So Who Had The Toughest Mirror Games?

[RockawayHoya]

With an unbalanced conference schedule, I was curious to see whether or not the imbalance played a pivotal role in where the teams currently are in the standings, given how so many teams are so closely bunched up together.

Here are the aggregate W-L records of mirror opponents, EXCLUDING record against that particular team (as to avoid artificially deflating opponents' W-L records for good teams and vice versa). Note that some teams have played 17 games, while others have played 16, which accounts for the differences in total W-L record. The top 11 teams (the ones that matter) are listed in order of the current standings:

-Pitt (Nova, WVU, USF): 18-26
-ND (UConn, Marquette, SJU): 27-16
-Lville (WVU, UConn, Prov): 20-24
-SJU (Cinci, GU, ND): 29-15
-Syracuse (GU, SHU, Nova): 21-23
-GU (Syracuse, Cinci, SJU): 28-16
-Marquette (ND, SHU, UConn): 25-19
-UConn (Lville, Marquette, ND): 29-15
-WVU (Pitt, Lville, Depaul): 22-19
-Cinci (SJU, GU, Depaul): 21-23
-Nova (Pitt, Syracuse, Rutgers): 25-19

Teams with the toughest schedules: UConn, SJU, GU
Teams with the easiest schedules: Pitt, Lville, Cuse, Cinci

[hoyabinx]

Nice work putting that together! Interesting to see it laid out like that.

[Coast2CoastHoya]

This is something I wonder about every season. Thanks Rockaway!

[superan]

So Syracuse sucks. Got it.

[FrazierFanatic]

Of course our schedule will be one game easier, and Cincy's one game harder, after Saturday.

[Jack]

Mar 1, 2011 15:52:18 GMT -5 FrazierFanatic said:

Of course our schedule will be one game easier, and Cincy's one game harder, after Saturday.

Of course that's not actually true under Rockaway's methodology:

RockawayHoya said:

Here are the aggregate W-L records of mirror opponents, EXCLUDING record against that particular team (as to avoid artificially deflating opponents' W-L records for good teams and vice versa).

[ScreamingHoya]

Rockaway- can you explain what this is a little bit more? I feel like I should be understanding it but I'm just not getting it.

[FrazierFanatic]

Mar 1, 2011 17:12:53 GMT -5 Jack said:

Mar 1, 2011 15:52:18 GMT -5 FrazierFanatic said:

Of course our schedule will be one game easier, and Cincy's one game harder, after Saturday.

Of course that's not actually true under Rockaway's methodology:

RockawayHoya said:

Here are the aggregate W-L records of mirror opponents, EXCLUDING record against that particular team (as to avoid artificially deflating opponents' W-L records for good teams and vice versa).

Yeah i kind of missed that part.

[RockawayHoya]

Sorry guys, let me put together something that will make it easier to understand. Will also give me a chance to re-check my numbers; I did this fairly quick earlier today when I was bored in class. Give me a few minutes...

[RockawayHoya]

OK, here's my shot at a better explanation.


-Pitt (Nova, WVU, USF): 9-8, 9-7, 3-13 = 21-28 -> (5-0) -> 21-23
-ND (UConn, Marquette, SJU): 9-7, 9-7, 11-5 = 29-19 -> (3-2) -> 27-16
-Lville (WVU, UConn, Prov): 9-7, 9-7, 3-13 = 21-27 -> (3-1) -> 20-24
-SJU (Cinci, GU, ND): 9-7, 10-7, 13-4 = 32-18 -> (3-3) -> 29-15
-Syracuse (GU, SHU, Nova): 10-7, 5-11, 9-8 = 24-26 -> (3-3) -> 21-23
-GU (Syracuse, Cinci, SJU): 11-6, 9-7, 11-5 = 31-18 -> (2-3) -> 28-16
-Marquette (ND, SHU, UConn): 13-4, 5-11, 9-7 = 27-22 -> (3-2) -> 25-19
-UConn (Lville, Marquette, ND): 11-5, 9-7, 13-4 = 33-16 -> (1-4) -> 29-15
-WVU (Pitt, Lville, Depaul): 13-3, 11-5, 1-15 = 25-23 -> (2-3) -> 22-21
-Cinci (SJU, GU, Depaul): 11-5, 10-7, 1-15 = 22-27 -> (4-1) -> 21-23
-Nova (Pitt, Syracuse, Rutgers): 13-3, 11-6, 4-12 = 28-21 -> (2-3) -> 25-19

I screwed up the numbers for Pitt and WVU, but they really don't change the overall picture much (except Pitt's schedule was not as easy as previously thought).

Here's what I did: I totaled up the BE wins and losses for each team's 3 mirror opponents. Then, I took out the records for each teams against their mirror opponents. So for example, Pitt has gone 5-0 so far this year against their three mirror opponents, so I subtracted 5 losses from their opponents' aggregate W-L total in order to show how Pitt's mirror opponents did against everybody not named Pitt.

This much still remains: the easiest schedules were played by Lville, Pitt, Cuse and Cinci (the only 4 contenders to have a sub .500 record using this methodology) and the hardest schedules were played by SJU, UConn, GU and ND.

