Yup. Example - i thought the ad review last cycle was absolutely brilliant.  Deep, simple, and eminently relevant & practical.  Wish more of the campaign "stategists" and "communications" directors paid attention to it cause i think it would have helped some of them...

Amen to that.  The "on the road" reporting was really interesting (and informative).  It fits right in with the "direct action," "close to the metal" attitude that i love about MyDD.

I totally agree with the your take on what is cool about myDD - the focus on concrete action within our political environment.  That said, i still see "community" in it.  Not necessarily a social-circle, but still a collection of like-minded individuals who share a common view on how citizens should relate to their government (and each other)

Admittedly i was drawn in to MyDD my a post of Matt's and then of Chris's.  But that was just the beginning (seems like forever ago.)  I stayed because of the across-the board thoughtful writing and perspective on the site and in the MyDD community.  That hasn't changed.  
I'm sure i'll check out what Matt and Chris are doing - it'll be great, i'm sure.  But I'll still keep visiting MyDD on a daily basis for my dose of perspective and analysis of the root of Democracy and Democratic-cy (or however ur supposed to spell it)

Dan

Now if i could just get my hands on the 3DMean on a district by district basis ;-)

This probably goes without saying, but we should expect a slew of rebublican retirements as well...

And for anyone still following this thread, here's some killer analysis that someone put together on districts that are "vulnerable."

http://www.fairvote.org/media/research/m onopoly/2006/monopoly_politics_2006_1115 06.xls

This analysis is pre-2006 election, so it's interesting to see how the projections went.  But the underlying data is still relevant for 2008...

You're right on a couple of counts.  Mid-term redistricting is probably pretty hard if you don't control all of the state-level bodies (of course it depends on the individual state rules i guess.)  And i didn't mean to say that democrats control ALL those elements in those states - just enough that given the inequities and current political leanings of the states that it might be possible (and worthwhile).
Also on IL, you're right that i'm missing a district (IL-04) - somehow that's missing in my dataset - (anyone want to tell me it?)  That said, looking at the raw data, that district has to have a PVI >+27D... But my main point still holds: If you look at IL as a whole, you'll see it's rediculously packed:

IL19-8R
IL15-6R
IL08-5R
IL13-5R
IL14-5R
IL18-5R
IL10-4R
IL16-4R
IL06-3R
IL11-1R
IL12-5D
IL17-5D
IL03-10D
IL05-18D
IL09-20D
IL04-27D
IL01-35D
IL02-35D
IL07-35D

If you look at state with a similar PVI margin in the GOP direction (like IN, GA, TN) you'll see a 2-1 margin in Rep-to-Dem leanings in seats...

So all of the states i mentioned need to be redistricted to make them less "packed."  Of course, FL is also on that list but my point there was that the state level is so dominated by the GOP it's probably unrealistic to expect anything remotely like fair play...

all good points.  And taken together they add up to a lot of negatives for such a move.  And all for what?  NM's 3 congressional seats - 2 of which are already Dem leaning (by PVI).  In the TX move, there were a whole lot of districts - if the Dems want to take the heat for such a move, they should at least get a lot out of it... Try OH (6D/12R by PVI), MI (5D/10R by PVI) or IL (8D/10R by PVI).  All of those states have a positive Democratic statewide margin, Democratic leadership at the state level, lots of seats and an unreasonable small (compared to other states) proportion of Democratic leaning districts.  Of course, FL is the most in need of correction (8D/17R by PVI in a state which is evenly split in popular vote) - but there's no way that'll ever get passed there...

This is an excellent point.  Not only it good for progressives, but it's good for the Democratic party.  In the races where progressives beat out DLC candidates, the progressives acquitted themselves pretty well.

I think yodafone captures the key point...
When i saw matt's post, i was wondering too.  Since i've doing some number crunching on the house races (over at DownWithTyranny), i had a lot of the data already loaded into a database and spreadsheets, so i added in the SUSA info, and crunched away...
Of course, the question Matt asks requires some speculation, so what i did is far from rigorous, but it passes the smell test (IMO).
I set up a simple model that used, incumbant status, overal environment (i.e. how strong was the Dem wind blowing), PVI, scandal presence and quality of candidates.  (This last factor uses a sitting senators approval combined with the challengers quality...)  Using these factors and 14 key senate races, i used some math to figure out what relationship between the factors would match the results.  [Aside: this whole process is a lot more guess-work than science since a lot of factors require subjective assessment - e.g., was Burn's scandal(s) worse than Allen's, and if so, how much more...  I used what i thought were reasonable values and then tweaked them until it all matched up...  Also, a came up with the simpliest model i could think of, but its very possible that the relationships aren't much more complex - like the Democratic wind might might have been more magnified in some states.   So please take this all with a grain (grainery) of salt...]

Here's what the model showed in terms of general rules (this year):
    1. the overall environment accounted for a 3 point shift toward democrats across the board.
    2. if an incumbent was unpopular (either generally or because of scandal), if he was in a state that was highly partison FOR him (by PVI), he had a chance
    3. if an incumbent was unpopular (either generally or because of scandal), if he was in a state that was highly partisan AGAINST him, he was in trouble
    4. if an incumbent was unpopular (either generally or because of scandal), if he was in a state that was general partisan neutral, then both a strong challenger and a sustained period of unpopularity (or scandal) was in trouble.
    5. scandal seemed to be something that could be overcome once the event was over (burns, allen) but not while it was still in the news (DeWine).

To matt's specific examples/questions:

Santorum vs. Burns:
This is mostly explained by the fact that Montana is +9R and PA is +2.9D.  That's a 12 point gap - the same ballpark as the 17 point election difference.

