What's new
By:
USCHO Fan Forum

This is a sample guest message. Register a free account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!

  • The USCHO Fan Forum has migrated to a new plaform, xenForo. Most of the function of the forum should work in familiar ways. Please note that you can switch between light and dark modes by clicking on the gear icon in the upper right of the main menu bar. We are hoping that this new platform will prove to be faster and more reliable. Please feel free to explore its features.

2012 NCAA Tournament: Bracketology

Status
Not open for further replies.
Re: 2012 NCAA Tournament: Bracketology

Off-season/next season I may link a google doc to program out odds and other things... treat it as "sorta open source" in the sense that the code will be there and maybe other stuff so that a good baseballprospectus type odds calculation can be done. Just thinking out loud.
 
Re: 2012 NCAA Tournament: Bracketology

Alaska88 said:
I think NMU is out. To get in they needed Western Michigan to lose both games.
Click to expand...

Not so. Try this: It gets 6 CCHA teams in yet.

WCHA Semifinal #1: Minnesota defeats North Dakota.
WCHA Championship game: Minnesota defeats Denver.
Hockey East Semifinal #2: Maine defeats Boston University.
Hockey East Championship game: Maine defeats Boston College.
ECAC Semifinal #2: Harvard defeats Cornell.
ECAC Championship game: Union defeats Harvard.
ECAC Consolation game: Colgate defeats Cornell.
CCHA Semifinal #1: Michigan defeats Bowling Green.
CCHA Championship game: Michigan defeats Western Michigan.
CCHA Consolation game: Miami defeats Bowling Green.
Atlantic Hockey Semifinal #2: RIT defeats Niagara.
Atlantic Hockey Championship game: Air Force defeats RIT.
 
Re: 2012 NCAA Tournament: Bracketology

I'm more of a math guy than a computer guy... but couldn't someone with the skills make a pretty easy program that runs every possibly outcome and tells you what percent of those outcomes have Team X in the field? There's what, 4.7 million or so possible outcomes at the beginning? You can't tell me one of you RPI geeks doesn't have a computer with decent enough power to set up a program and run it for like 5 minutes and have it spit out the % of times each team is in out of 4.7 million.

Or am I totally off base?
 
TonyTheTiger20 said:
I'm more of a math guy than a computer guy... but couldn't someone with the skills make a pretty easy program that runs every possibly outcome and tells you what percent of those outcomes have Team X in the field? There's what, 4.7 million or so possible outcomes at the beginning? You can't tell me one of you RPI geeks doesn't have a computer with decent enough power to set up a program and run it for like 5 minutes and have it spit out the % of times each team is in out of 4.7 million.

Or am I totally off base?
Click to expand...

You aren't off base but usually the computer guys arent math guys and vice versa. I'm only good enough to pull off what I need... Efficienctly... Oh, hell when I did it last it took 4-5 seconds... 1000 times over is an hour and a half.
 
Re: 2012 NCAA Tournament: Bracketology

Patman said:
You aren't off base but usually the computer guys arent math guys and vice versa. I'm only good enough to pull off what I need... Efficienctly... Oh, hell when I did it last it took 4-5 seconds... 1000 times over is an hour and a half.
Click to expand...
True, but you're pulling it online with high traffic. It's simple code, an RPI geek can just pull that shizz onto his own computer and run it offline. I have to figure even a halfway decent computer can run that multiple times a second.
 
Re: 2012 NCAA Tournament: Bracketology

If there were a hockey tournament but nobody attended and nobody could watch on TV, can we just pretend that it never actually happened? Please?
 
TonyTheTiger20 said:
True, but you're pulling it online with high traffic. It's simple code, an RPI geek can just pull that shizz onto his own computer and run it offline. I have to figure even a halfway decent computer can run that multiple times a second.
Click to expand...

I repeat...

The RPI is a little ****y... And COP and TUC is annoying too... It all comes down to the coder.

