[Lax] NCAA Prospects

[zg88]

Heh-heh... Matt and I are such good citizens -- we were practically tripping over each other to answer the eLF Faithful's questions!  :-D

[kingpin248]

A word on the Swami: his credibility has to take a hit.  You'll note he rigged his competition so his "hated rival," the Ivyman, got just enough of a "late pick penalty" to put Venus Lee on top.  Meanwhile, where did the Swami himself finish? Dead last.

[zg88]

Yeah, dead last on his own website -- that's pretty funny!  He is now relegated to comic relief.  Frankly, I wonder whether either Venus or Bunny actually exist (blow-up dolls don't count!).

[jtwcornell91]

Well, the games have all been played, and I just crunched the numbers.  See http://slack.net/~whelan/lax/rpi.shtml

We stay #6 in the RPI, but drop to #10 (ouch!) in Opp Pct which is apparently the NCAA's schedule strength measure.  (How dare we play Hobart!)  Unfortunately, staying at #6 is actually a slip: we passed UMass (who are in anyway on an AQ), but idle Hofstra somehow managed to pass us!  So if they went by RPI alone, we would be the last at-large bid: Hopkins, Syracuse, Virginia, Georgetown, Hofstra, and Cornell.

Behind us in the RPI but ahead in strength of schedule are Duke, Brown, Yale and UNC.  We have one win against 1-5 (Syracuse) one against 6-10 (Yale), and one against 11-15 (Penn).  Duke has one win against 1-5 (Virginia), one against 6-10 (Yale), and two against 11-15 (both Maryland).  Brown has two wins against 6-10 (us and Yale) and nothing else.  Yale has one win against 1-5 (Princeton), one against 6-10 (UMass) and one against 11-15 (Penn).  UNC has one win against one win against 6-10 (Duke) and nothing else.  I think Duke, at #9 in the RPI and #2 in schedule strength, is the team we have to worry about squeezing us out. (Of course, their 7-6 record might count against them.)

I have to say, this really makes you appreciate hockey, where we would know the definitive answer already.

Incidentally, we're comfortably ahead of Duke in both KRACH and PWR.

[kingpin248]

It doesn't affect the RPI, but Hobart is a TUC for PWR purposes.  The Patriot League announced that since they had a higher RPI than Army or Bucknell (the other two tri-champions), the Statesmen get the Patriot automatic bid despite being 6-7.

[jtwcornell91]

So, I was surprised to see that Hofstra passed us in the RPI even though we won and they were idle.  So I ran the numbers including all games except Cornell-Hobart, including all games, and including all games through Friday.  It turns out that the Hobart game did not effect our RPI at all (to four significant figures).  Our winning percentage went up by .0192 but our opponents' winning percentage went down by .0093 and our opponents' opponents' winning percentage went down by .0008.  With the 25/50/25 weighting, we end up at .6380 both with and without that game.  Meanwhile, the other results on Friday brought our schedule strength down and Hofstra's up.  (E.g., Hopkins beating Georgetown hurt the Hoya's winning percentage.)

Sadly, the other selection criteria don't actually correct for this problem with RPI (that beating a weak team does nothing for you, and may hurt a little), but make it worse.  Our strength of schedule was of course hurt by the Hobart game (since that criterion doesn't look at the result of the game) and since they're not in the top 15, they don't factor in the "vs top teams" criterion.  So I think this is a process which really can "overvalue" a team's strength of schedule.

BTW, our Bradley-Terry RRWP was .8612 without the Hobart win and .8615 with it.  Saturday's other results hurt our strength of schedule enough that our RRWP had dropped from .8615 to .8612, not including the Hobart game.  (So our RRWP was unchanged from Friday to Saturday.)  Meanwhile, Hofstra went from .8729 Friday to .8739 Saturday, with or without the Cornell-Hobart game being included, reflecting the stronger view of their strength of schedule.

Reference:
RPI as of Friday: http://slack.net/~whelan/lax/rpifri.shtml
RPI including all but Cornell-Hobart: http://slack.net/~whelan/lax/rpinoHb.shtml
Final RPI: http://slack.net/~whelan/lax/rpi.shtml
KRACH as of Friday: http://slack.net/~whelan/lax/krachfri.shtml
KRACH including all but Cornell-Hobart: http://slack.net/~whelan/lax/krachnoHb.shtml
Final KRACH: http://slack.net/~whelan/lax/krach.shtml

[jtwcornell91]

Thanks; I re-ran the script with Hobart as a TUC and modified the definition on the PWR page.  I wasn't going to bother with the AQ-as-TUC business, but I saw in the NCAA Lax Champs Handbook that the literal definition as "Team Under Consideration" applies, since only teams with .500 or better records are allowed to get at-large bids.

[Ben Doyle 03]

Howdy folks, here's the probable breakdown of the tournament selection (so if you can't watch it. . .here's the deal), the first six are already in:

Automatic Qualifiers:

AQ Ivy - Princeton (8-4) - #4
AQ ECAC - UMass (11-3) - #7
AQ MAAC - Manhatten (11-5) - #42
AQ Patriot - Hobart (6-7) - #23
AQ GWLL - Fairfield (7-5) - #20
AQ America East - Stony Brook (10-6) - #30

Interesting to note that these teams have a combined 30 losses and their combined ranking is a huge 126!!!!

