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.

ECAC Projected Standings 2013-2014

Re: ECAC Projected Standings 2013-2014

FlagDUDE08 said:
If Colgate get two points, the only team that can catch them is Quinnipiac. Number of wins would be exactly the same, so yes, we would go to top four. Union is 3-2 in favor of Colgate, and either Cornell OR CCT will give it to Colgate. If Yale gets the top 4, Quinnipiac is able to overtake. As Cornell holds the tiebreak over Yale, two points would give them the break over Yale and would overtake CCT, so you would be correct.
Click to expand...

It's worth pointing out that after the wins tiebreaker, my program says "lol **** it" and chooses a team at random to be higher. It also doesn't do four(or more)-way ties the way it should.

Ralph Baer said:
Am I correct that the Monte Carlo simulations does not take into account the likelihood of any occurrence. That is, are the odds of the favorite winning each game the same as the odds of the favorite losing each game? Otherwise I am surprised that the two sets of results are as different as they are. I realize that the averages points are very close.
Click to expand...

No, the Monte Carlo still used KRACH. I was fairly surprised to see how big some of the changes were. I don't know if that's because my computer's psuedo-RNG is bad or what.
 
Re: ECAC Projected Standings 2013-2014

Forget all this calculating lingo, Yale is getting fifth, they need the pairwise points. :p
 
Re: ECAC Projected Standings 2013-2014

LTsatch said:
Forget all this calculating lingo, Yale is getting fifth, they need the pairwise points. :p
Click to expand...

There's a pretty good chance you'll (we'll?) take 6th! And if you want pairwise calculations, talk to RHamilton.
 
Re: ECAC Projected Standings 2013-2014

I just re-ran the Monte Carlo (with 10^7 simulations) and there's definitely something funky going on with the Colgate/Quinnipiac/Cornell. Colgate is in second over 6 percentage points fewer times (apparently to Quinnipiac's benefit) in the exhaustive odds. Most of the code is identical (straight copy/paste), so I don't know why there should be such a discrepancy, especially when the average points are identical. The points thing makes me think it's tiebreak related, but who knows.
 
Re: ECAC Projected Standings 2013-2014

lugnut92 said:
No, the Monte Carlo still used KRACH. I was fairly surprised to see how big some of the changes were. I don't know if that's because my computer's psuedo-RNG is bad or what.
Click to expand...
I have trouble believing that even a poor pseudo random number generator would produce odds for teams being in certain places in the standings which differ by as much from the actual as they do when you are taking almost twice as many trials as there actually are possibilities, and yet the average number of points each gets is right on. Strange.

Sorry about the run-on sentence.
 
Re: ECAC Projected Standings 2013-2014

Ralph Baer said:
I have trouble believing that even a poor pseudo random number generator would produce odds for teams being in certain places in the standings which differ by as much from the actual as they do when you are taking almost twice as many trials as there actually are possibilities, and yet the average number of points each gets is right on. Strange.

Sorry about the run-on sentence.
Click to expand...

Yeah, the re-run (with nearly 19 times more trials than possible outcomes) basically eliminated the chance that the huge discrepancy was purely by chance. I'd look deeper now, but I've got actual classwork to do. So basically, everyone take these numbers with a grain of salt until I can figure out what's up.
 
Re: ECAC Projected Standings 2013-2014

lugnut92 said:
Yeah, the re-run (with nearly 19 times more trials than possible outcomes) basically eliminated the chance that the huge discrepancy was purely by chance. I'd look deeper now, but I've got actual classwork to do. So basically, everyone take these numbers with a grain of salt until I can figure out what's up.
Click to expand...

The only thing that I could think of was that you did the tiebreakers differently in the Monte Carlo from the case where you ran all 3^12 possibilities. However I suspect that you used the same code. You better get your homework done and fix this before next season gets to this point. :)
 
Re: ECAC Projected Standings 2013-2014

Ralph Baer said:
The only thing that I could think of was that you did the tiebreakers differently in the Monte Carlo from the case where you ran all 3^12 possibilities. However I suspect that you used the same code. You better get your homework done and fix this before next season gets to this point. :)
Click to expand...

What else is the off-season for, if not for making piles of code to simulate the next hockey season!
 
Re: ECAC Projected Standings 2013-2014

So in the wide world of who the **** knows, changing statements like "a = a + b" to "a += b' (which should be identical) caused numbers to change. Almost all numbers that changed brought the exhaustive numbers closer, but not equal, to the Monte Carlo numbers.

