Army LTC Gian Gentile gave an interview to World Politics Review, and says:
In my opinion, the two necessary and controlling reasons for lowering the violence in Baghdad in the second half of 2007 had little to do with the increased number of U.S. combat brigades practicing so-called new counterinsurgency tactics. Instead, the two necessary conditions were the decision by senior Americans to pay large amounts of money to our former enemies -- the non-al-Qaida Sunni insurgents -- to ally themselves with us to defeat al-Qaida and, as a by-product of this alliance, to stop killing Coalition Forces. That and Moqtada al-Sadr's decision to stand down attacks against American and coalition forces and against civilian Sunnis were the main causes. If those two conditions were not in place, I can not imagine how more American combat brigades using so called new methods would have lowered violence.
Gentile has emerged as a bit of a lightning rod for a piece he wrote for WPR a little while back where he said a change in tactics had little to do with the decline in Iraqi violence, that the factors he highlights in the interview, like paying Sunni insurgents to stop killing American troops, played a larger role. That part wasn't so controversial, as it kind of states the obvious, except to those who haven't spent much time in Iraq (Kagans). He went on to say that by focusing on counterinsurgency doctrine, the Army was losing its ability to fight large scale conventional battles and was fated to see a replay of Lebanon 2006 where Hezbollah schooled the IDF in hybrid warfare, that part was a little more controversial, among the counterinsurgency crowd in particular. That crowd asks: where are the big tank armies we will fight in the future?
I talked to Gentile recently, in part he resents the whole notion that counterinsurgency only arrived in Iraq in early 2007 with Gen. Petraeus, and that everything that came before was a waste of time and not in anyway associated with "true" COIN. Now, no question there were a lot of leaders doing stupid shit in their AOs, and that's been a problem in Iraq in that the highly localized nature of things means it is WAY too leader dependent and you can either end up with a superstar or a disaster, with the disaster being easier and probably leaving more lasting/pernicious ramifications.
But when I was in Baghdad in 2005, the Army guys I was with were doing COIN to the best of their ability. They weren't all just driving around waiting to get hit by an IED. Many/all commanders I knew were adamant about patrols getting out on the ground, getting to know the locals, trying to do local projects (although for some reason their money was cut way back in 2005 vice 2004) etc. But they were limited in what they could do by their small numbers relative to the AO: one Army battalion for all of West Rashid, with around a million people, mostly Sunni who resented their presence anyway. Take away guys on leave, resting, vehicles off line, what have you, and the actual numbers on the street at any give time were pretty minimal. So I'm not so sure what happened over the last 6-8 months was a function of more beat cops on the street than a real change in tactics, protect the people, doctrinaire COIN stuff, or, paying Sunni insurgents to stop fighting.