Talent vs. DeWine:
The main difference here was the Scandals in Ohio vs. the personal unpopularity of talent.  The scandals stayed in the news and DeWine couldn't shake them.  The personal unpopularity didn't help, but wasn't as strong of a factor as the scandal.  Also, I rated McCaskil vs. Talent was a slightly better matchup for the Dems than Brown vs. DeWine.

Allen vs. Chafee
Biggest difference here was the PVI 2.4R vs. 14.5D.  

Nelson/Menendez vs. Talent/DeWine:
First factor was the +3 overal environment.  Then for Menendez add in 8.5D PVI (vs. 1.2R and 0.2R for Talent/Dewine respectively) and for Nelson add a horrible challenger who was both unpopular and scandal ridden (and FL's 1.3D PVI didn't hurt) and neither was in any danger.

Kyl vs. Talent/Dewine
Two factors - 1. kyle's negatives weren't as strong as Talent/Dewine and 2. Pederson wasn't the same quality of challenger as McCaskill and Brown (not meant to be a ding, he just didn't seem to have a strong a campagn, not the same level of funding support, not an inspiring speaker, weaker debater, etc.)

Cantwell
For some reason things really broke for her in the last 2 weeks (approval-wise).  Totally unclear as to why.  Even if they hadn't, the model shows she would have had no problem (PVI of 3.8D, etc.)

Lieberman
To screwy to include in the model ;-)

Conclusion
As Matt, yodafone and other point out, relative popularity is only one factor.  The prevailing wind certainly helps, but PVI (or some similar calculation) is hugely important - it make the difference between being able to bounce back and being done for good.  In terms of strategy: targetting PVI of better than 2.0 Republican is good.  High quality candidates make a big difference (in the model a 12 point swing between Kyle-Pederson and Nelson-Harris).  

That was fun...

To bring cold hard facts to the argument:

I've been analyzing the spending of the DCCC and their net effectiveness.  I got all of the 2006 contributions from committees (and individuals) and looked at what races the DCCC gave money to and what the net result was.  

The short version is that the "swing state" strategy that Rahm was espousing (and following) initially, would probably have let the GOP keep the house.  The fact that other groups (like ActBlue, etc.) were pushing the 50 state strategy brought a number of new races into play that allowed the DCCC to jump in during the last month or two.  But if noone had been pushing those races, it's hard to imagine that they would have been positioned for a win in the final months.

Here's a summary of what i mean:

- Before 9/1: Of the races that were won (or who are still in doubt), the DCCC spent more than 10K on only 11 of them.  (Well, 14 if you count the contributions to the primary where they backed a challenger to the winner ;-) )  If you lower the threshold to 5K, they only had 17 winners.  Compare that to a total of 27 races where they spent 5K or more.  (I mean, 'cmon, 5K is nothing...)

- Between 9/1 and 10/1: Of the races that were won (or who are still in doubt), the DCCC spent more than 5K on only 15 of them.  This is well into where they were getting polling data and seeing which races were tight.  During that period they spent $7.3M.  They spent >5K on only 21 races!  They spent huge cash on a lot of races that were lost (like KY4, OH15, PA6, VA2, IL6).  

- After 10/1:  they got a little better, they spent >20K on 18 of the winning races.  During that period they spent $14M (and still blew huge amounts in loosing efforts in PA6, KY4, OH1, OH15, VA2, CT4, OR5, CO4, etc., etc.)

So in the last month, the DCCC did jump into a few races that they hadn't supported before, but they still ignore a number of them that were key to the margin of victory.  And more to the point, the only reason that the races were close was because others (like ActBlue, blogosphere, etc. had kept the campagns financed, staffed, and motivated.)

Now i realize that i'm just giving a summary and there are ways to quibble with details (like is it fair to ding the DCCC for not having spent much money on TX-22 which seemed like a sure pickup...)  But there is a lot of interesting stuff in looking at the big picture like this - the numbers show an strategic investment pattern by the DCCC which was very out of sync with the eventual margin of victory.  Rahm should be counting his lucky stars and thanking the various progressive organizations for keeping so many candidates close.  And in particular he owes Dean a big apology for getting upset about Deans spending practices.  

I plan to post more details of the cruched numbers (and the datasets behind them).  But all of this begs the question of "why?."  Why were progressive, underfinanced candidates so successful (relatively) while the "centrist" candidates underperformed.  I have some theories, but i'm still putting the numbers together on that...

Nice post Chris.  I couldn't agree more.  rightousness and truth resonate - and generally (but not always) people will choose candidates who they believe radiate it.

To all,
Along those line, i've been going through FEC and other data to try and analyze the impact of dollars spent by the DCCC and other groups in support of House candidates this cycle.  In particularly i'm interested in how well the DCCC initial swing-district strategy played out (and what the congress would have looked like if it had been the only strategy in play...)  
With that said, i've seen a number of references on blogs to the DCCC dumping 3M+ into IL-6 (Tammy Duckworth) but for the life of me, i can't find anything in the data to support this.  (in fact, the fec only has a record of 1.6M spent on behalf of BOTH candidateds in the general election by all PACs...)  Anyone have a better understanding of the data to explain all this.

PS.  i've spent time writing scripts and getting the data organized into oracle tables, etc.  If anyone is interested in getting the data or the scripts to create the data from the various data sources, email me at dandrasin@yahoo.com

Matt,

Wonderful, thoughtful, and insightful post.  Of course some might quibble with minor details in the body, but i think the frame you created is right-on - and more importantly - a necessary perspective for continuing to increase the influence of progressives.

I think the general advise of dos and don'ts is right.  However, the specifics are going to be more difficult.  The question that this post begs is: how do "we" figure all this stuff out?  In short, we first need processes and frameworks to work within.  Hopefully enough people will recognize that there's still more infrastructure to be built to make it all happen.

Dan