New stuff:
Code: 
EXHAUSTIVE (Every possible outcome (3^12) checked, weighted for odds of occurrence)

   |  KRACH |     1     2     3     4     5     6     7     8     9    10    11    12 |Avg Rk |AvgPts
-----------------------------------------------------------------------------------------------------
Un |  390.2 | [B]100.0[/B]     x     x     x     x     x     x     x     x     x     x     x |  [B]1.00[/B] | 35.81
Cg |  211.0 |     x  [B]87.8[/B]  10.5   1.7     x     x     x     x     x     x     x     x |  2.14 | 29.83
Qu |  256.0 |     x  10.7  [B]58.3[/B]  25.6   4.5   0.8     x     x     x     x     x     x |  3.26 | 27.61
Cr |  230.3 |     x   1.5  30.7  [B]64.1[/B]   3.2   0.6     x     x     x     x     x     x |  3.71 | 26.88
Ya |  170.2 |     x     x   0.2   2.9  [B]49.3[/B]  37.2  10.4     x     x     x     x     x |  5.55 | 23.77
Ck |  134.1 |     x     x   0.3   5.7  40.4  [B]49.7[/B]   3.8   0.1     x     x     x     x |  5.51 | 23.27
RP |  131.2 |     x     x     x     x   2.6  11.5  [B]60.3[/B]  20.6   4.9     x     x     x |  7.13 | 20.95
Br |  114.8 |     x     x     x     x     x   0.1  20.8  [B]43.7[/B]  32.6   2.7   0.1     x |  8.17 | 18.47
SL |  113.2 |     x     x     x     x     x     x   4.7  34.0  [B]54.1[/B]   6.2   0.9     x |  8.64 | 18.16
Ha |   87.8 |     x     x     x     x     x     x     x   0.6   3.6  37.1  [B]58.7[/B]     x | 10.54 | 15.26
Da |   62.5 |     x     x     x     x     x     x     x   1.0   4.8  [B]54.0[/B]  40.2     x | 10.33 | 15.03
Pr |   30.5 |     x     x     x     x     x     x     x     x     x     x     x [B]100.0[/B] | [B]12.00[/B] |  8.96

[b]Bold[/b] is each team's most likely rank.
0.0 means the team could place in that position, but there is less than 0.05% chance of that happening.
x means the team cannot place in that position.

Playoffs:
Code: 
   |    RP    Br    SL    Ha    Da    Pr |   BYE
------------------------------------------------
RP |         3.9  20.6  29.4  38.5   2.6 |     x
Br |   2.8        38.2   9.5  14.2     x |     x
Ya |     x   0.6   0.9  25.6  20.5  49.3 |   3.1
Ck |     x   0.2   0.8  30.6  22.0  40.4 |   6.0
SL |   2.1  29.8         3.4   3.3     x |     x
Cr |     x   0.0   0.0   0.3   0.2   3.2 |  96.3
Ha |     x   0.5   0.1         0.0     x |     x
Da |     x   0.4   0.6   0.0           x |     x
Qu |     x   0.0     x   0.5   0.3   4.5 |  94.6

[u]Bye[/u]
Union
Colgate

Hosts are listed top to bottom, visitors are across the top
(e.g., RPI has a 20.6% chance of hosting SLU in the first round).
 
Re: ECAC Projected Standings 2013-2014

After 28 February:

UC 35 - 37 [1]
COL 27 - 29 [2-3]
QU 27 - 29 [2-3]
--- Bye Lock - 25+ () ---
COR 24 - 26 [4-5]
CCT 23 - 25 [4-6]
--- Bye Eligible - 24+ () ---
YAL 22 - 24 [5-7]
RPI 21 - 23 [6-7]
--- Home Lock - 19 () ---
BRN 17 - 19 [8-11]
SLU 16 - 18 [8-11]
HVD 16 - 18 [8-11]
DC 16 - 18 [8-11]
--- Home Eligible - 17 () ---
PU 8 - 10 [12]



Remaining LEAGUE schedules:
UC: @BRN
QU: @CCT
COL: DC
COR: HVD
YAL: RPI
CCT: PU, QU
RPI: @YAL
SLU: PU
BRN: UC
HVD: @COR
DC: @COL
PU: @CCT, @SLU