The at-large bids will most likely go to the following:

At- large #1 - Syracuse (12-2) - #1
At- large #2 - Johns Hopkins (11-1) - #2
At- large #3 - Virginia (10-3) - #3
At- large #4 - Georgetown (11-2) - #5
At- large #5 - Hofstra (11-3) - #6
At- large #6 - Maryland (9-4) - #8
                          Cornell (10-3) - #9
                          Duke (7-6) - #10
                          Loyola (9-4) - #11
 
Even if you take Duke (and their 6 losses) over the other three teams (Maryland, Cornell, and Loyola) the at-large field still only has 17 losses and a combined ranking on the high end of 28.

This should clearly illustrate the need for the 2003 expansion of the tournament field. If AQ's are going to conferences like the MAAC & America East resulting in two teams in the top 10 not making the field. . .we have a problem. A 16 team NCAA lax tournament will do justice to all.
Sorry. . .I'm getting off my soap box now ;-)


Rankings were taken from www.Laxpower.com
 

[jtwcornell91]

I got tired of trying to read my tables of selection criteria to see how many "quality wins" Cornell, Duke, and Loyola had, so I did a new version of the lacrosse RPI page which includes records against the top 5, 10, and 15 (inclusive).

http://slack.net/~whelan/lax/newrpi.shtml

Also, the team names are now links to a list of the team's opponents in RPI order, with 1-5, 6-10, and 11-15 teams highlighted in different colors.  For example:

http://slack.net/~whelan/lax/current/rpi.Cr.shtml

Enjoy!

[jtwcornell91]

How do you get Maryland in?  They're #13 in the RPI and 2-4 against the top 15, with the only wins coming against #10 Yale and #12 UNC.

[Ben Doyle 03]

Thanks JTW, I was just showing the teams that could potentially bump us from the tournament, the key word being could. I hope, as I'm sure anyone reading this hopes we get the nod (as we should). I was more intent on showing the teams that should NOT be in the tournament through the AQ process (Manhattan, Stony Brook, and Hobart). Thanks for the input with regard to the RPI, it definitely shows we have the best shot for the last at-large. Enjoy the sunshine. . .well at least it's sunny in Ithaca:-)

[jtwcornell91]

Teams in the running for the last two at-large berths.  I'm still not sure if the opponents' percentage column is the right one for strength of schedule, nor exactly what the committee is supposed to do with it.

                                    ___vs_Top__
               ___RPI___  _Opp_Pct_  5   10  15 ________Top_15_wins_________
Hofstra (11-3)  #6 .6394  #12 .7857 0-2 2-2 3-3 #8 UMass, #9 Duke, #12 UNC
Cornell (10-3)  #7 .6380  #10 .6115 1-2 2-2 3-3 #2 Syr, #10 Yale, #15 Penn
Duke (7-6)      #9 .6299   #2 .6923 1-3 2-4 4-6 #3 UVa, #10 Yale, #13 UMd (x2)
Yale (9-4)     #10 .6219   #7 .6155 1-0 2-2 3-4 #5 Prin, #8 UMass, #15 Penn
Loyola (9-4)   #11 .6110  #14 .5844 0-3 2-3 3-3 #6 Hof, #9 Duke, #14 Brown
UNC (8-5)      #12 .6060   #8 .6139 0-3 1-4 1-5 #9 Duke
Maryland (9-4) #13 .6032  #16 .5652 0-2 1-4 2-4 #10 Yale, #12 UNC

For all people say about Cornell only having the one quality win against Syracuse, no one else in this group has more than one win against the top 5 or two against the top 10.  Duke has one more top 15 win, but they won 4 out of 10 and we won 3 out of 6.  Yale is the only team with a better winning percentage against the top 5, but that was one game against Princeton.  There are teams with tougher schedules, but all it seems to have got them is more losses.

If Duke gets in despite our .0081 RPI advantage over them, you'd have to think we'd get the nod ahead of Hofstra despite their .0014 RPI advantage over us, since we seem to have the edge over Hofstra in the strength of schedule and quality wins departments.  Conversely, if Hofstra's in on the strength of their RPI, we should be too.  It's hard to think of a single perspective from which two of the teams above are better than us.

[neil shapiro]

John,

two questions about the rankings:

Is it possible to set it up so that a team that is 0-0 against a group (eg. top 5) is not considered equivalently ranked to a team that is 1-1 against that group?

Is it possible for teams with identical records to have identical rankings, rather than sequential rankings sorted alphabetically?  What I am trying to ask is for Hopkins and UVA to both be ranked third against the top 5, with identical 2-1 records, rather than Hopkins 3rd and UVA 4th because Hopkins is first alphabetically.

[jtwcornell91]

These things are possible (in fact I'm planning to write a pseudo-sorting routine for next season's hockey rankings which allows for ties in the rankings), but I haven't taken the time to implement them.  You should pay attention to the records rather than the rankings for the top 5/10/15 columns.

[jtwcornell91]

Cornell and Duke are in.  Hofstra and Loyola are out.  We play Stony Brook at Brown on Saturday.