Now and Again

Discussion in 'Middle East - General' started by Annie, Sep 1, 2007.

  1. Annie
    Offline

    Annie Diamond Member

    Joined:
    Nov 22, 2003
    Messages:
    50,847
    Thanks Received:
    4,644
    Trophy Points:
    1,790
    Ratings:
    +4,770
    Then there is the news, then there is reality:

    http://engram-backtalk.blogspot.com/2007/09/casualties-in-iraq-for-month-of-august.html

    [​IMG]

    [​IMG]

    Um yeah

     
  2. onedomino
    Offline

    onedomino SCE to AUX

    Joined:
    Sep 14, 2004
    Messages:
    2,677
    Thanks Received:
    474
    Trophy Points:
    98
    Ratings:
    +476
    The two charts above are the same. The number of casualties is somewhat skewed due to the unusual massive attack that killed 400 in far northern Iraq. I still do not think we are doing those things necessary to win. We are not attacking the sources of insurgent funding (no matter if they are outside Iraq), and we are not going after the Iranian training and weapons sources that are negatively impacting the situation. We are letting Iran influence the level of violence, and we are in reactive mode concerning their training of killers and shipments of weapons.
     
  3. DeadCanDance
    Offline

    DeadCanDance Senior Member

    Joined:
    May 29, 2007
    Messages:
    1,414
    Thanks Received:
    127
    Trophy Points:
    48
    Ratings:
    +127

    LOL

    Who does this blogger think he's kidding?

    I took his same data and plotted it in excel.

    In terms of trend analysis, Excel give you about a dozen different options for trend analysis - and only two of them give you the result this guy got (trending down in the last two months of "the surge): a third order polynomial trend analysis, or a fourth order polynomial analysis.

    EVER OTHER trend analysis you can do in Excel shows the TREND going ever upward. In other words, this guy had to cherry pick one trend analysis, out of at least a dozen options, to get the line on the graph he wanted.

    Now, the problem with him using a third or fourth polynomial trend analysis, is that it doesn't show us the overall trend. The polynomial he's using, oversamples the data. What his cherry picked trend analysis is actually showing is seasonal trends. There tends to be less casualties in June and July, because those are the HOTTEST months in Iraq, and there's less violence. Its too freakin' hot to fight in other words.

    The polynomial trend actually shows that violence always tends to peak in the spring months, and die down in the Summer. The spring being the most active month for ethnic cleansing and violence. And this years data is no different than previous years: more violence in the Spring, and less in the hot Summer. Except the overall trend is worse this year in 2007, than it was in 2006 or 2005. With the outlier exceptions of a couple of especially violent months, where there were over 3000 casualites in a single month, which oddly doesn't shoe up on his graph.
     
    • Thank You! Thank You! x 1

Share This Page