Texas A&M is up 4-0 over Arizona and it's only the first half. Wow. Also, Illinois state is down 0-2 to Pepperdine at half time. I love the underdog, but there's usually reasons why they're the underdog.
Whoop! Aggies now up 5-0 on Arizona (halftime). It's a good night out here at Ellis Field. Kelley Monogue's 2nd goal, the deflected header into the top corner from Leigh Edwards, was absolutely superb.
Second Round Scores Human Division So with the second round all put to bed, (I'm assuming that Harvard will not climb out of their 4-0 hole to take down UCLA) we have a new leader in uncchamps2012 Hot on his Heels is bluesky222 followed by Kolabear and HeadSpun. This is a tight contest (given that most contestants were agreed on the vast majority of places), so the scores are tightly packed for the moment. Contestants should see more daylight in the next two rounds. 68 uncchamps2012 67 Bluesky222 66 Kolabear 66 HeadSpun 65 keithscarlett 65 cachundo 64 SoCalSun 64 GrassTopper 64 Ingoldsby 63 BruBru 63 RtD! 63 raiderD15 62 Justdoit 61 hykos1045 60 Got Jukes? 60 soccerhunter 59 Tom81 58 Kickithard 57 dmthomas49 56 Glove Stinks 55 liesse00 55 Bosco 51 Hooked003 Statistical Machine Division In the Statistical Machine division the Iteration 5 and Massey are tied for first place, followed by Bennett, the Adjusted RPI, and the NCAA Standards. Bringing up the rear is the Old Recruiting Scores, the only statistical approach not taking into account this season's play. As was true last year, the top human players beat the statistics, but on the whole, the statistics beat the human average. We'll have to see how this ends up. 65 Massey Machine 65 Iteration 5 URPI 64 Bennett Machine 63 Adjusted RPI 61 NCAA Standards 60 Albyn Jones/Massey 57 Old Recruiting Scores (PS. I'll be driving all weekend and will not be able to post the official 3rd round scores until late in the evening on Sunday. Thanks for your patience.)
by my quick count, only 6 of the 23 non-computer entries picked all #1 and #2 seeds to advance. 17 out of 23 picked at least one of the top seeds to fall by the end of the 3rd round. The odds of at least one #1 or #2 seed falling is quite good; the question is of course which one and the risk of getting it wrong. Of course to win the prediction contest takes some sort of risk. It might be interesting to note that 9 out of the 17 who picked a lower-seeded or unseeded team to advance past the 3rd round have already seen their picks eliminated. That doesn't necessarily hurt them that much. For instance, those who picked Cal or Florida Gulf Coast in place of #2 seed Florida may not be hurt in the standings much if Florida falls to someone else (in this case Texas Tech). At least they'll stay even with the 13 who picked Florida. Florida led the #1 and #2 seeded teams in terms of who was expected to fall with 10 out of 23 picking another team besides Florida. The most expected "upsets" among the 1st and 2nd seeds in order are: Florida 10 out of 23 entries Texas A&M 7/23 Penn State 6/23 North Carolina 3/23 Virginia 1/23 Unless I missed something, UCLA, Florida State and Stanford were picked by all the entrants. Notre Dame led the underdogs with 7 entrants picking them. However it's quite possible that only (3) would have a net benefit by picking the lower seeded team because (4) entrants also selected another underdog who's already eliminated. Texas Tech got 5 picks and 4 of them would certainly have a net gain compared to those who selected all the top seeds. Virginia Tech would definitely benefit 3 entrants while Kentucky and South Carolina would benefit one each in case those teams advanced.
