Posted by tarvolon on 11/21/2016 4:33:00 PM (view original):

Posted by kcsundevil on 11/21/2016 4:13:00 PM (view original):

Posted by pkoopman on 11/21/2016 2:31:00 PM (view original):
I think it would be a good idea for them to release some big picture data, now that 3.0 recruiting has run at least once in every world. If it is working as intended, it'd be nice to see what normal is supposed to look like.

A couple things on limited, user-run snapshots. The sample size is always going to be pretty small. And what many of us (I don't know about mully) would be looking at is top 100 battles, because that's what is the easiest and quickest to sort through. I'd guess that top 100 battles are going to skew toward the middle, i.e. a low VH vs high H, in 2-way battles. I suspect coaches are more likely to go all in on those battles, and when both guys go all in, they're likely to even out.

We've got to keep in mind that those categories are aesthetic, when you get past who is in signability range. They are not supposed to be a representation of effort credit.

Yes, I remember seble saying the difference was cosmetic. It would be useful for the current WIS staff to reconfirm that, since it has caused so much confusion.

The difference between VH and H? I'm 100% sure that wasn't cosmetic but that there was supposed to be a real difference in signing odds.

If you have only one High and one Very High .. there is a 30%-35% that the High team will sign the player.

If you have one Very High and another Very High the lower VH has 35% - 45% chance of singing the player.

mullycj .. do you agree?

Note, if you are VH and looking at difference in your odds to sign against a VH or H second team, there could be as little as almost zero difference (if the other VH was barely above the threshold and the H was really close to being a VH.. you would have a 65% chance to sign ... the most you ever will have against a H is 70%. That is NOT all that different.

Posted by hughesjr on 11/21/2016 6:01:00 PM (view original):

Posted by tarvolon on 11/21/2016 4:33:00 PM (view original):

Posted by kcsundevil on 11/21/2016 4:13:00 PM (view original):

Posted by pkoopman on 11/21/2016 2:31:00 PM (view original):
I think it would be a good idea for them to release some big picture data, now that 3.0 recruiting has run at least once in every world. If it is working as intended, it'd be nice to see what normal is supposed to look like.

A couple things on limited, user-run snapshots. The sample size is always going to be pretty small. And what many of us (I don't know about mully) would be looking at is top 100 battles, because that's what is the easiest and quickest to sort through. I'd guess that top 100 battles are going to skew toward the middle, i.e. a low VH vs high H, in 2-way battles. I suspect coaches are more likely to go all in on those battles, and when both guys go all in, they're likely to even out.

We've got to keep in mind that those categories are aesthetic, when you get past who is in signability range. They are not supposed to be a representation of effort credit.

Yes, I remember seble saying the difference was cosmetic. It would be useful for the current WIS staff to reconfirm that, since it has caused so much confusion.

The difference between VH and H? I'm 100% sure that wasn't cosmetic but that there was supposed to be a real difference in signing odds.

If you have only one High and one Very High .. there is a 30%-35% that the High team will sign the player.

If you have one Very High and another Very High the lower VH has 35% - 45% chance of singing the player.

mullycj .. do you agree?

This would go against what seble said during beta. He said the "high" range went from 0 to x.

Posted by tarvolon on 11/21/2016 6:05:00 PM (view original):

Posted by hughesjr on 11/21/2016 6:01:00 PM (view original):

Posted by tarvolon on 11/21/2016 4:33:00 PM (view original):

Posted by kcsundevil on 11/21/2016 4:13:00 PM (view original):

Posted by pkoopman on 11/21/2016 2:31:00 PM (view original):
I think it would be a good idea for them to release some big picture data, now that 3.0 recruiting has run at least once in every world. If it is working as intended, it'd be nice to see what normal is supposed to look like.

A couple things on limited, user-run snapshots. The sample size is always going to be pretty small. And what many of us (I don't know about mully) would be looking at is top 100 battles, because that's what is the easiest and quickest to sort through. I'd guess that top 100 battles are going to skew toward the middle, i.e. a low VH vs high H, in 2-way battles. I suspect coaches are more likely to go all in on those battles, and when both guys go all in, they're likely to even out.

We've got to keep in mind that those categories are aesthetic, when you get past who is in signability range. They are not supposed to be a representation of effort credit.

