B1G Tournament Forecasting Thread

Status
Not open for further replies.
#26      

JFGsCoffeeMug

BU:1 Trash cans:0
Chicago
Every time we lose at home, we're giving away one of our hard-fought road wins. We need this one.
 
#27      

danielb927

Orange Krush Class of 2013
Rochester, MN
73.4% chance at a Top 4 Seed per Torvik with the next being Rutgers at 42%. If we win vs MSU I'd expect that gap to grow. If we lose this game vs MSU I'd expect the Rutgers / Illinois percent to be approximately equal. Need to get a W against MSU and make progress on securing at Top 4 finish.

I'm just noticing that his top-4 forecast includes ties, so that 73% chance is reflective of chances of a top-4 finish, not a top-4 seed. A decent fraction of that 73% is situations where we tie for 4th (or 3rd) and are seeded 5th or worse. As evidence, if you check out the sum of the top-4 column, you get something like 500%. That means on average, 5 teams are going to end up tied for the top 4 spots.
 
#28      
I'm just noticing that his top-4 forecast includes ties, so that 73% chance is reflective of chances of a top-4 finish, not a top-4 seed. A decent fraction of that 73% is situations where we tie for 4th (or 3rd) and are seeded 5th or worse. As evidence, if you check out the sum of the top-4 column, you get something like 500%. That means on average, 5 teams are going to end up tied for the top 4 spots.
Important distinction.

Thanks
 
#29      

danielb927

Orange Krush Class of 2013
Rochester, MN
The latest from model runs overnight:

Perceived Clarity — 51% (up from 49%)
True Clarity — 43% (up from 39%)

1581386213172.png


I'm going to stick with Mean Seed as the sort order this season just for consistency's sake. Illinois/MSU is really showing how odd that can be. We have a lower shot at the #1 seed, but quite a bit higher chance for a 2–4 seed and end up with a higher overall shot at a top-4 slot (60% to 55%). Illinois still maintains our sub-1% chance at a no-bye catastrophe despite Friday's loss.

The other big change is Indiana, who has played themselves into a more-likely-than-not 12th seed after their home loss. Elsewhere in the B1G, this is an incredible race where four teams (Iowa, Wisconsin, Purdue, and Rutgers) have a 1-3% chance of both a 1 seed or a 12 seed. I can't imagine we've seen that too often with 8 conference games left too often.

1581386650948.png


If we look at the equal-strength scenario, it's apparent how much of IU's fall is due to the drop in their ratings. Based purely on record and the alignment of remaining games, they'd be expected to finish roughly on par with OSU and Michigan.

I've changed the rightmost column here to highlight the difference in mean seeding versus the normal case. Rutgers finds themselves in much the same boat as IU — they're expected to finish almost 2 whole seeds worse than what we see with all teams equal. In their case it's not so much how good they are as how hard their last 7 games are: @OSU, IL, Mich, @Wisc, @PSU, MD, @Purdue.

Lastly, I'll just note without the whole chart that running the "regular" scenario with random tiebreakers nets us a 10% chance at the 1 seed instead of 6%. It's hard (but not impossible) to imagine us winning any tiebreaker involving Maryland, so let's say 3/4 of that bump is due to our two losses to them. That tells us something interesting:
  • 6% chance if we always lose the tiebreaker to Maryland
  • 10% chance if we have a 50/50 chance in all tiebreakers
  • at least 3/4 of that change attributable to tiebreakers — i.e., a 3% bump if our tiebreaker chance goes from none to half
  • = something like 6% (roughly 1 in 15) scenarios involve Illinois and Maryland tied for the Big Ten crown.
 
#30      
The latest from model runs overnight:

Perceived Clarity — 51% (up from 49%)
True Clarity — 43% (up from 39%)

View attachment 5676

I'm going to stick with Mean Seed as the sort order this season just for consistency's sake. Illinois/MSU is really showing how odd that can be. We have a lower shot at the #1 seed, but quite a bit higher chance for a 2–4 seed and end up with a higher overall shot at a top-4 slot (60% to 55%). Illinois still maintains our sub-1% chance at a no-bye catastrophe despite Friday's loss.

The other big change is Indiana, who has played themselves into a more-likely-than-not 12th seed after their home loss. Elsewhere in the B1G, this is an incredible race where four teams (Iowa, Wisconsin, Purdue, and Rutgers) have a 1-3% chance of both a 1 seed or a 12 seed. I can't imagine we've seen that too often with 8 conference games left too often.

View attachment 5677

If we look at the equal-strength scenario, it's apparent how much of IU's fall is due to the drop in their ratings. Based purely on record and the alignment of remaining games, they'd be expected to finish roughly on par with OSU and Michigan.

I've changed the rightmost column here to highlight the difference in mean seeding versus the normal case. Rutgers finds themselves in much the same boat as IU — they're expected to finish almost 2 whole seeds worse than what we see with all teams equal. In their case it's not so much how good they are as how hard their last 7 games are: @OSU, IL, Mich, @Wisc, @PSU, MD, @Purdue.

