PDA

View Full Version : ECAC Pick the Standings 2017-18



Pages : 1 2 3 [4] 5 6

DrDemento
10-27-2017, 05:19 PM
Just got back in time!!!! Will have to put both my picks and Jenny's together on this one post as she cannot seem to remember her login information password and is too pressed for time so here goes.

Me

1. Harvard
2. Quinnipiac
3. Cornell
4. Yale
5. RPI
6. Union
7. Clarkson
8. SLU
9. Princeton
10. Dartmouth
11. Colgate
12. Brown


Jenny

1. Harvard
2. Quinnipiac
3. Cornell
4. Union
5. RPI
6. Clarkson
7.Yale
8. Princeton
9. Dartmouth
10.Colgate
11. St. Lawrence
12. Brown

jmhusker
10-27-2017, 05:25 PM
1. Harvard
2. Clarkson
3. Quinnipiac
4. RPI
5. Union
6. Cornell
7. Yale
8. Dartmouth
9. St. Lawrence
10. Princeton
11. Colgate
12. Brown

lugnut92
10-29-2017, 04:57 PM
Here are the final statistics for all human entries (not including the Random Number, Mr. Last Year, and the Fan Average; n=46):

Team | Mean±SD | Median(Range) |
-----------------------------------------
Harvard | 1.59±1.44 | 1 ( 1-10) |
Quinnipiac | 2.59±1.24 | 2 ( 1- 6) |
Clarkson | 3.80±2.07 | 3 ( 1-12) |
Cornell | 3.83±1.27 | 4 ( 2- 7) |
Princeton | 5.91±1.98 | 5 ( 1-12) |
Yale | 6.57±1.63 | 7 ( 3-11) |
Union | 6.98±2.01 | 7 ( 3-12) |
RPI | 8.13±2.86 | 9 ( 1-12) |
SLU | 8.30±2.40 | 9 ( 1-12) |
Dartmouth | 9.48±1.95 | 10 ( 2-12) |
Colgate | 9.89±1.87 | 10 ( 3-12) |
Brown | 10.98±1.33 | 11.5 ( 7-12) |

The most consensus around any given team was Quinnipiac, though Harvard received more votes for a single position (1st, 31) than any other team. RPI had the least consensus, evidenced with a very wide distribution and the fewest votes for its mode (10th, 8) of any team. As the saying goes, never bet for or against RPI!

No team was picked to finish in every position, and no position had every team picked at least once.

hockeyplayer1015's entry most closely matches the fan average (only flipping the positions of Harvard and Quinnipiac), while the RPI Standings Board is the most different (ColSkate65 and Rjm7272 hold that crown for those who actually post on the board). As I mentioned in a previous post, LTsatch and The Coaches made identical entries.

Visualizations
<img src="https://i.imgur.com/bBbgBll.png" height="350"></img><img src="https://i.imgur.com/VpOVxeu.png" height="350"></img>

Data on the first graph represent the mean ± SD ranking of all human entries. The second shows the frequency with which each team was predicted for each position.

ColSkate65
10-29-2017, 05:13 PM
lugnut, TMI, you are way over my head. i was an English major, not a math major ;)

lugnut92
10-29-2017, 10:24 PM
Well we're five games in, and we have to fake a whole hell of a lot of tiebreakers to make a set of standings that even remotely makes sense, but that's what happens when half of the league has no games played and 2 more split their home and home (with even goal differential). sshablak has the early lead, and (I think I'll be saying this an awful lot) RPI Standings Board is bringing up the rear. So here are your basically meaningless initial standings:

Rk Post# Contestant ρ
1 48 sshablak 0.5734
2 23 Mike Hutter 0.5385
3 60 TchrBill 0.5315
4 3 Mr. Last Year 0.4895
46 goldenbriggsy 0.4895
6 16 BigAl 0.4825
7 3 The Media 0.4615
2 TBRW 0.4615
18 bothman 0.4615
10 40 Dutchman 0.4545
11 33 Rainman 0.4196
12 53 johnk 0.3986
13 30 Section5forLife 0.3846
52 kdiff77 0.3846
15 32 cudmud 0.3776
59 TheValleyRaider 0.3776
17 3 The Coaches 0.3636
5 Kepler 0.3636
22 techtodenver 0.3636
41 LTsatch 0.3636
54 BIG Green 0.3636
61 JennyRN 0.3636
23 9 IsThereADog 0.3601
24 17 ecachawkeyfan 0.3566
63 TheFanAverage 0.3566
26 20 Muskieman 0.3497
42 FireKnight 0.3497
56 hockeyplayer1015 0.3497
29 61 DrDemento 0.3427
30 47 GoldenBear 0.3357
31 21 alslammerz 0.3252
32 28 UFan 0.3077
33 10 slumulva88 0.3007
34 44 QUAlum2004 0.2867
35 34 Saints#1 0.2727
57 1988 0.2727
37 49 ColSkate65 0.2587
38 8 miker 0.2308
39 14 Go Puckman 0.2098
40 31 Turk 0.1818
39 drshoen 0.1818
42 62 jmhusker 0.1469
43 58 red133883 0.1399
44 11 lugnut92 0.0909
45 26 FlagDUDE08 0.0699
50 Rjm7272 0.0699
47 15 Ralph Baer -0.0140
48 3 The Random Number -0.5315
49 25 RPIStandingsBoard -0.7413

Okay, so I mentioned arbitrary tiebreakers, and here they are. If we extend past those tiebreakers provided by the ECAC, I'll basically try and apply the same thought process to teams' overall records. So right now, I'm just comparing win percentages and games played. That doesn't even fly for Brown and Yale, so I ranked them by shot differential. None of this matters, no one is going to care what tiebreakers I used 127 games from now.

Fun fact: Any time the scores here are negative, it means the prediction would be more accurate if it were reversed. So the RPI Standings Board would be #1 if it were literally backwards. But again, none of this matters, we're five games in.

Finally, I want to ask if people prefer the [code] tags I use for the formatted table or if they like the old [pre] tags better? [code] is smaller and more compact, but you end up having to scroll to the bottom. Does anyone care?

Ralph Baer
10-30-2017, 01:47 AM
I am the only actual entrant who is currently negative. :)

DrDemento
10-30-2017, 01:51 PM
I am the only actual entrant who is currently negative. :)

But you keep a positive outlook!!

DrDemento
10-30-2017, 01:52 PM
Whatever Lug is doing, and I have no understanding of the math involved, it must be correct since Jenny is, as usual, about 10-20 slots above me.

lugnut92
10-30-2017, 02:36 PM
Whatever Lug is doing, and I have no understanding of the math involved, it must be correct since Jenny is, as usual, about 10-20 slots above me.

It's actually pretty simple. The rankings this year will be identical to if I listed the scores the old way, but just scaled slightly differently. If you were to divide the old scores by 286 (i.e., multiply by 6/(n(n2-1)) ) and subtract the result from 1, you get the new scores. It's all just a way to limit the scores to a much smaller range.

Ralph Baer
10-30-2017, 03:08 PM
It's actually pretty simple. The rankings this year will be identical to if I listed the scores the old way, but just scaled slightly differently. If you were to divide the old scores by 236 (i.e., multiply by 6/(n(n2-1)) ) and subtract the result from 1, you get the new scores. It's all just a way to limit the scores to a much smaller range.

Is there something wrong with the formula, or am I confused? I think you said that three of the entrants aren't counted. I.e., n = 46. That doesn't give 1/236, but rather 1/16215.

lugnut92
10-30-2017, 03:16 PM
Is there something wrong with the formula, or am I confused? I think you said that three of the entrants aren't counted. I.e., n = 46. That doesn't give 1/236, but rather 1/16215.

Oh sorry, the n is the number of ranks being compared (i.e., 12). I suppose reusing variables isn't very good practice.

Also, it should be 286, not 236. That was a mistake on my part.

Kepler
11-01-2017, 10:22 AM
Oh sorry, the n is the number of ranks being compared (i.e., 12). I suppose reusing variables isn't very good practice.

Also, it should be 286, not 236. That was a mistake on my part.

Thank you again for doing this; it's great!

Kepler
11-01-2017, 10:23 AM
Finally, I want to ask if people prefer the [code] tags I use for the formatted table or if they like the old [pre] tags better? [code] is smaller and more compact, but you end up having to scroll to the bottom. Does anyone care?

I like the code tags better than the pre tags, personally.