Yes, I remember seble saying the difference was cosmetic. It would be useful for the current WIS staff to reconfirm that, since it has caused so much confusion.

The difference between VH and H? I'm 100% sure that wasn't cosmetic but that there was supposed to be a real difference in signing odds.

If you have only one High and one Very High .. there is a 30%-35% that the High team will sign the player.

If you have one Very High and another Very High the lower VH has 35% - 45% chance of singing the player.

mullycj .. do you agree?

This would go against what seble said during beta. He said the "high" range went from 0 to x.

No .. he said high was 0 chance below 35% probability .. that at < 35% probability you go to moderate .. which is 0 chance to sign Then there is a 5% adjustment to the leader, which bumps that 35% down to 30%.

Now if you are talking more than 2 teams in the H and VH categories, then the actual probability for the H teams can go less than 30%. But in a two team battle, H is defined as a > 35% probability of getting the player (which is then adjusted down to 30% for leader credit). Very High is defined as > 40% probability .. and that would go up to 49.9%. This would also be adjusted down 5%, so to 35% - 45%.

It is just that simple.

Posted by hughesjr on 11/21/2016 6:08:00 PM (view original):

Posted by tarvolon on 11/21/2016 6:05:00 PM (view original):

Posted by hughesjr on 11/21/2016 6:01:00 PM (view original):

Posted by tarvolon on 11/21/2016 4:33:00 PM (view original):

Posted by kcsundevil on 11/21/2016 4:13:00 PM (view original):

Posted by pkoopman on 11/21/2016 2:31:00 PM (view original):
I think it would be a good idea for them to release some big picture data, now that 3.0 recruiting has run at least once in every world. If it is working as intended, it'd be nice to see what normal is supposed to look like.

A couple things on limited, user-run snapshots. The sample size is always going to be pretty small. And what many of us (I don't know about mully) would be looking at is top 100 battles, because that's what is the easiest and quickest to sort through. I'd guess that top 100 battles are going to skew toward the middle, i.e. a low VH vs high H, in 2-way battles. I suspect coaches are more likely to go all in on those battles, and when both guys go all in, they're likely to even out.

We've got to keep in mind that those categories are aesthetic, when you get past who is in signability range. They are not supposed to be a representation of effort credit.

Yes, I remember seble saying the difference was cosmetic. It would be useful for the current WIS staff to reconfirm that, since it has caused so much confusion.

The difference between VH and H? I'm 100% sure that wasn't cosmetic but that there was supposed to be a real difference in signing odds.

If you have only one High and one Very High .. there is a 30%-35% that the High team will sign the player.

If you have one Very High and another Very High the lower VH has 35% - 45% chance of singing the player.

mullycj .. do you agree?

This would go against what seble said during beta. He said the "high" range went from 0 to x.

No .. he said high was 0 chance below 35% probability .. that at < 35% probability you go to moderate .. which is 0 chance to sign Then there is a 5% adjustment to the leader, which bumps that 35% down to 30%.

I'm 99% sure you've got it wrong. I specifically asked whether there is any effort threshold where your signing odds drop from 30 (or 35) to 0. Seble said no. This is just another way of saying that the low end of the high range is significantly lower than 30%. Now when you drop below 35% effort, your signing odds drop to 0. But seble said that the signing odds don't jump from 0 to 35.

Posted by tarvolon on 11/21/2016 10:44:00 PM (view original):

Posted by hughesjr on 11/21/2016 6:08:00 PM (view original):

Posted by tarvolon on 11/21/2016 6:05:00 PM (view original):

Posted by hughesjr on 11/21/2016 6:01:00 PM (view original):

Posted by tarvolon on 11/21/2016 4:33:00 PM (view original):

Posted by kcsundevil on 11/21/2016 4:13:00 PM (view original):

Posted by pkoopman on 11/21/2016 2:31:00 PM (view original):
I think it would be a good idea for them to release some big picture data, now that 3.0 recruiting has run at least once in every world. If it is working as intended, it'd be nice to see what normal is supposed to look like.

A couple things on limited, user-run snapshots. The sample size is always going to be pretty small. And what many of us (I don't know about mully) would be looking at is top 100 battles, because that's what is the easiest and quickest to sort through. I'd guess that top 100 battles are going to skew toward the middle, i.e. a low VH vs high H, in 2-way battles. I suspect coaches are more likely to go all in on those battles, and when both guys go all in, they're likely to even out.