Well, I think hykos was the only one to pick South Carlina over North Carolina. Nice call! I'm pretty sure that will boost him or her to within one point of the lead! (and no, he/she doesn't pick South Carolina to win in the quarterfinals) I said earlier my picks were "boring" in that I wound up going with the top seeds in this round (all the #1 and #2 seeds). But I believe, all the leaders after this round will be entrants who picked the #1 and #2 seeds to advance, with the exception of hykos! This also means the machines will do well, as I believe all the ranking systems forecast the #1 and #2 seeds to advance. Shows it's tough to beat the machines. (There's a few games to be played as I write this but it won't affect the top of the standings as nearly everyone picked the remaining #1 seeds, UCLA, Stanford, and Florida St, to advance along with #2 seed Virginia. The only entrant to pick Kentucky over Virginia I think also took Florida Gulf Coast and Notre Dame in this round so they can't move up towards the top of the standings even if Kentucky prevails)
What freshman Abbey Crider is doing for the Gamecocks in goal is just incredible. She only had three save opportunities today against UNC, but they were big ones and she corralled them. I did not foresee these events with Crider but got very lucky with my upset pick which was based on Sabrina's ability to turn a clean sheet. Now that they have actually surpassed UNC, I'm rooting for SC to continue on to the College Cup and bust my bracket, as well as everyone else's today. What an incredible 1 goals to none run to the Elite 8!
The big news today, of course, was the upset at UNC. Only Hykos got this figured out. I didn't see the game, but from everything I've heard, SC keeper Abbey Crider came up big today with some excellent saves. Kudos to her. Complete 3rd round scores follow. As can be seen, the statistic based entries are doing solidly well, but the top 4 humans currently have them beat. However, it could get interesting going forward. Uncchamps is set to fall down over the last three rounds, but could stay near the top if key upsets occur going forward. Likewise, Hykos will have had to correctly predict more upsets to stay in contention. Maybe he has! If Standord takes it all Bluesky will win, and Kolabear looks to remain very strong until the last game but he failed to [pick an eventual winner, so will have to see Florida Sate win to stay ahead of the Statistical entries. This will be interesting yet. 3rd Round Human Division 110 uncchamps2012 109 Bluesky222 109 hykos1045 108 Kolabear 107 cachundo 107 keithscarlett 105 BruBru 102 Got Jukes? 100 GrassTopper 100 Ingoldsby 99 raiderD15 98 Justdoit 96 HeadSpun 95 Tom81 94 Kickithard 94 SoCalSun 93 dmthomas49 90 soccerhunter 87 Hooked003 87 RtD! 86 Glove Stinks 85 Bosco 85 liesse00 3rd Round Machine Statistical Division 107 Iteration 5 URPI 107 Massey Machine 106 Bennett Machine 105 Adjusted RPI 103 NCAA Standards 102 Albyn Jones/Massey 93 Old Recruiting Scores
MACHINES DIVISION: Pts. Massey……… 110 Iteration5uRPI. 110 aRPI……….. 108 Standards……. 106 Bennett…….. 106 HomefieldJones. 105 Recruiting….. 90 Below is an attachment showing your picks history, as well as all the remaining bracket picks. Red highlight in the finals means an earlier round bracketbuster occurred.
Virginia all the way! There are video highlights of the SC-UNC match. Looks a little shaky. Could be the buffering on my end. My pick was for Sabrina D'Angelo to figure out UNC, and Abbey Crider bailed SC (and me) out once again! http://www.gamecocksonline.com/…/w-so…/recaps/112314aab.html
Quarterfinals will be interesting. A lot can happen. There's enough points in these last 3 rounds to really shake up the leaderboard. I don't even think Bluesky is guaranteed to win if Stanford wins because I think there are people close enough to him who can make up points by picking Virginia, Penn State or Florida St instead of UCLA and Texas A&M, for instance. (as well as the "burnt card" North Carolina) This is also where people have a chance to beat the computers and get ahead of them Speaking of computer entries, I thought about but forgot to enter a bracket that I would simply call "The Committee". But conceivably there's a couple ways to do this - one which incorporates the adjusted RPI and the other which could use cpthomas' "Standards" analysis. Last year, I think I entered a bracket for The Committee using the adjusted RPI as the default choice. Put simply, The Committee's picks are based on seeds as announced , regardless of how they came up with them, and the RPI where seeds weren't involved. Rule 1) Seeds are picked over non-seeds. 2) Higher seeds are picked over lower seeds 3) when it comes to the #1 seeds, it's assumed the bracket is drawn like a similar bracket. So the seed in the upper-left (for instance) would be placed like a 1.1 or a traditional #1 seed. Where teams aren't seeded, the team with the higher adjusted RPI is picked. I think this is commonsense inference of as to which of two teams the Committee thinks is stronger and therefore more likely to win. Using this method I think TheCommittee would currently have a point total of 109, tied for 2nd among humans.