Tie-breaks (3pts or more):
UC: CCT, COR, YAL, RPI, HVD, DC, PU
QU: YAL, RPI, HVD, DC,
COL: UC, CCT, COR, SLU, PU
COR: YAL, BRN, SLU, PU
YAL: COL, HVD, SLU, PU
CCT: RPI, BRN, SLU, PU
RPI: COL, COR, DC,
SLU: HVD,
BRN: SLU, HVD,
HVD: RPI, PU
DC: COR, YAL
PU:

ADVANTAGE (2pts with game in hand):
UC: BRN
QU: CCT
COL: DC
COR: HVD
YAL: RPI
CCT:
RPI:
SLU: PU
BRN:
HVD:
DC:
PU:

H2H, #Wins, Record v Top4, Record v Top8

I will dispense with the "specials" and discuss tie-breaks:

Colgate/Quinnipiac: Both would have the same number of wins, so we'd go to Top 4. Regardless of if Cornell or CCT get the last bye spot, Colgate will take a tie-break over Quinnipiac.
CCT/Cornell, CCT/Yale: Clarkson would hold both of these tiebreaks due to number of ECAC wins should one of these occur.
With Cornell holding the tiebreak over Yale, Yale cannot get the bye.
Cornell/CCT/Yale: Cornell 6, Clarkson 4, Yale 2.
Brown/SLU/Harvard/Dartmouth: OK, time for this clusterblank to be explained:
2-way: Brown holds over SLU and Harvard, with Dartmouth it comes to record vs. Top 4, both have 2 on Colgate, Brown has 2 on Quinnipiac and 0-1 on Union. Dartmouth has 3 on Cornell and 2 on CCT, both of which Brown has zero. In top 8, Dartmouth has 3 on Yale to Brown's 2, and if it is for 8th place, would give Dartmouth the tie-break on top 8; Brown beats Dartmouth 3-2 on both SLU and Harvard, so Brown would take those tie-breaks on Top 8. SLU takes over Harvard, and Dartmouth would take the others by virtue of ECAC wins.
3-way: B/D/H: Brown 6, Dartmouth 4, Harvard 2. B/D/S: Brown 5, Dartmouth 4, SLU 3. B/H/S: Brown 7, SLU 4, Harvard 1. D/H/S: SLU 5, Dartmouth 4, Harvard 3.
4-way: Brown 9, Dartmouth 6, SLU 6, Harvard 3; Dartmouth takes SLU by ECAC wins.
 
Last edited:
Re: ECAC Projected Standings 2013-2014

In a 4-way tie between Brown, Dartmouth, Harvard, and SLU, Brown would win the 4-way tie, but then you have to break a 3-way tie between Dartmouth, Harvard, and SLU. SLU would then win that tiebreaker based on cumulative head-to-head points. Then it goes down to a Dartmouth-Harvard tiebreaker. Dartmouth would take that on ECAC wins.

So it would go Brown (8th), SLU (9th), Dartmouth (10th), and Harvard (11th).
 
Re: ECAC Projected Standings 2013-2014

Rainman said:
In a 4-way tie between Brown, Dartmouth, Harvard, and SLU, Brown would win the 4-way tie, but then you have to break a 3-way tie between Dartmouth, Harvard, and SLU. SLU would then win that tiebreaker based on cumulative head-to-head points. Then it goes down to a Dartmouth-Harvard tiebreaker. Dartmouth would take that on ECAC wins.

So it would go Brown (8th), SLU (9th), Dartmouth (10th), and Harvard (11th).
Click to expand...

Not exactly. The ECAC states on tiebreakers involving three or more teams: "In the case of ties among three or more schools, the criteria will be used in order until a team, or teams, is separated from the pack. At that point, the process will begin anew to break the "new" tie. In other words, when a four-way tie becomes a three-way tie, the three-way tie is treated as a "new" tie and the process begins with the first criterion." As shown by the points (and it's actually 9-6-6-3 instead of 7-6-6-3, I can't add), two teams are separated from the pack: Brown 8th and Harvard 11th. THEN the two-way tie-break for 9th occurs, with Dartmouth getting 9th and SLU getting 10th.
 
Re: ECAC Projected Standings 2013-2014

Mid-weekend update:

Union is locked into the #1 seed.
Colgate and Quinnipiac can finish as high as 2nd, but no lower than 3rd. Colgate controls its own destiny for 2nd because it wins the tiebreaker with the Bobcats.
Cornell can finish either 4th or 5th. They control their own destiny for 4th. If they finish tied for 4th with Clarkson, Clarkson takes the final bye with the ECAC wins tiebreaker.
Clarkson can finish 4th through 6th. They control their own destiny for 5th, but need some help from Harvard to get the bye. They hold the head-to-head tiebreaker with RPI.
Yale can finish 5th through 7th. They are guaranteed a 1st-round home series.
RPI can finish 6th or 7th. They are also guaranteed a 1st-round home series.
Brown can finish 8th through 11th. They control their own destiny for the final home-ice spot.
St. Lawrence, Dartmouth, and Harvard also can finish between 8th and 11th. But they need help if they want to play at home next weekend.
Princeton is locked into the 12th seed.
 