We've got to keep in mind that those categories are aesthetic, when you get past who is in signability range. They are not supposed to be a representation of effort credit.

Yes, I remember seble saying the difference was cosmetic. It would be useful for the current WIS staff to reconfirm that, since it has caused so much confusion.

The difference between VH and H? I'm 100% sure that wasn't cosmetic but that there was supposed to be a real difference in signing odds.

If you have only one High and one Very High .. there is a 30%-35% that the High team will sign the player.

If you have one Very High and another Very High the lower VH has 35% - 45% chance of singing the player.

mullycj .. do you agree?

This would go against what seble said during beta. He said the "high" range went from 0 to x.

No .. he said high was 0 chance below 35% probability .. that at < 35% probability you go to moderate .. which is 0 chance to sign Then there is a 5% adjustment to the leader, which bumps that 35% down to 30%.

I'm 99% sure you've got it wrong. I specifically asked whether there is any effort threshold where your signing odds drop from 30 (or 35) to 0. Seble said no. This is just another way of saying that the low end of the high range is significantly lower than 30%. Now when you drop below 35% effort, your signing odds drop to 0. But seble said that the signing odds don't jump from 0 to 35.

Mully?

I read it the same as tarvolon. If you're behind 65-35 *in effort credit*, you have 0% signing odds. Somewhere above 35% effort credit (of total between two signing range teams, not 35% of effort leader) gets you into high consideration in a 2 way battle. Say, 63-37. That discrepancy isn't signing odds, they "stretch" for signing odds. The specifics were never disclosed, that I know of. But it was implied (I think) that they stretch farther the more discrepancy there is. So a team barely making the cut for signing consideration could only be behind 63-37 in effort credit, but have much lower signing odds. I don't know how low it goes. I never saw that disclosed. I'm assuming that's basically what's being asked in the OP. I don't know of a good way for users to test that though, without a lot of collusion.

Posted by bathtubhippo on 11/21/2016 5:14:00 PM (view original):
i feel like i recall him saying that as a response to people complaining about high winning battles, saying he could remove H/VH distinction as some wanted, but it would really just be cosmetic, because the odds "behind the curtain" wouldn't change at all. basically, the odds/outcomes would be the same as with the labels, but coaches would know less going into the RNG.

I believe that this is correct.

Posted by chapelhillne on 11/21/2016 11:54:00 PM (view original):

Posted by bathtubhippo on 11/21/2016 5:14:00 PM (view original):
i feel like i recall him saying that as a response to people complaining about high winning battles, saying he could remove H/VH distinction as some wanted, but it would really just be cosmetic, because the odds "behind the curtain" wouldn't change at all. basically, the odds/outcomes would be the same as with the labels, but coaches would know less going into the RNG.

I believe that this is correct.

^^^ that.

This has some benefits, in that coaches would not feel so slighted. However, it would make it hard to know if you are ahead or behind, because the range would be higher.

Posted by pkoopman on 11/21/2016 11:54:00 PM (view original):
I read it the same as tarvolon. If you're behind 65-35 *in effort credit*, you have 0% signing odds. Somewhere above 35% effort credit (of total between two signing range teams, not 35% of effort leader) gets you into high consideration in a 2 way battle. Say, 63-37. That discrepancy isn't signing odds, they "stretch" for signing odds. The specifics were never disclosed, that I know of. But it was implied (I think) that they stretch farther the more discrepancy there is. So a team barely making the cut for signing consideration could only be behind 63-37 in effort credit, but have much lower signing odds. I don't know how low it goes. I never saw that disclosed. I'm assuming that's basically what's being asked in the OP. I don't know of a good way for users to test that though, without a lot of collusion.

This is how I read it. So, to make up total guess numbers, a 64/36 battle in effort might be 95/5 in signing odds. Whereas a 60/40 battle in effort might be 70/30 in signing odds. But both would be VH vs H.

But if this were the case (which, unless seble or I was extremely confused, it is supposed to be), we would expect to see H only winning (in a two-way battle) something like 25% of the time (because some of the VH vs H battles may be 65/35, but others may be 95/5), and certainly not almost 50% of the time, as has been indicated earlier in this thread.

What we need is some good data. Probably only look at two-team battles and get a good sense of the winning percentage.

