Pomeroy Projections on Feb. 14

[MChambers]

Interesting to see how Pomeroy projects the ACC regular season playing out:

Duke 13-3
WF 11-5
MD 11-5
VT 10-6
FSU 9-7
Clemson 8-8
GT 8-8
VA 7-9
Miami 5-11
BC 5-11
UNC 5-11
NCSU 4-12

(Sorry, don't do tables.) It's very interesting that Clemson and Georgia Tech are projected to finish 8-8, which is below what I would have expected before the season. Also interesting that the teams are not projected to be as bunched in the middle with 8-8 and 7-9 records as I would have thought.

Finally, seven teams are in the top 25 in defense, but only two on offense. UNC has the most balanced team, 63 on offense and 62 on defense.

http://kenpom.com/conf.php?c=ACC

[uncwdevil]

Interesting to see how Pomeroy projects the ACC regular season playing out:

Duke 13-3
WF 11-5
MD 11-5
VT 10-6
FSU 9-7
Clemson 8-8
GT 8-8
VA 7-9
Miami 5-11
BC 5-11
UNC 5-11
NCSU 4-12

(Sorry, don't do tables.) It's very interesting that Clemson and Georgia Tech are projected to finish 8-8, which is below what I would have expected before the season. Also interesting that the teams are not projected to be as bunched in the middle with 8-8 and 7-9 records as I would have thought.

Finally, seven teams are in the top 25 in defense, but only two on offense. UNC has the most balanced team, 63 on offense and 62 on defense.
http://kenpom.com/conf.php?c=ACC

That's like how I tell my friends that I have the most balanced golf game, since I slice every drive into the woods and 3-putt every green.

[MChambers]

That's like how I tell my friends that I have the most balanced golf game, since I slice every drive into the woods and 3-putt every green.

I couldn't find an emoticon for tongue in cheek.

I guess it means that Ol' Roy is doing an equally good job on both ends of the floor.

[JStuart]

Does this mean UNC could play in the Les Robinson Invitational game Thursday in the ACC Tournament?

[ChrisP]

Does this mean UNC could play in the Les Robinson Invitational game Thursday in the ACC Tournament?

I've seen the "Les Robinson" game mentioned several times recently with regard to UNC's wonderfully awful season and maybe people are just having a bit of fun at the 'holes expense but...there is no longer a "play in game". Since the powers that be decided, in their infinite wisdom, that the conference should be expanded to 12 teams, what we get is the bottom 8 teams playing on the first day (Thurs) of the ACC Tourney and the top 4 teams getting a bye until the real first round on Friday. I actually forgot that myself so...I looked it up on the ACC's website.

Bottom line, UNC's having to play that first day won't be nearly as embarrassing as it might if they were one of only two teams competing. Sad, but true

[roywhite]

Does this mean UNC could play in the Les Robinson Invitational game Thursday in the ACC Tournament?

Well, there are 4 games on ACC Tournament Thursday with the 12-team league.

As of now, UNC could play in one of the evening games, #7 vs #10, or #6 vs #11. Actually, it's too early to tell where they will end up in the standings, but playing on Thursday in the ACC Tournament is likely.

[jipops]

Anyone have an idea to the historical accuracies of the pomeroy projections?

[CDu]

Does this mean UNC could play in the Les Robinson Invitational game Thursday in the ACC Tournament?

Well, only the top-4 teams avoid playing on Thursday. So the odds are pretty good that they'll be playing that day. And there's no longer a game between the last two teams in the conference.

As for the likelihood of UNC playing on Thursday, it's darn near 100% at this point. The most ACC wins that UNC can get is 9. Five teams already have 6 ACC wins, and three teams already have 7 wins. It's very unlikely that UNC will win out, and it's fairly likely that at least three other teams will reach 9-7 along with Duke.

[JasonEvans]

Anyone have an idea to the historical accuracies of the pomeroy projections?

Well, seeing as the projections get adjusted throughout the season (based on the most up-to-date game results), I would imagine the projections are not all that reliable in December but are really, really accurate around the first week of March.

--Jason "He has had us at 14-2 or 13-3 pretty much since early January" Evans

[airowe]

Who do you guys think our last loss will be, if Pomeroy is correct?