I use them myself and wish there was a way of overwriting the title (code or pre) with custom text (e.g., a caption). Is there a way to do that?

lugnut92
11-01-2017, 12:51 PM
I like the code tags better than the pre tags, personally.

I use them myself and wish there was a way of overwriting the title (code or pre) with custom text (e.g., a caption). Is there a way to do that?

I don't think so. I've always wanted to do it.



Nope


<code="neither does this">
Nope
</code>

vicb
11-05-2017, 07:50 AM
Thanks for taking over the contest this year. I just did not have the time or energy to do this again (it is tough getting old). Good luck to all the prognosticators and may the luckiest one win ;) .

lugnut92
11-05-2017, 04:41 PM
Well, no one is doing particularly well after 17 games. It doesn't help that Quinnipiac is in last and Dartmouth is sitting in the top 4, but I imagine those small errors will be rectified soon enough. Anyway, goldenbriggsy has jumped up a few spots to take over the top spot, while the RPI Standings Board has a pretty tight grip on the bottom spot already (FlagDUDE08 holds that honor for real posters).

Here are your new standings:

Rk Post# Contestant ρ
1 46 goldenbriggsy 0.4615
2 48 sshablak 0.3497
3 20 Muskieman 0.3287
4 21 alslammerz 0.3182
5 54 BIG Green 0.2937
6 3 Mr. Last Year 0.2657
7 60 TchrBill 0.2238
8 16 BigAl 0.2168
9 58 red133883 0.2098
62 jmhusker 0.2098
11 30 Section5forLife 0.1608
33 Rainman 0.1608
13 17 ecachawkeyfan 0.1469
14 15 Ralph Baer 0.1329
15 18 bothman 0.1259
56 hockeyplayer1015 0.1259
17 2 TBRW 0.1189
49 ColSkate65 0.1189
19 61 JennyRN 0.0839
20 22 techtodenver 0.0699
21 3 The Coaches 0.0629
8 miker 0.0629
23 Mike Hutter 0.0629
41 LTsatch 0.0629
42 FireKnight 0.0629
26 5 Kepler 0.0559
63 TheFanAverage 0.0559
28 3 The Media 0.0420
29 14 Go Puckman 0.0350
30 47 GoldenBear 0.0210
31 53 johnk 0.0140
32 9 IsThereADog 0.0035
33 28 UFan -0.0210
34 44 QUAlum2004 -0.0280
35 10 slumulva88 -0.0350
36 31 Turk -0.0490
34 Saints#1 -0.0490
38 57 1988 -0.0629
39 40 Dutchman -0.0699
61 DrDemento -0.0699
41 59 TheValleyRaider -0.0769
42 52 kdiff77 -0.0839
43 39 drshoen -0.1049
44 50 Rjm7272 -0.1399
45 32 cudmud -0.2028
46 11 lugnut92 -0.2238
47 26 FlagDUDE08 -0.4056
48 3 The Random Number -0.4196
49 25 RPIStandingsBoard -0.6643

Pretty much every current tie in the ECAC table was broken by games played, with only one tie falling to H2H points.

FlagDUDE08
11-05-2017, 06:00 PM
For at least the purposes of first and last place, do we know what the classic-style scores are so we can keep tabs on how we're doing when compared to last year?

lugnut92
11-05-2017, 07:41 PM
For at least the purposes of first and last place, do we know what the classic-style scores are so we can keep tabs on how we're doing when compared to last year?

I'll start reporting it as "Raw".