Re: ECAC Projected Standings 2013-2014

FlagDUDE08 said:
When a four-way tie becomes a three-way tie, the three-way tie is treated as a "new" tie and the process begins with the first criterion.
Click to expand...

But when Brown is extrapolated as the winner of the 4-way tie, you then have a 3-way tie with the other teams and you start over, comparing the head-to-head only with the remaining teams. In a SLU-Dartmouth-Harvard 3-way tie, SLU wins on head-to-head.

Are also simultaneously extrapolating Harvard from the 4-way tie as the "loser", as opposed to extrapolating teams one-by-one?
 
Re: ECAC Projected Standings 2013-2014

AS OF GAMES COMPLETED BEFORE 10 PM FRIDAY NIGHT:

Exhaustive:
Code: 
EXHAUSTIVE (Every possible outcome (3^6) checked, weighted for odds of occurrence)

   |  KRACH |     1     2     3     4     5     6     7     8     9    10    11    12 |Avg Rk |AvgPts
-----------------------------------------------------------------------------------------------------
Un |  402.2 | [B]100.0[/B]     x     x     x     x     x     x     x     x     x     x     x |  [B]1.00[/B] | 36.50
Cg |  189.2 |     x  [B]55.3[/B]  44.7     x     x     x     x     x     x     x     x     x |  2.45 | 28.39
Qu |  267.6 |     x  44.7  [B]55.3[/B]     x     x     x     x     x     x     x     x     x |  2.55 | 28.27
Cr |  200.8 |     x     x     x  [B]83.9[/B]  15.0   1.1     x     x     x     x     x     x |  4.17 | 25.30
Ck |  138.6 |     x     x     x  14.5  [B]58.4[/B]  27.0     x     x     x     x     x     x |  5.12 | 23.73
Ya |  161.2 |     x     x     x   1.6  26.5  31.6  [B]40.2[/B]     x     x     x     x     x |  6.10 | 23.06
RP |  140.0 |     x     x     x     x     x  40.2  [B]59.8[/B]     x     x     x     x     x |  6.60 | 21.94
Br |  107.1 |     x     x     x     x     x     x     x  [B]38.9[/B]  34.7  22.0   4.4     x |  8.92 | 17.50
SL |  106.9 |     x     x     x     x     x     x     x  [B]38.3[/B]  32.9  18.8   9.9     x |  9.00 | 17.49
Ha |   98.2 |     x     x     x     x     x     x     x   6.5  22.2  14.2  [B]57.1[/B]     x | 10.22 | 16.70
Da |   70.7 |     x     x     x     x     x     x     x  16.3  10.1  [B]45.0[/B]  28.6     x |  9.86 | 16.61
Pr |   29.4 |     x     x     x     x     x     x     x     x     x     x     x [B]100.0[/B] | [B]12.00[/B] |  8.51

[b]Bold[/b] is each team's most likely rank.
0.0 means the team could place in that position, but there is less than 0.05% chance of that happening.
x means the team cannot place in that position.

Playoffs:
Code: 
   |    Br    SL    Ha    Da    Pr |   BYE
------------------------------------------
RP |  15.3  14.3  32.6  37.8     x |     x
Br |        23.4   6.8   8.7     x |     x
Ya |   9.7  11.4  24.2  26.6  26.5 |   1.6
Ck |   1.1   2.9  14.5   8.6  58.4 |  14.5
SL |  25.1        12.2   1.1     x |     x
Cr |   0.2   0.2     x   0.6  15.0 |  83.9
Ha |   5.3   0.8         0.4     x |     x
Da |   4.3   8.7   3.3           x |     x

[u]Bye[/u]
Union
Colgate
Quinnipiac

Hosts are listed top to bottom, visitors are across the top
(e.g., RPI has a 15.3% chance of hosting Brown in the first round).
 
Re: ECAC Projected Standings 2013-2014

Rainman said:
But when Brown is extrapolated as the winner of the 4-way tie, you then have a 3-way tie with the other teams and you start over, comparing the head-to-head only with the remaining teams. In a SLU-Dartmouth-Harvard 3-way tie, SLU wins on head-to-head.