I've been saying all along we're going to lose to Maryland in College Park...

[uncwdevil]

Anyone have an idea to the historical accuracies of the pomeroy projections?

The team that has won the title since 2004 has always been in the top 5 in offense and top 15 in defense at the end of the year (post-tourney).

Duke is currently #1 in offense and #12 in defense. Kansas is currently the only other team in the top 5 in offense and top 15 in defense. They are #2 in offense and #3 in defense.

It would be interesting to see what prior year ratings looked like at the beginning of March before the championship teams went on their runs.

[mike88]

I have become a real believer in Pomeroy's statistical methods. It was hard to see how Duke was going to finish 13-3 right after the NCSU loss, but we are in a position to get there now. We have to take care of business this week by beating Miami on the road and VaTech at home; then (after Tulsa) we have UVa on the road, Maryland on the road and UNC at home. I don't see UNC beating us, but I could see us losing to either UVa (playing for the NCAA bid) or Maryland (playing for revenge and seeding).

Some other observations from Pomeroy: We have the highest offensive rating in the country, so I am surprised about how many posters are calling for changes in our offense- I would say we should just keep on doing exactly what we have done this year. It is interesting, too, that our opponents have one of the highest collective defensive ratings, which may account for why our offense has looked disjointed at times.

To increase our chances of winning out (!), the most important thing for Duke to do is continue to play defense at a high level. In our losses (at least G-town, Wisc, and NCSU), it was our inability to stop the other team that made the difference. Our defensive rating (now 13th) has improved over the past few games. We are not really too far behind Kansas for the overall #1 rating per Pomeroy. Of course, predictive indices are not perfect, and while they improve with more data, they are also better at predicting aggregate events (like the rest of the ACC regular season) rather than single games (will we beat team x).

[MChambers]

The team that has won the title since 2004 has always been in the top 5 in offense and top 15 in defense at the end of the year (post-tourney).

Duke is currently #1 in offense and #12 in defense. Kansas is currently the only other team in the top 5 in offense and top 15 in defense. They are #2 in offense and #3 in defense.

It would be interesting to see what prior year ratings looked like at the beginning of March before the championship teams went on their runs.

Although Duke's offense has been rated highly all year, the defense has fluctuated quite a bit, and was in the 20s after the Georgetown game. It's interesting how quickly it can change.

[CDu]

The team that has won the title since 2004 has always been in the top 5 in offense and top 15 in defense at the end of the year (post-tourney).

Duke is currently #1 in offense and #12 in defense. Kansas is currently the only other team in the top 5 in offense and top 15 in defense. They are #2 in offense and #3 in defense.

It would be interesting to see what prior year ratings looked like at the beginning of March before the championship teams went on their runs.

Well, except for last year, in which UNC ended up #16 in defense. It's also important to note that we've finished in the top-15 in both offense and defense in all but two of those years.

But yeah, it's been relatively reflective of tournament success (as predictive as can be given the single-elimination format).

[MChambers]

This year, it seems that it is Duke's defense that is more key than the offense, in that Duke's 4 losses have come in its 5 worst defensive efficiency games. (The other bad defensive game was St. Johns.)

Duke has survived some bad offensive games (UNC, UConn, ASU, @Clemson).

If Brian's and Mason's offensive games continue to improve, maybe we'll have a better chance of winning a game with relatively weak defense, but I'd prefer not to take that chance. I'd like to see the defense continue to improve.

[sagegrouse]

Interesting to see how Pomeroy projects the ACC regular season playing out:

Duke          13-3
WF            11-5
MD            11-5
VT            10-6
FSU            9-7
Clemson        8-8
GT             8-8
VA             7-9
Miami          5-11
BC             5-11
UNC            5-11
NCSU           4-12

(Sorry, don't do tables.)

But the Grouse does...

[COYS]

I have become a real believer in Pomeroy's statistical methods. It was hard to see how Duke was going to finish 13-3 right after the NCSU loss, but we are in a position to get there now. We have to take care of business this week by beating Miami on the road and VaTech at home; then (after Tulsa) we have UVa on the road, Maryland on the road and UNC at home. I don't see UNC beating us, but I could see us losing to either UVa (playing for the NCAA bid) or Maryland (playing for revenge and seeding).

