Pairwise and Bracketology 2013 Edition

[goblue78]

Re: Pairwise and Bracketology 2013 Edition

I've got PWR code written in Stata if anyone is interested. It's not yet ready to do extensive Monte Carlo simulations yet, because it's w-a-a-a-y too slow on the CommOpp criterion. Everything else is working pretty well, much faster than I'd have expected.

[RHamilton]

Re: Pairwise and Bracketology 2013 Edition

I found an odd thing in JimDahl's blog tonight. The assumption here has been that St Cloud can now clinch with a win.

However, on his blog, updated with today's results, we find this for St Cloud, under the 1 win column:

#13 0.0%

Which in interesting to me, because the #14 row is totally empty. And, if that really means 0.01%, where there might be a strange case yet, then a win tomorrow afternoon does NOT guarantee them in, because we all know that CC, Mich, Prov and Brown could yet win tourneys, in which case #13 in the PWR misses.

Can, RHamilton or PatMan add more information?

I see 10 outcomes (out of 49,152) where St. Cloud ends up #17 when winning their semi.

The scenarios all involve:
- Niagara losing to Mercyhurst or Connecticut
- Miami losing to Ohio State or Notre Dame
- Union DEFEATING Brown in the ECAC Final
- Quinnipiac and Yale TYING in the ECAC Consolation
- 8 of 10 feature BU winning hockey east, the other two have Providence beating BU -- the Providence wins require Connecticut to beat Niagara
- CC winning the WCHA

EDIT: Putting this in the Pairwise Predictor, I see that St. Cloud ends up in a 4-way tie for #9, becomes #12 on tiebreakers. Union finishes #13, and BU finishes #14, but they both move up one seed to make room for Ohio State, Colorado College, and Mercyhurst. This bumps SCSU down to #17.

Ugly is all I can say!

[RHamilton]

Re: Pairwise and Bracketology 2013 Edition

To help all the pairwise prognosticators out there, I put together this:

http://pwr.reillyhamilton.com/query.php
Simply input your choice, or fill in games as the day goes on, and receive the weighted (and unweighted) outcomes.

It may take up to a minute or so for some queries. The more teams you switch from "Don't Care" to something else, the quick it will run. Impossible conditions (a team winning a final than won neither semi) will not return any results. Results are cached, though, so repeated visits to the same query should be instantaneous. Use the links to share.

Let me know if you have any issues running it. We'll see how long it stays up, and if my webhost complains!

[Umileated]

Re: Pairwise and Bracketology 2013 Edition

Maybe I'll prepare a readme later.

Yes, please!

[sshablak]

Re: Pairwise and Bracketology 2013 Edition

Much easier....you're good. Tx

[RHamilton]

Re: Pairwise and Bracketology 2013 Edition

Okay, I've got one more tool written up. This one is far less polished, and a bit harder to exmplain, but it can be used to help find what individual game outcomes lead to a particular result.
http://pwr.reillyhamilton.com/query2.php

Take the St. Cloud example earlier. Inputting St. Cloud winning their semi and a seed of > 16 gives us the following:
http://pwr.reillyhamilton.com/pub_re...&opp=1&seed=16
Going game by game, the page lists what must happen for that outcome to happen. Anything 100% must happen for this scenario to occur, and any game that is "missing" a result (example: Niagara winning the AHA final) can NOT happen if the scenario listed will occur.

It also somewhat provides teams to root for. If you submit the form as is, you'll get the list of game outcomes when RPI's seed is <= 16, or, when RPI is in the tournament. There's not a single missing game or 100% game beyond the two completed WCHA semis, so no one result can knock the Engineers out of the tournament. The number of outcomes for each result illustrates how many outcomes would remain if that result happened. Since larger numbers are generally better, one should root for the outcome with the larger number of outcomes.

Another example: Robert Morris in the tournament (seed <= 16):
http://pwr.reillyhamilton.com/pub_re...&opp=5&seed=16

Since many of these scenarios are dependent on each other, you can/should return to the query2 page to fine tune game outcomes to get a better picture of what exactly must happen.

Let me know if you have any question about this tool; I hope someone finds it useful!

