View Full Version : NCAA Hockey Tournament Odds

Catamount93

03-24-2014, 02:45 PM

Has anyone seen any odds posted for this year's tournament. I think in the past I've seen odds on going all the way but not on individual games.

Great to see the Catamounts back in the dance!

goblue78

03-24-2014, 07:32 PM

First round KRACH odds:

BC 0.769

Lowell 0.586

Wisconsin 0.516

Ferris 0.557

Union 0.695

Quinnipiac 0.533

Minnesota 0.933

St Cloud 0.506

sshablak

03-25-2014, 06:38 AM

First round KRACH odds:

BC 0.769

Lowell 0.586

Wisconsin 0.516

Ferris 0.557

Union 0.695

Quinnipiac 0.533

Minnesota 0.933

St Cloud 0.506

The other teams don't count ??

mookie1995

03-25-2014, 06:41 AM

The other teams don't count ??

Count as much as these don't :p

jcarter7669

03-25-2014, 07:04 AM

First round KRACH odds:

BC 0.769

Lowell 0.586

Wisconsin 0.516

Ferris 0.557

Union 0.695

Quinnipiac 0.533

Minnesota 0.933

St Cloud 0.506

These odds were created while someone was on crack I think... Union has better odds of beating UVM then Lowell does of taking out Mankato... that doesn't even pass the crack cocaine sniff test.

The other teams don't count ??Those appear to be probabilities, not odds. The probabilities in each game add up to 1. If the probability that BC will win is 0.769, the probability that Denver will win is .231.

These odds were created while someone was on crack I think... Union has better odds of beating UVM then Lowell does of taking out Mankato... that doesn't even pass the crack cocaine sniff test.That the probability that a one seed will be a four seed is greater than the probability that a two seed will beat a three seed is exactly what you'd expect.

Catamount93

03-25-2014, 09:13 AM

I think both of you are correct above.

Per KRACH, the odds are exactly as you'd expect.

But they are far from the odds a Vegas oddsmaker would develop for gambling purposes, which is why they don't pass the "sniff test".

I was trying to find Vegas gambling odds, as in the past they were posted at one or two casinos, but I have not found them this year yet.

Felger

03-25-2014, 09:32 AM

I was trying to find Vegas gambling odds, as in the past they were posted at one or two casinos, but I have not found them this year yet.

My rule of thumb is that if I'm trying to find odds on a collegiate sporting event, Chestnut Hill is the logical place to start

goblue78

03-25-2014, 02:22 PM

But they are far from the odds a Vegas oddsmaker would develop for gambling purposes, which is why they don't pass the "sniff test".

I'm not sure exactly what you mean by "far from," beyond the fact that these are indeed probabilities, not odds. Las Vegas bookmakers simply try to get equal action on both sides and collect the vig. They do this by putting out preliminary lines (which tend to be quite similar to KRACH odds, btw) and then open very limited betting to see which way to move the lines. So the percentages here aren't exactly lines, because they haven't been adjusted for the 10 percent vig, but otherwise they ought to be fairly close to fair, with the proviso that strong bettor interest on one side or another can move the odds for reasons that have little to do with fundamentals. Roughly, speaking, the odds will simply be the win probabilities expressed as odds ratios (p/(1-p)) and then rounded for vig. For example, BC at .769 win probability becomes 1:3 odds, betting 3 to return 4.

johnk

03-25-2014, 02:37 PM

Ok, if Vegas doesn't want to put out odds, I will:

Minn 3/1

BC 7/2

Union 7/2

Wisc 4/1

UM Lowell 5/1

Ferris St 6/1

Quinnipiac 8/1

SCSU 10/1

ND 10/1

Prov 12/1

Minn St. 12/1

Colgate 15/1

North Dakota 18/1

Vermont 20/1

Denver 25/1

Robert Morris 100/1

5mn_Major

03-25-2014, 08:24 PM

First round KRACH odds:

BC 0.769

Lowell 0.586

Wisconsin 0.516

Ferris 0.557

Union 0.695

Quinnipiac 0.533

Minnesota 0.933

St Cloud 0.506

I gotta laugh at these. No way MN is a 93/7 likelihood to win the game against RM. The Gophers women's team was a 90%+ likelihood to win an NC and I laughed at that one too.

I'd put quite a bit of money on UW though.

Kdaddy1968

03-25-2014, 08:35 PM

Ok, if Vegas doesn't want to put out odds, I will:

Minn 3/1

BC 7/2

Union 7/2

Wisc 4/1

UM Lowell 5/1

Ferris St 6/1

Quinnipiac 8/1

SCSU 10/1

ND 10/1

Prov 12/1

Minn St. 12/1

Colgate 15/1

North Dakota 18/1

Vermont 20/1

Denver 25/1

Robert Morris 100/1

I would move Ferris down and put Colgate higher..(above Prov) and UML above Wisc but looks about right otherwise.

goblue78

03-25-2014, 10:21 PM

Unfortunately, 5mn_Major, there's no way to tell by itself if the 0.933 is right. Even if RM wins, that just means that the 7 percent chance came through. That happens... wait for it... 7 percent of the time! All the KRACh probabilities actually means is that when teams as good as Minnesota play teams as below-average as RM, the good teams wins over 9 out of ten times. What part of that seems unlikely to you?