[gujake]

Mar 1, 2011 12:39:21 GMT -5 RockawayHoya said:

-Pitt (Nova, WVU, USF): 18-26
-ND (UConn, Marquette, SJU): 27-16
-Lville (WVU, UConn, Prov): 20-24
-SJU (Cinci, GU, ND): 29-15
-Syracuse (GU, SHU, Nova): 21-23
-GU (Syracuse, Cinci, SJU): 28-16
-Marquette (ND, SHU, UConn): 25-19
-UConn (Lville, Marquette, ND): 29-15
-WVU (Pitt, Lville, Depaul): 22-19
-Cinci (SJU, GU, Depaul): 21-23
-Nova (Pitt, Syracuse, Rutgers): 25-19

Very nice Rockaway.

I think we can take this one step further and do a very crude analysis to give us a rough sense of how many wins or losses an average Big East team would expect to get in the mirror games that each team was given.

So, for example, Pitt was given Nova/WV/USF, which has an average winning percentage of 41% with your numbers above (18/44). So an average Big East team would expect to win 59%, or .59*3 = 1.77 games in the 3 mirror games against those teams.

Expanding that out to every set we get:

-Pitt (Nova, WVU, USF): 1.77 wins
-ND (UConn, Marquette, SJU): 1.12 wins
-Lville (WVU, UConn, Prov): 1.64 wins
-SJU (Cinci, GU, ND): 1.02 wins
-Syracuse (GU, SHU, Nova): 1.57 wins
-GU (Syracuse, Cinci, SJU): 1.09 wins
-Marquette (ND, SHU, UConn): 1.30 wins
-UConn (Lville, Marquette, ND): 1.02 wins
-WVU (Pitt, Lville, Depaul): 1.39 wins
-Cinci (SJU, GU, Depaul): 1.57 wins
-Nova (Pitt, Syracuse, Rutgers): 1.30 wins

Now, there are a whole bunch of statistical problems with this (I can think of at least 4) – but it is just meant to give a very rough idea of what it means to get good or bad teams for your mirror games.

So, in essence, an average Big East team might expect to win .75 more games if they got Pitt’s schedule compared to St. John’s schedule this year.

[RockawayHoya]

HoyaJake, you might want to update your numbers. I caught a couple of goofs on my part re: Pitt and WVU's numbers.

Interesting to also see your analysis as well. The difference between one win for nearly half the league right now could be the difference in several seeds for the BET.

[gujake]

Mar 1, 2011 19:18:36 GMT -5 RockawayHoya said:

HoyaJake, you might want to update your numbers. I caught a couple of goofs on my part re: Pitt and WVU's numbers.

Interesting to also see your analysis as well. The difference between one win for nearly half the league right now could be the difference in several seeds for the BET.

Will try to throw together a new post before I go to bed. I'll attempt to account for a couple of the statistical problems in that post as well.

[bmartin]

This is not that complicated. If you drew a home and home with Depaul or USF, you had a very good chance for a sweep. There was a fair chance of sweeping Providence, Rutgers, or Seton Hall. But home and home between the top 11 teams were most likely to end in a split. The teams that drew three other top teams for their home and homes had harder schedules than teams that drew one of the bottom five teams.

I can only see two completed sweeps so far of a top team, Pitt over WVU and Louisville over UConn. All the other home and homes between top teams were split or have not been completed.

Also, though, Seton Hall split with Syracuse, Rutgers split with Villanova, and Providence won the 1st game vs. Louisville.

[gujake]

Here are new numbers. Besides the incorrect records, there were two major problems with the previous #s: (1) they did not account for remaining games and (2) the records for each team was skewed based on their own mirror games. Maybe I'll put up a long methodology post tomorrow, but in order to keep from cluttering the board I'll just summarize how I dealt with these problems:
1. team records were adjusted for remaining games using KenPom projected results
2. team records were adjusted as if each team played a 15-game schedule (1 game vs each opponent) by averaging results against teams played twice

Adjusted Records:
Pitt: 11.85-3.15
ND: 11.7-3.3
Lville: 10.15-4.85
SJ: 11–4
Cuse: 10.5-4.5
GU: 9.2-5.8
Marq: 8.5-6.5
Uconn: 9.1-5.9
WV: 9-6
Cincy: 7.7-7.3
Nova: 8.1-6.9
SH: 4.7-10.3
Rutg: 3.35-11.65
Prov: 2.4-12.6
USF: 1.65-13.35
Dep: 1.35-13.65

Expected # of wins for an average Big East team given a set of mirror opponents:
Pitt (Nova, WVU, USF): 1.75 wins
ND (UConn, Marquette, SJU): 1.09 wins
Lville (WVU, UConn, Prov): 1.63 wins
SJU (Cinci, GU, ND): 1.09 wins
Syracuse (GU, SHU, Nova): 1.53 wins
GU (Syracuse, Cinci, SJU): 1.05 wins
Marquette (ND, SHU, UConn): 1.30 wins
UConn (Lville, Marquette, ND): 0.98 wins
WVU (Pitt, Lville, Depaul): 1.44 wins
Cinci (SJU, GU, Depaul): 1.56 wins
Nova (Pitt, Syracuse, Rutgers): 1.29 wins

[joey0403p]

This works for the toughest mirror games but not the toughest schedule . We d have to factor in the home vs away games into it. Just my two cents