[UND92UMNLaw99]

Re: Pairwise and Bracketology 2013 Edition

Okay, I've got one more tool written up. This one is far less polished, and a bit harder to exmplain, but it can be used to help find what individual game outcomes lead to a particular result.
http://pwr.reillyhamilton.com/query2.php

Take the St. Cloud example earlier. Inputting St. Cloud winning their semi and a seed of > 16 gives us the following:
http://pwr.reillyhamilton.com/pub_re...&opp=1&seed=16
Going game by game, the page lists what must happen for that outcome to happen. Anything 100% must happen for this scenario to occur, and any game that is "missing" a result (example: Niagara winning the AHA final) can NOT happen if the scenario listed will occur.

It also somewhat provides teams to root for. If you submit the form as is, you'll get the list of game outcomes when RPI's seed is <= 16, or, when RPI is in the tournament. There's not a single missing game or 100% game beyond the two completed WCHA semis, so no one result can knock the Engineers out of the tournament. The number of outcomes for each result illustrates how many outcomes would remain if that result happened. Since larger numbers are generally better, one should root for the outcome with the larger number of outcomes.

Another example: Robert Morris in the tournament (seed <= 16):
http://pwr.reillyhamilton.com/pub_re...&opp=5&seed=16

Since many of these scenarios are dependent on each other, you can/should return to the query2 page to fine to game outcomes to get a better picture of what exactly must happen.

Let me know if you have any question about this tool; I hope someone finds it useful!

Your newest creation is fantastic. Thanks!

I see that Quinnipiac and Minnesota are locked into being #1 and #2 respectively. I wonder if two top teams have ever had so little to play for in their conference tourneys?

[Numbers]

Re: Pairwise and Bracketology 2013 Edition

I thought I would take a little look at what the bracket might be on Sunday night.

For this I am assuming "more or less normal" results, meaning not a lot of upsets.

A summary: QU = 1; Minny = 2; Miami =~3, but almost for sure 4 otherwise; and the most likely thing is BC/Lowell to #4.

So, let's set the #1s:
Providence = Quinnipiac (unless Brown wins, but that is a long shot yet)
Rapids = Minny
Toledo = Miami
Manchester = HE Champ
There is a very very good chance of this happening!!

The #2s are a problem right now, because we don't know yet whether UNH ends up a 2 or a 3. Likewise NoDak can be 7,8,9 with 'normal' results. I did not run lots of cases, but I believe that if NoDak falls to a #3 under normal results, it is because Mankato ends up as a #2. Again, no guarantees here, but this is a "most likely" scenario.

It seems that, whichever WCHA team ends up as a #2, that most of the 3 band is going to be WCHA teams (St Cloud, Denver +). This changes if SCSU wins the Broadmoor (please do not be upset SCSU fans, SCSU can get to a #2 win the Broadmoor, but then we still end up with 3 WCHA teams in the 3 band).

So, there are basically 2 scenarios: 1 with UNH as a 2, the other as a 3.

UNH = #2 seed
Providence: QU (1); #7 seed (most likely WCHA #2)
Rapids: Minny (2); #6 seed (could be HE#2, or UND/SCSU)
Toledo: Miami (3 or 4); #5 seed (could be HE #2, Yale or Niagara)
Manchester: HE Champ (4 or 3); UNH (host)
And, we could guess the rest, because the 3 WCHA teams in the 3 bands have to matchup with UNH, and the non-WCHA teams in the 2 bands. This is not so interesting yet. Committee cold move WCHA2 to Rapids for attendance, etc......

In the case where UNH is a 3 seed (9 or 10)
Then, we get the usual serpentine:
Prov: QU (1), #8 (0ften, WCHA #2)
Rapids: Minny (2), #7 (also, commonly WCHA #2, but not guaranteed)
Toledo : Miami and either 5/6
Manchester: HE Champ and 6/5
Then, UNH goes to Manchester, and the WCHA #2 is moved over to join them to avoid an all WCHA matchup in round one.

So, the top of the bracket then is:
Prov: QU; lowest ranked non-WCHA team in the 2-band;highest non-UNH team in 3-band (WCHA team); 4seed=??
Rapids: Minny; next-lowest non-WCHA team in the 2-band; next-highest non non-UNH team (WCHA team); 4seed=??
Toledo: Miami; other non-WCHA team in 2-band; other non-UNH team in 3-band; 4seed=??
Manchester: HE Champ; WCHA team from 2-band; UNH; 4seed=??