If you're making the somewhat more sophisticated point that RM must have played better lately to get through the AHA tournament and MN must be playing worse to not even get to the Big10 final, then that's something that can be estimated. There doesn't actually seem to be a lot of momentum in college hockey, although writers love to write about it, but there is some. But if you want to make a bet on RM and are willing to take, say, 5-1, let me know and maybe we can make a deal.

mnstate0fhockey

03-25-2014, 10:37 PM

Unfortunately, 5mn_Major, there's no way to tell by itself if the 0.933 is right. Even if RM wins, that just means that the 7 percent chance came through. That happens... wait for it... 7 percent of the time! All the KRACh probabilities actually means is that when teams as good as Minnesota play teams as below-average as RM, the good teams wins over 9 out of ten times. What part of that seems unlikely to you?

If you're making the somewhat more sophisticated point that RM must have played better lately to get through the AHA tournament and MN must be playing worse to not even get to the Big10 final, then that's something that can be estimated. There doesn't actually seem to be a lot of momentum in college hockey, although writers love to write about it, but there is some. But if you want to make a bet on RM and are willing to take, say, 5-1, let me know and maybe we can make a deal.

Momentum, no. But there is such a thing as playing with confidence and playing without. Playing loose and gripping your stick a bit too tight. Teams that are said to have "momentum" are usually playing with a lot of confidence, don't seem to press, and do the things that win you games (like get shots to the net and get bodies there). Teams playing without "momentum" seem to press, get away from the things that got them there, and play too tight.

I am a firm believer that confidence is one of the most important and most underrated aspects to playing good hockey. It's something that takes time to build, but can be lost in an instant.

Patman

03-26-2014, 12:02 AM

Momentum, no. But there is such a thing as playing with confidence and playing without. Playing loose and gripping your stick a bit too tight. Teams that are said to have "momentum" are usually playing with a lot of confidence, don't seem to press, and do the things that win you games (like get shots to the net and get bodies there). Teams playing without "momentum" seem to press, get away from the things that got them there, and play too tight.

I am a firm believer that confidence is one of the most important and most underrated aspects to playing good hockey. It's something that takes time to build, but can be lost in an instant.

I agree with you... and with that, leadership. Its amazing to see how much of hockey is the last step or move you are willing to take. This is something that I know Lowell knows.

5mn_Major

03-26-2014, 12:53 AM

Unfortunately, 5mn_Major, there's no way to tell by itself if the 0.933 is right. Even if RM wins, that just means that the 7 percent chance came through. That happens... wait for it... 7 percent of the time! All the KRACh probabilities actually means is that when teams as good as Minnesota play teams as below-average as RM, the good teams wins over 9 out of ten times. What part of that seems unlikely to you?

If you're making the somewhat more sophisticated point that RM must have played better lately to get through the AHA tournament and MN must be playing worse to not even get to the Big10 final, then that's something that can be estimated. There doesn't actually seem to be a lot of momentum in college hockey, although writers love to write about it, but there is some. But if you want to make a bet on RM and are willing to take, say, 5-1, let me know and maybe we can make a deal.

Of course I understand how KRACH is calculated. I doubted KRACH odds before...and I was pretty much on target for the women (including the fact that they had too high of odds for the UW women).

KRACH chances of the winning the whole women's championship: Minnesota 72.8%; Wisconsin 9.0%; Cornell 6.6%; Clarkson 5.6%; BC 2.8%; Harvard 1.6%; Mercyhurst 1.3%; BU 0.4%

Hard to say that MN's odds are quite that high. I'd also give UW a chance more in line with eastern schools (Cornell, Clarkson, etc): MN has one more opportunity at an L before they would play an eastern school not named BU...and the game is out east giving a hair of an advantage to eastern schools.

Regarding the men, #1 seed MN (#2 in the country) lost to #4 seed Yale last year. MN just lost to both MI and OSU (at the X) back to back. This is nowhere near a slam dunk. 5:1 in favor of the gophers is about right. Those odds are not .933 or 13:1.

Patman

03-26-2014, 01:16 AM

Of course I understand how KRACH is calculated. I doubted KRACH odds before...and I was pretty much on target for the women (including the fact that they had too high of odds for the UW women).

Regarding the men, #1 seed MN (#2 in the country) lost to #4 seed Yale last year. MN just lost to both MI and OSU (at the X) back to back. This is nowhere near a slam dunk. 5:1 in favor of the gophers is about right. Those odds are not .933 or 13:1.

KRACH is a system of equations set out to solve a maximum likelihood problem. ML estimates are known in certain cases to be over-zealous. Further models are just that, models. Each with their own assumptions built in. Each observation is assumed independent. If there is some latent pattern (say, health, for example) then the observations are not independent but rather dependent on some latent process that itself will often correlate in time.

KRACH does not incorporate uncertainty in the parameter. It can't. KRACH probabilities are nothing more than the most likely probability as resulting from a optimization of the likelihood (invariance principle of the maximum likelihood estimator). It is not, in itself, a probability statement on the overall.

*I am not limiting the means of solution to just Bayesian solutions. Nonetheless, if the underlying uncertainty in a parameter covers a range then so should any probability statement. If I were to approach this in a bayesian posterior-predictive form I am sure that estimates would likely slump closer towards .5. More of a mathematical reality of taking an expected value.

----

If I were so inclined to produce estimates I'd probably use a bayesian approach to soften them up. Even then, it will not account for streaks and runs. I mean, say Connor Hellebuyck and Doug Carr break hands in the world's best high-five. The model isn't going to adapt for that and even then, what information is put in its place.

5mn_Major

03-26-2014, 01:24 AM

Fine. Its still off.

Tipsy McStagger

03-26-2014, 07:42 AM

Nothing to see here folks. 5Mn_major doesn't like the KRACH results so we have to find another way.

SJHovey

03-26-2014, 01:26 PM

Google is your friend, people.

http://www.sportbet.com/lines/hockey_college

Powered by vBulletin® Version 4.2.2 Copyright © 2019 vBulletin Solutions, Inc. All rights reserved.