I actually have 110 for iteration5URPI, not 107. I can't find the 3 pt swing. I have verified my math (I think), and my brackets are printed with tallies next to each result. UCLA 1 HARVARD 1 (south florida) PEPPERDINE 1 KENTUCKY 1 AZ ST 1 RUTGERS 1 VIRGINIA 1 PENNST 1 UCONN 1 VA TECH 1 (west virginia) NOTRE DAME 1 TEXAS 1 ARIZONA 1 A&M 1 FLORIDA ST 1 (boston u) CENTRAL FLA 1 WISC 1 SOUTH CAR 1 (wash st ) (brigam young) UNC 1 FLA 1 CAL 1 AUBURN 1 TEX TECH 1 WASH 1 (kansas) ARKANSAS 1 STANFORD 1 UCLA 3 PEPP 3 KENTUCKY 3 VIRGINIA 3 PENN ST 3 (west virginia) NOTRE DAME 3 A&M 3 FLORIDA ST 3 (wisc) SOUTH CAR 3 UNC 3 FLA 3 TEX TECH 3 WASH 3 STAN 3 UCLA 6 VA 6 PENNST 6 A&M 6 FLORIDA ST 6 (unc) FLA 6 STANFORD 6 110
I also come up with 110 for Iteration 5. Also 108 for Adjusted RPI But 105 for "Standards" Also Massey 110 Massey/homefield 105 * I found a mistake in my own The Committee bracket (RPI would place LaSalle over Rutgers). So The Committee would have a score of 108, not 109.
Thanks to you sharp eyed folks I have now caught the error. (I even see where I got it if you look at the NCAA scoring bracket http://www.ncaa.com/interactive-bracket/soccer-women/d1 in that they dropped the first round score with Washington and I mistakenly followed the bracket down.) So the corrected third round scores: 110 Iteration 5 URPI 110 Massey Machine 108 Adjusted RPI 106 Bennett Machine 106 NCAA Standards 105 Albyn Jones/Massey 96 Old Recruiting Scores
Scores going into the final four.... Well. Well. Big surprise which sure mare a significant difference in the score sheet. Only 4 humans (and no statistical entries) got the UVA-UCLA winner. This made no difference to the stat entries since all contenders were identical. But it made a big difference in then Human division. The big winner was Bosco who went from a tie for dead last to within striking distance. He was the only contestant to pick the final four teams and if he is correct for his three remaining picked teams, he will tie for first place with Hykos. This would be the most remarkable turnaround this contest has ever seen as far as I know. Of those who picked Stanford to take it all, RaderD15 is in a strong position because he also had Virginia picked correctly and if he hits again with the Wahoos, he will separate himself from Cachundo and Bluesky. And, of course, if Virginia goes all the way, Hykos is the big winner. Below are the current scores. Human Division 1. 133 hykos1045 2. 132 Kolabear 3. 125 Bluesky222 4. 123 cachundo 4. 123 keithscarlett 4. 123 raiderD15 7. 122 Justdoit 8. 121 BruBru 9. 118 Got Jukes? 9. 118 uncchamps2012 11. 117 Bosco 12. 116 GrassTopper 12. 116 Ingoldsby 14. 112 HeadSpun 15. 111 Hooked003 15. 111 Tom81 16. 110 SoCalSun 17. 109 dmthomas49 17. 109 liesse00 19. 106 soccerhunter 20. 103 RtD! 21. 102 Glove Stinks 21. 102 Kickithard Machine Statistical Division 134 Iteration 5 URPI 134 Massey Machine 132 Adjusted RPI 130 Bennett Machine 130 NCAA Standards 129 Albyn Jones/Massey 104 Old Recruiting Scores (My apologies for being later that I would have liked with these postings. My family and I have been celebrating Thanksgiving with good friends who live in Canada and we were all at a lodge some 150 miles north of Montreal where there was no internet. I just got back home to North Carolina at midnight and saw the results!)