If a few people want to get serious about this we can talk in the off-season

Dr. Joyce
 
Re: 2012 NCAA Tournament: Bracketology

20 games = 1 million possibilities, even not accounting for ties in consolation games. If you could evaluate 10 per second (pretty optimistic), that's still 100,000 seconds, or 27 hours. It is, however, a problem which lends itself to parallel computing since the results of one case don't depend on the others, so if you had 27 computers, you could do it in an hour. There are pretty clearly at least 27 of us who could participate, and most of us probably have multiple computers - sounds like we need a distributed computing solution!

Even so, not exactly the kind of thing you can update in real time as the results come in. At best, you could do it overnight each night before the games the next day.
 
Re: 2012 NCAA Tournament: Bracketology

In a few situations I have Northern in a tie for 15th with Merrimack and Northern has a higher RPI, also all the higher seeds win (AHA factored in). What's the tie-breaker?
 
Re: 2012 NCAA Tournament: Bracketology

LynahFan said:
20 games = 1 million possibilities, even not accounting for ties in consolation games. If you could evaluate 10 per second (pretty optimistic), that's still 100,000 seconds, or 27 hours. It is, however, a problem which lends itself to parallel computing since the results of one case don't depend on the others, so if you had 27 computers, you could do it in an hour. There are pretty clearly at least 27 of us who could participate, and most of us probably have multiple computers - sounds like we need a distributed computing solution!

Even so, not exactly the kind of thing you can update in real time as the results come in. At best, you could do it overnight each night before the games the next day.
Click to expand...
This is what you'd do, and keep in mind, the number of possibilities drops exponentially after each night. You'd just need 27 people the first night (or however many), honestly.
 
LynahFan said:
20 games = 1 million possibilities, even not accounting for ties in consolation games. If you could evaluate 10 per second (pretty optimistic), that's still 100,000 seconds, or 27 hours. It is, however, a problem which lends itself to parallel computing since the results of one case don't depend on the others, so if you had 27 computers, you could do it in an hour. There are pretty clearly at least 27 of us who could participate, and most of us probably have multiple computers - sounds like we need a distributed computing solution!

Even so, not exactly the kind of thing you can update in real time as the results come in. At best, you could do it overnight each night before the games the next day.
Click to expand...

Which is why I believe in models and monte carlo for stuff like this.

Edit: the funny thing is most home machines have multiple cores and this isn't a memory intensive job... Feed the problem to the other cores and you get 2-4 fold improvement.
 
Last edited:
Re: 2012 NCAA Tournament: Bracketology

TonyTheTiger20 said:
This is what you'd do, and keep in mind, the number of possibilities drops exponentially after each night. You'd just need 27 people the first night (or however many), honestly.
Click to expand...
Right, which is all the more reason no (sane) person would ever put in the effort to program it. Just going into SUNDAY'S games, there were 29 games left that had to have a winner plus 2 that can end in ties. That's 4.8 billion possibilities, which, at 10 per second is 15 YEARS of computing time. I think we can agree that there's no point in thinking about going any further back?

After tonight's games, there are 5 games with a winner and 2 with possible ties, for a whopping 288 possibilities left. A toddler with a calculator could work all those out.

So this awesome distributed computing solution would be useful exactly one time per season. Yawn. Totally not worth it - nobody will ever do it.

Edit: yes, Patman - Monte Carlo is definitely the way to go. I'm working on the only aircraft system (fuel tank inerting) which is FAA certified by Monte Carlo methods, in fact.
 
Last edited:
Re: 2012 NCAA Tournament: Bracketology

Merrimack is suddenly looking a lot more possible.

They need 3 things to happen now, rather than like 11 earlier. They need Cornell to lose to Colgate, and they need Havard and Bowling Green to not win their tournaments.

Still slim chance, but less slim than earlier.
 
Re: 2012 NCAA Tournament: Bracketology

Eaglefan06 said:
Merrimack is suddenly looking a lot more possible.

They need 3 things to happen now, rather than like 11 earlier. They need Cornell to lose to Colgate, and they need Havard and Bowling Green to not win their tournaments.

Still slim chance, but less slim than earlier.
Click to expand...

Well, they got BG.
 
Status
Not open for further replies.
Back
Top