• right here:

        The Pentagon said the weapons they would send to Ukraine had a failure rate of 2.35 percent or less, far better than the usual rate that is common for cluster weapons.

        But the Pentagon’s own statements indicate that the cluster munitions in question contain older grenades known to have a failure rate of 14 percent or more.

        here is the original document they are quoting from. page 31/32 (35/36 in the .pdf’s numbering) states that:

        U.S. ammunition stockpile sample testing also indicated that DOD has experienced past problems with submunition reliability rates. For example, in 1990, testing of artillery-delivered nonland-mine submunitions identified two lots that had duds in excess of 40 percent. According to a testing document, one way to compensate for this high dud rate is to increase the quantity fired. Instructions contained in the testing document were to “Notify the user of the increase in submissile defect rate so that he can make adjustments in the tactical employment plans.” The July 2000 Army study of dud rates for ammunition reports that the dud rate for artillery-fired M42/46 submunitions is over 14 percent.

        the Pentagon’s actual, data based estimate (and not its uh, lying) correspond well to other data reported by groups such as Human Rights Watch (which have routinely found submunitions to have rates like that across the board), and specialists in the field who anecdotally report rates of as high as 30%.

        • the Pentagon’s actual, data based estimate (and not its uh, lying)

          The Pentagon’s five actual, data-based and more recent estimates, which indicate a much lower rate, are classified. Otherwise I take your point, seems you’re essentially right about the failure rate.

          I suspect the failure rate is higher than what the Pentagon is saying, probably ranging from 2-25% depending on conditions. At least it’s much safer than the cluster munitions the Kremlin is using.

          • The Pentagon’s five actual, data-based and more recent estimates, which indicate a much lower rate, are classified.

            oh, fascinating. i’m sure there’s a banal reason why the Pentagon totally can’t release these supposed much lower rates and has to classify them—but we definitely have lower rates of duds now, believe us this time! the Pentagon would never do things like selectively classify or release data to manipulate narratives, misrepresent how dangerous things are or the severity of certain weapons or political trespasses, or generally and systemically lie about everything. that’s why, for example, whenever we audit where our money is going and to what things, they fail said audit. i’m definitely going to take their classified word for it here instead of all the actual data (including some of their own previous data) which strongly implies they are lying as they usually do.