Attached are the last bracket updates before the finals. Green is a win, Yellow is a still pending result, Red is a loss. My hat's off to Bosco for that perfect 4 for 4 round!
I like the scoring format that soccerhunter is using but participating in a volleyball bracket at volleytalk made me wonder why I like it. Volleytalk doesn't weight games any differently from round to round. All matches are worth one point. I was going to comment over there on soccerhunter's scoring system but I wasn't sure what to say as to why I like it. I have a few preliminary thoughts (and I was thinking of tinkering with soccerhunter's system) but anyone else have any?
Just some comments: The scoring system tries to achieve a correct balance between long-range and sort range predictions. (Or in other words, just going with the favorites and playing it safe, or taking some calculated risks.) The player is rewarded for both getting the big picture right and predicting the final 4 and the winner, but also is significantly rewarded by getting a lot of the "tough to predict" games in the "upset rounds" (2nd and 3rd rounds). Soo.... the idea is to weight the final rounds enough such that those who correctly get the "basics" right will have a chance (ie Bosco, this year) but also arrange the numbers such that picking a lot of the "up set rounds" correctly will keep others in the hunt even if they have missed on one or two of the elite 8 or final four. In short, it's all about trying to get the balance correct and not making it a foregone conclusion after the third round (a half dozen players are the only ones who have a chance at winning) or, conversely, discounting the early rounds and making the only thing important getting the last three rounds essentially correct. If I were to make any changes in the scoring system, it might be to de-emphasize the last few rounds even more, but basically, I think that I'm pretty close to the correct balance with the present system.
It's all over but the shoutin' now... (as far as this contest is concerned, that is.) Hykos was the only one to pick Virginia and Florida State to win to night and consequently the only one to garner 16 points. Thus he has distanced himself from (what was last week) a tight pack to put himself into the winner's circle regardless of the championship game. Congratulations to Hykos, for his prescient picks! Out of the 23 contestants he is the only one to have picked the two finalists and it has paid off. And, Hykos, this victory gives you the right to lord it over the rest of us for one grand post! (But then we expect your natural intelligence and humility to return.) Bosco's crystal ball failed him in this round as his pick of TAMU over VA cost him 16 points versus Hykos. Had the Texans won he would have been poised to possibly tie Hykos for the accolades. But it looks like his meteor flamed out. The guaranteed second place finisher is kolabear. Like most everyone else, he chose UCLA last week and this cost him a shot at the top spot. (Especially if he could have brought himself to name an overall winner.) And congratulations to all 23 of the human entrants. This is a fun (but humbling) exercise! Thanks for giving it a shot. Maybe next year it will be YOU who correctly prognosticates the winners! (I believe that I remember that two or three years ago Kickithard narrowly missed ending up at the top of the heap, so his last place finish this year should somehow give us hope for next year??) As to the statistical/machine entrants, their contest is all over as all of them picked UCLA as the champion. As a group they out-performed the average human scores, but once again, the top humans out-did their average. (I'll comment on the recruiting-scored entry later...) The current scores are below. Although we know who the winers are at this time, I'll post the final scores late on Sunday evening. My best to all of you faithful women's college soccer fans! Human Division 149 hykos1045 140 Kolabear 131 keithscarlett 131 raiderD15 130 Justdoit 129 BruBru 126 Got Jukes? 125 Bluesky222 125 Bosco 124 Ingoldsby 123 cachundo 119 Hooked003 119 Tom81 118 uncchamps2012 117 liesse00 116 GrassTopper 114 soccerhunter 112 HeadSpun 111 RtD! 110 SoCalSun 109 dmthomas49 102 Glove Stinks 102 Kickithard Statistical Machine Division 142 Massey Machine 138 Bennett Machine 137 Albyn Jones/Massey 134 Iteration 5 URPI 132 Adjusted RPI 130 NCAA Standards 104 Old Recruiting Scores