Posted by tarvolon on 11/21/2016 10:44:00 PM (view original):

Posted by hughesjr on 11/21/2016 6:08:00 PM (view original):

Posted by tarvolon on 11/21/2016 6:05:00 PM (view original):

Posted by hughesjr on 11/21/2016 6:01:00 PM (view original):

Posted by tarvolon on 11/21/2016 4:33:00 PM (view original):

Posted by kcsundevil on 11/21/2016 4:13:00 PM (view original):

Posted by pkoopman on 11/21/2016 2:31:00 PM (view original):
I think it would be a good idea for them to release some big picture data, now that 3.0 recruiting has run at least once in every world. If it is working as intended, it'd be nice to see what normal is supposed to look like.

A couple things on limited, user-run snapshots. The sample size is always going to be pretty small. And what many of us (I don't know about mully) would be looking at is top 100 battles, because that's what is the easiest and quickest to sort through. I'd guess that top 100 battles are going to skew toward the middle, i.e. a low VH vs high H, in 2-way battles. I suspect coaches are more likely to go all in on those battles, and when both guys go all in, they're likely to even out.

We've got to keep in mind that those categories are aesthetic, when you get past who is in signability range. They are not supposed to be a representation of effort credit.

Yes, I remember seble saying the difference was cosmetic. It would be useful for the current WIS staff to reconfirm that, since it has caused so much confusion.

The difference between VH and H? I'm 100% sure that wasn't cosmetic but that there was supposed to be a real difference in signing odds.

If you have only one High and one Very High .. there is a 30%-35% that the High team will sign the player.

If you have one Very High and another Very High the lower VH has 35% - 45% chance of singing the player.

mullycj .. do you agree?

This would go against what seble said during beta. He said the "high" range went from 0 to x.

No .. he said high was 0 chance below 35% probability .. that at < 35% probability you go to moderate .. which is 0 chance to sign Then there is a 5% adjustment to the leader, which bumps that 35% down to 30%.

I'm 99% sure you've got it wrong. I specifically asked whether there is any effort threshold where your signing odds drop from 30 (or 35) to 0. Seble said no. This is just another way of saying that the low end of the high range is significantly lower than 30%. Now when you drop below 35% effort, your signing odds drop to 0. But seble said that the signing odds don't jump from 0 to 35.

Agreed.

I am so confused.

I know that VH and H each have a chance.

I know that VH has a better chance than H.

I know that mod has no chance.

I know that the actual probabilities vary from case to case.

https://www.youtube.com/watch?v=KX5jNnDMfxA

When y'all have reached common ground, could you write down what you think the 30-35-40% stuff comes out to in a couple of examples - like two person VH v H and three person VH, VH, H?

and thanks!

It sounds like the next updates are going to offer us something in the way of post-battle data.

Posted by tarvolon on 11/22/2016 5:02:00 PM (view original):

Posted by pkoopman on 11/21/2016 11:54:00 PM (view original):
I read it the same as tarvolon. If you're behind 65-35 *in effort credit*, you have 0% signing odds. Somewhere above 35% effort credit (of total between two signing range teams, not 35% of effort leader) gets you into high consideration in a 2 way battle. Say, 63-37. That discrepancy isn't signing odds, they "stretch" for signing odds. The specifics were never disclosed, that I know of. But it was implied (I think) that they stretch farther the more discrepancy there is. So a team barely making the cut for signing consideration could only be behind 63-37 in effort credit, but have much lower signing odds. I don't know how low it goes. I never saw that disclosed. I'm assuming that's basically what's being asked in the OP. I don't know of a good way for users to test that though, without a lot of collusion.

This is how I read it. So, to make up total guess numbers, a 64/36 battle in effort might be 95/5 in signing odds. Whereas a 60/40 battle in effort might be 70/30 in signing odds. But both would be VH vs H.

But if this were the case (which, unless seble or I was extremely confused, it is supposed to be), we would expect to see H only winning (in a two-way battle) something like 25% of the time (because some of the VH vs H battles may be 65/35, but others may be 95/5), and certainly not almost 50% of the time, as has been indicated earlier in this thread.

What we need is some good data. Probably only look at two-team battles and get a good sense of the winning percentage.

Yes - agree.

To be conservative, VH should win >66% of the time (more than 2 out of every 3 battles).

Also would like to see from CS if this is working as intended.