WIS needs to do a better job of explaining VH vs. H.

Posted by possumfiend on 10/1/2016 11:55:00 AM (view original):

Posted by tarvolon on 10/1/2016 11:22:00 AM (view original):
Good stuff, possum. We just learned too important things there:

1. A 64/36 split is very high vs high (although it is unclear whether this was before seble boosted the odds of the leading team. anyone want to sleuth it out?)
2. The addition of the third team reduces everyone's changes proportionally. The top very high team would've had 77% better odds than the high team in a two team battle, and the top very high team had 77% better odds than the high team in the three-team battle.

I believe he made changes to the odds AFTER this discussion (it appears he made changes in early August and the discussion above was in July). It's also possible that cwis and redhawk have made further changes since taking over.

I also came across this little tidbit that seble posted on 8/6 ... thought this was interesting to pass along.

2. If credit is 65/35, the team at 35 wouldn't even be at High interest, so odds would be 100%.

That is what I was trying to remember in another thread--I think he posted it in response to a question that I asked.


So the cutoff for high/medium in terms of recruiting credit (in a two team battle) is somewhere between 35 and 45% (my guess is it's in the 38-40% range, but that's just a guess based on how I'd program it). The cutoff between high and medium for signing odds, of course, is 0%

I think he posted it in response to a question that I asked.

He did.