RPI...SOS

[rosconey]

who the @ck cares- this team could get a 16 seed and still win it all -

 

[jordoo]

who the @ck cares- this team could get a 16 seed and still win it all -

Hey as long as its in Buff/MSG I'm cool with that.

 

[Marsh01]

who the @ck cares- this team could get a 16 seed and still win it all -

I agree. I pay very little attention to rpi and sos and all of the other sites when we have seasons like this.

 

[jncuse]

the RPI/SOS is broken. of the top 30 teams how many have losses to teams below 100-150? does it really matter if SU plays team 150 or team 300 they should win more than 99.9% of the time. so the formula shouldnt differentiate it so much like it does. if one team plays 150-159 and one team plays 300-309 the reality is they both played the same cupcake schedule but the RPI/SOS will say one team played a much tougher schedule..
.

You have figured out the secret to the RPI game. Avoid sub 200 games.

 

[jncuse]

when are people going to realize that we will be at MSG as long as we are a top 4 seed, just like last year

As a 4 seed, I really doubt we would be in MSG. MSG is the home for the #1 ACC team, and if we are a 4 seed, I suspect some other team in the ACC would be a 3.

Remember that three top teams from the ACC must be in a different region.

But I really doubt we are a 4 seed.

 

[jncuse]

Nova hopped Cuse overnight in KenPom's rankings.

Its .9515 vs .9510. Really close - if Nova wins by a point or two less than expected against Creighton, we probably jump back ahead. But its a bigger gap vs #4, which would need to make up about 15-18 points from expectation to catch Cuse or Nova.

 

[Knicks411]

As a 4 seed, I really doubt we would be in MSG. MSG is the home for the #1 ACC team, and if we are a 4 seed, I suspect some other team in the ACC would be a 3.

Remember that three top teams from the ACC must be in a different region.

But I really doubt we are a 4 seed.

Yeah, last year was different, we were the 4th team out of the BE, so we could be placed with Marquette. If we are second or third team out of the ACC this year, not likely we get MSG

 

[jncuse]

We win the ACC with no more than 4 losses and we will be the #1 seed in the East regardless of our RPI. Take it to the bank.

http://www.rpiforecast.com/index2.html

Its a great resource for seeing our #1 seed chances now that they project us at 4 losses.

You can see what we could be up against with 4 losses pre conference tournament. In this case our RPI is 7, and SOS is 50.

In the scenario above, Arizona and Michigan St would be #1's, and its a battle between Syracuse, Nova, Kansas and Wisconsin for the last two #1 seed.

 

[Knicks411]

How do they project the games out?

 

[jncuse]

With regard to the impact of the sub 200 games we can see it in the final SOS numbers.

Duke 8
Florida St 23
Syracuse 50
Pitt 72

That is not only due to sub 200. Duke played a few more top 25 teams as well. At the end of the day Syracuse may have only played 1.

 

[jncuse]

How do they project the games out?

They use win probabilities based on Sagarin.

 

[jncuse]

Largest board blow up scenario ever:

Syracuse goes 13-5 shares the regular season title, loses in conference tourney semi. It slips out of 1 seed discussion.
Duke goes 12-6 in the regular season and loses the conference tourney final.

Syracuse is projected as a 2/3, and ends up as 3.
Duke is a borderline 3/4, and ends up as 3.

Duke gets MSG

 

[Knicks411]

They use win probabilities based on Sagarin.

Was just curious, thanks. Was just checking Pomeroy, we're not up to 15-3 on the ACC projection, fwiw.

 

[LennX6]

With regard to the impact of the sub 200 games we can see it in the final SOS numbers.

Duke 8
Florida St 23
Syracuse 50
Pitt 72

That is not only due to sub 200. Duke played a few more top 25 teams as well. At the end of the day Syracuse may have only played 1.

At the end of the day Syracuse may have only played 1
There is an inherent fallacy with these data based rating systems. Every datapoint has several other independent variables. At the end of the day it still is only a guess, but they can call it "data-driven" and somehow noone questions its validity.
The point I want to make is that when you say "at the end of the day..." , one of the factors why we have only played 1 Top 25 team is because all our opponents have one thing in common, we have beaten them. Is that fact in there somewhere in the data or when you get to that deep level is it then lost?

 

[omniorange]

I will preface this post by saying that we pass the "eye test" and even Kenpom has Syracuse ranked 2nd... Which is probably the most important statistical evaluation of a team.