Some other observations from Pomeroy: We have the highest offensive rating in the country, so I am surprised about how many posters are calling for changes in our offense- I would say we should just keep on doing exactly what we have done this year. It is interesting, too, that our opponents have one of the highest collective defensive ratings, which may account for why our offense has looked disjointed at times.

To increase our chances of winning out (!), the most important thing for Duke to do is continue to play defense at a high level. In our losses (at least G-town, Wisc, and NCSU), it was our inability to stop the other team that made the difference. Our defensive rating (now 13th) has improved over the past few games. We are not really too far behind Kansas for the overall #1 rating per Pomeroy. Of course, predictive indices are not perfect, and while they improve with more data, they are also better at predicting aggregate events (like the rest of the ACC regular season) rather than single games (will we beat team x).

One thing I would be very interested to see from Pomeroy is conference efficiency statistics. I know you can go through the game plan and manually calculate each team's stats for conference games, but I think it would help give us a sense of which teams are playing well in conference and which teams have really padded their stats with weak out of conference games. Of course, it would be even better if you could manually exclude games. Last year, Memphis was number one going into the tourney, if I remember, but that was largely due to their ridiculous defensive efficiency rating which was the result of a weaker than normal C USA schedule. Home and road sortable stats would also be awesome. But hey, I'll take them as is for free any day.

[pfrduke]

One thing I would be very interested to see from Pomeroy is conference efficiency statistics. I know you can go through the game plan and manually calculate each team's stats for conference games, but I think it would help give us a sense of which teams are playing well in conference and which teams have really padded their stats with weak out of conference games. Of course, it would be even better if you could manually exclude games. Last year, Memphis was number one going into the tourney, if I remember, but that was largely due to their ridiculous defensive efficiency rating which was the result of a weaker than normal C USA schedule. Home and road sortable stats would also be awesome. But hey, I'll take them as is for free any day.

Here's the season to date in the ACC:

		ORTG
1	Duke            110.55
2	Maryland        107.73
3	Wake Forest	101.55
4	Virginia        101.35
5	Boston College	100.72
6	Virginia Tech	99.80
7	Georgia Tech	99.01
8	Miami           98.60
9	North Carolina	98.25
10	NC State        97.44
11	Florida State	97.30
12	Clemson	        96.29
		
		DRTG
1	Duke          	94.04
2	Virginia Tech	94.40
3	Clemson	        94.83
4	Virginia       	95.54
5	Maryland     	96.91
6	Wake Forest	98.22
7	Georgia Tech	100.16
8	Florida State	100.67
9	North Carolina	104.69
10	NC State     	108.67
11	Miami         	109.40
12	Boston College	109.74
		
		E Margin
1	Duke          	16.51
2	Maryland     	10.81
3	Virginia       	5.81
4	Virginia Tech	5.39
5	Wake Forest	3.34
6	Clemson      	1.46
7	Georgia Tech	-1.15
8	Florida State	-3.37
9	North Carolina	-6.43
10	Boston College	-9.02
11	Miami         	-10.81
12	NC State    	-11.23

[pfrduke]

One thing I would be very interested to see from Pomeroy is conference efficiency statistics. I know you can go through the game plan and manually calculate each team's stats for conference games, but I think it would help give us a sense of which teams are playing well in conference and which teams have really padded their stats with weak out of conference games. Of course, it would be even better if you could manually exclude games. Last year, Memphis was number one going into the tourney, if I remember, but that was largely due to their ridiculous defensive efficiency rating which was the result of a weaker than normal C USA schedule. Home and road sortable stats would also be awesome. But hey, I'll take them as is for free any day.

Also, if you if you check Basketball Prospectus every Tuesday for the "Tuesday Truths," they go through the conference-only performances for the top 10 or so conferences.

[COYS]

Also, if you if you check Basketball Prospectus every Tuesday for the "Tuesday Truths," they go through the conference-only performances for the top 10 or so conferences.

Thank you very much. I check the blog irregularly and had seen that a couple of times but did not realize that it is a weekly update.