Rk Post# Contestant Raw ρ
1 46 goldenbriggsy 154 0.4615
2 48 sshablak 186 0.3497
3 20 Muskieman 192 0.3287
4 21 alslammerz 195 0.3182
5 54 BIG Green 202 0.2937
6 3 Mr. Last Year 210 0.2657
7 60 TchrBill 222 0.2238
8 16 BigAl 224 0.2168
9 58 red133883 226 0.2098
10 62 jmhusker 226 0.2098
11 30 Section5forLife 240 0.1608
33 Rainman 240 0.1608
13 17 ecachawkeyfan 244 0.1469
14 15 Ralph Baer 248 0.1329
15 18 bothman 250 0.1259
56 hockeyplayer1015 250 0.1259
17 2 TBRW 252 0.1189
49 ColSkate65 252 0.1189
19 61 JennyRN 262 0.0839
20 22 techtodenver 266 0.0699
21 3 The Coaches 268 0.0629
8 miker 268 0.0629
23 Mike Hutter 268 0.0629
41 LTsatch 268 0.0629
42 FireKnight 268 0.0629
26 5 Kepler 270 0.0559
63 TheFanAverage 270 0.0559
28 3 The Media 274 0.0420
29 14 Go Puckman 276 0.0350
30 47 GoldenBear 280 0.0210
31 53 johnk 282 0.0140
32 9 IsThereADog 285 0.0035
33 28 UFan 292 -0.0210
34 44 QUAlum2004 294 -0.0280
35 10 slumulva88 296 -0.0350
36 31 Turk 300 -0.0490
34 Saints#1 300 -0.0490
38 57 1988 304 -0.0629
39 40 Dutchman 306 -0.0699
61 DrDemento 306 -0.0699
41 59 TheValleyRaider 308 -0.0769
42 52 kdiff77 310 -0.0839
43 39 drshoen 316 -0.1049
44 50 Rjm7272 326 -0.1399
45 32 cudmud 344 -0.2028
46 11 lugnut92 350 -0.2238
47 26 FlagDUDE08 402 -0.4056
48 3 The Random Number 406 -0.4196
49 25 RPIStandingsBoard 476 -0.6643

If the season ended today, these would be the worst winning and losing scores of all time.

FlagDUDE08
11-05-2017, 09:03 PM
I'll start reporting it as "Raw".

If the season ended today, these would be the worst winning and losing scores of all time.

I'm almost sorry I asked. :eek::p

lugnut92
11-12-2017, 03:37 PM
Not much changing right now in the standings. goldenbriggsy maintains his hold on first overall, improving slightly over his position last week. Behind him, alslammerz leapfrogged sshablak to take second. Elsewhere in the standings, slumulva88 has a score of exactly 0, meaning that their prediction has exactly no correlation with the real standings. Down at the bottom, RPIStandingsBoard is really bad and FlagDUDE08 is too.

Here are your new standings:

Rk Post# Contestant Raw ρ
1 46 goldenbriggsy 146 0.4895
2 21 alslammerz 167 0.4161
3 48 sshablak 186 0.3497
4 20 Muskieman 194 0.3217
5 3 Mr. Last Year 210 0.2657
6 49 ColSkate65 216 0.2448
7 60 TchrBill 218 0.2378
8 2 TBRW 220 0.2308
54 BIG Green 220 0.2308
10 33 Rainman 226 0.2098
11 5 Kepler 232 0.1888
12 16 BigAl 234 0.1818
58 red133883 234 0.1818
14 22 techtodenver 236 0.1748
30 Section5forLife 236 0.1748
61 JennyRN 236 0.1748
17 56 hockeyplayer1015 242 0.1538
62 jmhusker 242 0.1538
19 18 bothman 244 0.1469
20 8 miker 248 0.1329
17 ecachawkeyfan 248 0.1329
22 15 Ralph Baer 250 0.1259
23 47 GoldenBear 256 0.1049
53 johnk 256 0.1049
25 28 UFan 258 0.0979
26 63 TheFanAverage 262 0.0839
27 3 The Coaches 264 0.0769
41 LTsatch 264 0.0769
29 42 FireKnight 266 0.0699
30 23 Mike Hutter 268 0.0629
31 3 The Media 274 0.0420
14 Go Puckman 274 0.0420
40 Dutchman 274 0.0420
34 31 Turk 280 0.0210
35 61 DrDemento 284 0.0070
36 10 slumulva88 286 0.0000
37 34 Saints#1 290 -0.0140
52 kdiff77 290 -0.0140
39 44 QUAlum2004 292 -0.0210
9 IsThereADog 293 -0.0245
41 57 1988 296 -0.0350
42 32 cudmud 300 -0.0490
43 11 lugnut92 306 -0.0699
44 59 TheValleyRaider 308 -0.0769
45 39 drshoen 320 -0.1189
46 50 Rjm7272 342 -0.1958
47 3 The Random Number 358 -0.2517
48 26 FlagDUDE08 388 -0.3566
49 25 RPIStandingsBoard 464 -0.6224