But our RPI and SOS schedule leave something to be desired for a 18-0 team, whose main goal should be to get the 1 seed in the East.

Our RPI is only 8th (or 4th depending on which site you prefer) and that is really driven by SOS which is currently 66th (or 101 depending one which site you prefer)
And according to ESPN's BPI (ranked 4th in the BPI overall) our SOS is 82.

With all the talk about how this might be the best schedule we have ever played, one might think our SOS would be much higher.. But if you dig deeper into those numbers there is on reason for that. The bad teams on our schedule were really really really bad.
We have played 6 teams with RPIs 200 worse, and that is what is really killing our numbers. 3 of which are 300 or worse
If you compare that to teams that the RPI loves
Kansas 0 games against 200 or less
Wisky 2 games 200 or less 0 vs 300 or less
Zona 1 game vs 200 or less 0 vs 300 or less
UMass (ranked 5th in RPI) 1 game vs 200 or less (that team was 208)

And you can see where our terrible RPI/SOS is coming from, but of course it will be a moot point if we continue to win.
I know we still have a lot of good games on our schedule which should help our numbers a bit (so does everyone) but these bad teams we have played will continue to kill our numbers.
Again we pass the eye test, but if we have a hiccup or 4 down the road these numbers might come back to bite us, and force us to make a tourney run in some bracket that doesn't run through MSG.

Top seeds are not decided by these factors though, not even by overall RPI, since rarely are the Top 8 RPI teams the top 8 seeds in the tourney.

As of this moment in time, our Top 50 wins compare favorably to who will be the likely competition for those overall top 8 seeds. And we still have two cracks at Duke, another crack at Pitt, plus UVa and FSU to build upon those five victories, all currently Top 25 and unlikely to fall below Top 50. If we go at least 3-2 in those 5 games, even a What loss to someone out of the Top 100 in the ACC isn't likely to cause us to lose an East 1 or 2 seed.

Cheers,
Neil

 

[omniorange]

Its .9515 vs .9510. Really close - if Nova wins by a point or two less than expected against Creighton, we probably jump back ahead. But its a bigger gap vs #4, which would need to make up about 15-18 points from expectation to catch Cuse or Nova.

And we own the win over Nova. So the committee would weigh that more heavily than either RPI or Kenpom standing if Nova and Cuse remain close to each other and near the top of both as well.

Cheers,
Neil

 

[Knicks411]

At the end of the day Syracuse may have only played 1
There is an inherent fallacy with these data based rating systems. Every datapoint has several other independent variables. At the end of the day it still is only a guess, but they can call it "data-driven" and somehow noone questions its validity.
The point I want to make is that when you say "at the end of the day..." , one of the factors why we have only played 1 Top 25 team is because all our opponents have one thing in common, we have beaten them. Is that fact in there somewhere in the data or when you get to that deep level is it then lost?

I doubt it's going to make that much of a difference over the course of a season.

But we've played more than 1 top 25 team, at least if we're going by RPI. Baylor is 24th, Pitt is 17th, Nova is 4. Minnesota and Cal are right on the verge

 

[omniorange]

Largest board blow up scenario ever:

Syracuse goes 13-5 shares the regular season title, loses in conference tourney semi. It slips out of 1 seed discussion.
Duke goes 12-6 in the regular season and loses the conference tourney final.

Syracuse is projected as a 2/3, and ends up as 3.
Duke is a borderline 3/4, and ends up as 3.

Duke gets MSG

I just don't see us losing 6 games the rest of the way.

But the Duke scenario, regardless of 1, 2, or 3 seed is the only scenario where I currently see a potential threat to the MSG placement. But Duke is going to have beat someone more substantial than NC State to make me believe that is a real probability. Right now, I see us at worse splitting with Duke and I see any other of our possible ACC losses as being against better competition than Duke has already lost to (ND and Clemson). The fact they beat UVa while we might lose to them won't help Duke's cause. The only thing I think that might help Duke is them beating us twice.

Cheers,
Neil

 

[Madbiker]

You have figured out the secret to the RPI game. Avoid sub 200 games.

Actually you need to avoid the really bad cupcakes with 300+ RPI. Those teams usually play several other 300+ RPI teams and they all make their way into your RPI calculation (via opponents of your opponent). We probably should not play BOTH Colgate and Cornell in the same season or if we want both make sure not to include any other likely 300+ teams like High Point.

 

[Madbiker]

I just don't see us losing 6 games the rest of the way.