Lastly, I'll just note without the whole chart that running the "regular" scenario with random tiebreakers nets us a 10% chance at the 1 seed instead of 6%. It's hard (but not impossible) to imagine us winning any tiebreaker involving Maryland, so let's say 3/4 of that bump is due to our two losses to them. That tells us something interesting:
  • 6% chance if we always lose the tiebreaker to Maryland
  • 10% chance if we have a 50/50 chance in all tiebreakers
  • at least 3/4 of that change attributable to tiebreakers — i.e., a 3% bump if our tiebreaker chance goes from none to half
  • = something like 6% (roughly 1 in 15) scenarios involve Illinois and Maryland tied for the Big Ten crown.

what does your model have against Rutgers? Others have them selected to finish better
 
#31      

danielb927

Orange Krush Class of 2013
Rochester, MN
what does your model have against Rutgers? Others have them selected to finish better

For now, all I've been able to incorporate is KenPom projections. So, I'd guess his system thinks a bit less of Rutgers than others' (or maybe just thinks more of other mid-pack B1G teams, like Iowa). I should probably hesitate before calling my setup a "model" per se — basically just simulating the rest of the season many times under a few different conditions.
 
Last edited:
#32      
For now, all I've been able to incorporate is KenPom projections. So, I'd guess his system thinks a bit less of Rutgers than others' (or maybe just thinks more of other mid-pack B1G teams, like Iowa). I should probably hesitate before calling my setup a "model" per se — basically just simulating the rest of the season many times under a few different conditions.

With 2 losses to Maryland and 1 loss (hopefully only 1) to MSU our tiebreaker scenarios for the title are not good at all. We would need to win outright and that would probably require a 14-6 record. Which means going 6-2 the rest of the way. Not impossible but a real challenge.
 
#33      
With 2 losses to Maryland and 1 loss (hopefully only 1) to MSU our tiebreaker scenarios for the title are not good at all. We would need to win outright and that would probably require a 14-6 record. Which means going 6-2 the rest of the way. Not impossible but a real challenge.

5-3 will be.
 
#34      
With 2 losses to Maryland and 1 loss (hopefully only 1) to MSU our tiebreaker scenarios for the title are not good at all. We would need to win outright and that would probably require a 14-6 record. Which means going 6-2 the rest of the way. Not impossible but a real challenge.
Isn't it shared title if teams are tied at the top? I know there are tiebreaker scenarios needed for BTT seeding, but the conference title would still be shared.
 
#40      
what does your model have against Rutgers? Others have them selected to finish better
It's probably about the remaining schedule. Per Pomeroy, here are the in-conference SoS rankings to date for the top half of the league:

2. Wisconsin
3. Illinois
5. Penn St.
7. Maryland
11. Purdue
12. Michigan St.
13. Iowa
14. Rutgers

The unbalanced schedule makes it an imperfect measure, but since we're a little more than halfway through the season the inverse of those rankings are a decent proxy for how difficult the rest of everyone's schedule is. Rutgers last seven games are:

at OSU
vs Illinois
vs Michigan
at Wisconsin
at Penn St.
vs. Maryland
at Purdue

No games against the bottom two. Same for Iowa.
 
#42      

danielb927

Orange Krush Class of 2013
Rochester, MN
Updated after last night's games. Illinois' mean BTT seed is now 5.2, but that's still 4th best in the conference. 42% chance of a double-bye at this point (obviously this does not take into account any time Ayo misses).

1581612283619.png


Also surprising to see Penn State ahead of Maryland — two reasons for that. First, PSU won in their lone meeting way back on December 10. Second, PSU has a much easier schedule from here on out, with 4 home games and 2 games against Northwestern. That adds up to a major difference versus the case where every game from here out is a toss-up:

1581612448796.png


True clarity is up to 46% and perceived clarity up to 54%. Both of those numbers still feel about right, IMO.
 
#43      
So, we are either a 4-seed or a 5-seed and either way we play Iowa on Friday unless the 5-seed loses to Indiana on Thursday?
 
#44      

danielb927

Orange Krush Class of 2013
Rochester, MN
So, we are either a 4-seed or a 5-seed and either way we play Iowa on Friday unless the 5-seed loses to Indiana on Thursday?

With about 40% certainty, definitely maybe probably (y)
 
#46      
A lot has changed in the last six days. We are in a five way tie for third, and of that group, only Wisconsin seems to have a remaining schedule as easy as ours. If we go 4-1 from here, it seems almost certain we will wind up with a two-round bye in the BTT. If Penn State goes 3-2 while Wisconsin and Illinois go 4-1, we could have a three-way tie for second. If so, we win the tiebreak and get the second seed. Our final game against Iowa could be absolutely huge.
 
#50      

haasi

New York
I’m going to guess we have about a 50% shot of a top 4 seed and a 98% chance of top 8.

1 seed is all but out of reach - need to gain 4 games on Maryland with 5 left.

2 should be in play if we win out, but probably not otherwise.

My guess is our most likely outcomes will be focused in the 3-7 range, where we’re currently tied with 4 other teams, including MSU which has a tie break on us. If there’s a log jam at say 12-8, tie breaks will come into play. I think we’re reasonably likely to finish in top 4 because of our relatively easy remaining schedule, but it could totally come down to our finale against Iowa at home. I feel pretty good that we’ll be in top 4 if we finish 4-1, hard to see more than 4 teams finishing 13-7 and beating us on tie breaks.

in terms of 8 and 9, were 1.5 games ahead of Michigan and OSU. 8 could be in play, but 9 seems very unlikely because it almost certainly would mean Michigan has to pass us, and we have the relatively easy schedule and a tie break on them.

10-12 should be extremely unlikely with us 2.5 games ahead of Minny, Indiana, and Purdue.

13/14 is already impossible.
 
Status
Not open for further replies.