Are also simultaneously extrapolating Harvard from the 4-way tie as the "loser", as opposed to extrapolating teams one-by-one?
Click to expand...

It doesn't have to go 4-3-2, it says "separated from the pack". Clearly, Harvard has been separated from the pack in 4th of 4 in the case of a four-way.
 
Re: ECAC Projected Standings 2013-2014

lugnut92 said:
AS OF GAMES COMPLETED BEFORE 10 PM FRIDAY NIGHT:

Exhaustive:
Code: 
EXHAUSTIVE (Every possible outcome (3^6) checked, weighted for odds of occurrence)

   |  KRACH |     1     2     3     4     5     6     7     8     9    10    11    12 |Avg Rk |AvgPts
-----------------------------------------------------------------------------------------------------
Un |  402.2 | [B]100.0[/B]     x     x     x     x     x     x     x     x     x     x     x |  [B]1.00[/B] | 36.50
Cg |  189.2 |     x  [B]55.3[/B]  44.7     x     x     x     x     x     x     x     x     x |  2.45 | 28.39
Qu |  267.6 |     x  44.7  [B]55.3[/B]     x     x     x     x     x     x     x     x     x |  2.55 | 28.27
Cr |  200.8 |     x     x     x  [B]83.9[/B]  15.0   1.1     x     x     x     x     x     x |  4.17 | 25.30
Ck |  138.6 |     x     x     x  14.5  [B]58.4[/B]  27.0     x     x     x     x     x     x |  5.12 | 23.73
Ya |  161.2 |     x     x     x   1.6  26.5  31.6  [B]40.2[/B]     x     x     x     x     x |  6.10 | 23.06
RP |  140.0 |     x     x     x     x     x  40.2  [B]59.8[/B]     x     x     x     x     x |  6.60 | 21.94
Br |  107.1 |     x     x     x     x     x     x     x  [B]38.9[/B]  34.7  22.0   4.4     x |  8.92 | 17.50
SL |  106.9 |     x     x     x     x     x     x     x  [B]38.3[/B]  32.9  18.8   9.9     x |  9.00 | 17.49
Ha |   98.2 |     x     x     x     x     x     x     x   6.5  22.2  14.2  [B]57.1[/B]     x | 10.22 | 16.70
Da |   70.7 |     x     x     x     x     x     x     x  16.3  10.1  [B]45.0[/B]  28.6     x |  9.86 | 16.61
Pr |   29.4 |     x     x     x     x     x     x     x     x     x     x     x [B]100.0[/B] | [B]12.00[/B] |  8.51

[b]Bold[/b] is each team's most likely rank.
0.0 means the team could place in that position, but there is less than 0.05% chance of that happening.
x means the team cannot place in that position.

Playoffs:
Code: 
   |    Br    SL    Ha    Da    Pr |   BYE
------------------------------------------
RP |  15.3  14.3  32.6  37.8     x |     x
Br |        23.4   6.8   8.7     x |     x
Ya |   9.7  11.4  24.2  26.6  26.5 |   1.6
Ck |   1.1   2.9  14.5   8.6  58.4 |  14.5
SL |  25.1        12.2   1.1     x |     x
Cr |   0.2   0.2     x   0.6  15.0 |  83.9
Ha |   5.3   0.8         0.4     x |     x
Da |   4.3   8.7   3.3           x |     x

[u]Bye[/u]
Union
Colgate
Quinnipiac

Hosts are listed top to bottom, visitors are across the top
(e.g., RPI has a 15.3% chance of hosting Brown in the first round).
Click to expand...

Tie-breaker is still broken; Yale can't get the bye because Cornell would take the tie-break. Even in a three-way, Cornell would have 6 points to Clarkson's 4 and Yale's 2.
 
Re: ECAC Projected Standings 2013-2014

Rainman is correct on the tiebreaker procedure. History shows that multi-way ties in ECAC play are broken in a "top-down" fashion.

Brown would be granted 8th in a 4-way tie and then the tiebreaker procedure would start over as a three-way tie between SLU, Dartmouth, and Harvard, which SLU wins on head-to-head play to take 9th.

Since FlagDUDE's Brown / SLU / Dartmouth / Harvard tiebreaking guide was either incomprehensible or wrong, you can also check out WaP's tiebreaking guide to see who actually wins what.
 
You must log in or register to reply here.
Back
Top