But the Duke scenario, regardless of 1, 2, or 3 seed is the only scenario where I currently see a potential threat to the MSG placement. But Duke is going to have beat someone more substantial than NC State to make me believe that is a real probability. Right now, I see us at worse splitting with Duke and I see any other of our possible ACC losses as being against better competition than Duke has already lost to (ND and Clemson). The fact they beat UVa while we might lose to them won't help Duke's cause. The only thing I think that might help Duke is them beating us twice.

Cheers,
Neil

If Duke were to beat us twice, avoid any additional bad loses, and finish ahead of us in the ACC, they would very likely get MSG over us. Never under estimate the the likelihood of the Committee to award Duke a very favorable seeding/placement in the tourney. It's critical that we beat Duke in the Dome.

 

[omniorange]

If Duke were to beat us twice, avoid any additional bad loses, and finish ahead of us in the ACC, they would very likely get MSG over us. Never under estimate the the likelihood of the Committee to award Duke a very favorable seeding/placement in the tourney. It's critical that we beat Duke in the Dome.

I believe that is what I said: The only thing I think that might help Duke is them beating us twice.

Cheers,
Neil

 

[Quazzum69]

There has to be a way in those rating services to preclude a team from jumping a team that has beaten it,. or at least have a "factor" in there to make it honest if it does happen.
Also , if whe "whupped" 'Nova and now they are pounding others, that should somehow make it to our column.

If team A (who is ranked higher than B) can beat everyone but team B (say Syracuse can beat everyone but Pitt), but team B can only beat teams ranked below it, should team B considered better than team A just because they can/have beat them but no one else?

In theory, I suppose, you could have a team that hasn't won a game be rated first in Pomeroy if in each game they had ridiculous points scored per possession and points allowed per possession but their opponents somehow managed to do better in each of those games. Highly unlikely that each opponent would play its best game of the year against you and you would lose each game but it's still possible.

 

[moqui]

Nova hopped Cuse overnight in KenPom's rankings.

yup . . . although we played a great game defensively, holding Pitt well below its season average in PPP, our defensive PPP in the game was also quite a bit higher than our own season average. That hurt us a bit in Pomeroy, just enough that Nova was able to overtake us by 5 ten thousandths of a point.

 

[Cuse315Pat]

Yeah playing teams such as Colgate (7-10), and Fordham (7-10), and High Point (6-10) all dumbed down our SOS and such, but Cornell (1-14) and Binghamton (4-13) really push us out of the schedule talk much more than you'd expect two games to.

Main thing is, we are 8-0 against current projected tournament teams, which currently gives us the best overall resume because Kansas is the only team who has done "as well" against their projected tournament oppponents. Although they have lost a few, they have played them on the road as well. Michigan St has also won like 6 or 7 but lost a couple as well, and Wiscy.

 

[jncuse]

The point I want to make is that when you say "at the end of the day..." , one of the factors why we have only played 1 Top 25 team is because all our opponents have one thing in common, we have beaten them. Is that fact in there somewhere in the data or when you get to that deep level is it then lost?

Your point is incorrect. A loss to Syracuse will not have much negative impact (if any) on RPI or KP of a team like Baylor. There are many games, but the way the formulas work as well will not impact Baylor.

But let's go further with the impact on Baylor of losing to Syracuse. I am not a fan of RPI but its obviously relevant because it is the basis of top 25 and top 50 teams for the committee.

RPI

RPI is 25% your winning%, 50% your opponent winning%, and 25% their opponents winning%. I guess your point could only be applied to the last 25% ...but all top teams would have same impact on cumulative record of their opponents . Let's say we won 2 more games than another top contender. It would be 2 losses out of a sample of about 1200. So no impact.

By the formula itself, merely playing us, will have a positive impact on your RPI. Its huge for the 50% factor, and our SOS is still projected at #50 in the nation so the last 25% is strong as well.

Losing to us, is better than beating a Cupcake. If you beat a cupcake the first 25% is great, but then your getting much less for the last 75% of the formula.

The only thing better than losing to Syracuse would be beating a team like Dayton. It would get less of that last 75%, but not so much less to offset the first 25% you now get. (Of course you better beat Dayton!)

But at the end of the day, losing to us really has no negative impact on your RPI. We are not taking teams out of the top 25 RPI.

KP

KP is based on efficiency, but you can convert the rankings to points. We are currently 8 points better than Baylor on a neutral court. We beat them by 7 on a neutral court. So in fact that game against Syracuse slightly improved their KP by about 1 point over their course of the season -- from .8292 to.8298, but still 44th either way.