Now, admittedly, this is a preliminary only. But, this is the strong pattern: 1 WCHA team in 2-band, 3 WCHA teams in 3-band, and then if UNH is a 3-seed, the WCHA 2-seed HAS to go east.

Watch for that. I will keep this up. I think I can be back after the AHA/WCHA/ECAC early games with more information about how this looks.


EDIT: I tried Reilly's tool for UNH being a 2 seed. It gave me ~60,000 combos, which I can't amake sense of, given that Reilly said last night that he ran 49,000 combos to find where SCSU wins today and still misses. So, I am confused by the numbers. I find, looking at the results of my query, that when Niagara wins AHA, UNH has far fewer opportunities to be a #2. This of course, makes sense, because Niagara would move ahead of UNH in the PWR.
So, as a first approximation, if Niagara wins, it gets 'quite likely' (whatever that really means??) that UNH will be a 3, and the WCHA situation discussed above will come into play.

[MaizeRage]

A really basic question that I've just realized I never knew the answer to and never inquired about:

How long have the selection committee used the PWR system, and have they ever veered from it for at-large selection or seeding? I've only been paying attention to this kind of detail for the last decade or so, and I can't think of an instance where they've veered from the formula.

No, they've never veered from it on the men's side. The PWR is actually codified in the rulebook now, so their hands are tied in instances where the PWR goes against all logic.

They have veered on the women's side because the flaws of the PWR system are a little more glaring.

[RHamilton]

Re: Pairwise and Bracketology 2013 Edition

EDIT: I tried Reilly's tool for UNH being a 2 seed. It gave me ~60,000 combos, which I can't amake sense of, given that Reilly said last night that he ran 49,000 combos to find where SCSU wins today and still misses. So, I am confused by the numbers. I find, looking at the results of my query, that when Niagara wins AHA, UNH has far fewer opportunities to be a #2. This of course, makes sense, because Niagara would move ahead of UNH in the PWR.
So, as a first approximation, if Niagara wins, it gets 'quite likely' (whatever that really means??) that UNH will be a 3, and the WCHA situation discussed above will come into play.

The SCSU result was run with SCSU designated as winning their semi, resulting in ~50,000 total combos. Running on a blank slate (other than the WCHA quarters) will result in ~100,000 combos.

Is that what you meant?

[Numbers]

Re: Pairwise and Bracketology 2013 Edition

That helps Reilly. So, right now about 1000,000 total combos. 60,000 yield UNH to be a 2 seed, but the KRACH odds are against that because of the Niagara results.

[Priceless]

you know what is the most annoying part of this? we've got it boiled down to a science with actual computer scientists and statisticians who can tell ahead of time what a given result will yield and we'll still have to listen to supposed experts on TV tell us total nonsense.

[RHamilton]

Re: Pairwise and Bracketology 2013 Edition

That helps Reilly. So, right now about 1000,000 total combos. 60,000 yield UNH to be a 2 seed, but the KRACH odds are against that because of the Niagara results.

Right; there's a fairly major difference the weighted vs. unweighted results for UNH's chance at a 9 seed. The 40,000 outcomes that result in a 3 seed are more likely aggregately than the 60,000 that result in a 2 seed, which is part of the reason I chose to KRACH weight.

[RHamilton]

Re: Pairwise and Bracketology 2013 Edition

you know what is the most annoying part of this? we've got it boiled down to a science with actual computer scientists and statisticians who can tell ahead of time what a given result will yield and we'll still have to listen to supposed experts on TV tell us total nonsense.

Sometimes I forget that the games even have to be played at all!

[Patman]

you know what is the most annoying part of this? we've got it boiled down to a science with actual computer scientists and statisticians who can tell ahead of time what a given result will yield and we'll still have to listen to supposed experts on TV tell us total nonsense.

Its almost like they can't read.

The other thing of course is that they guys make their money pretending to be experts, once they figure out that they aren't then they're going to figure which of us is the most telegenic and put us on TV.

Of course, a lot of them are flat out allergic to the geekery... Even the NFL tiebreakers the TV guys go "aw shucks" and move on.