Numbers in War: a thread. I wanted to write this because it highlights some significant dilemmas in writing about ongoing conflict, and can hopefully help readers interpret and properly use reports. 1/17
Today my colleague Nick Reynolds and I published āMeatgrinder: Russian Tactics in the Second Year of Its Invasion of Ukraine rusi.org In which we report that the AFU are losing 10,000 UAVs a month. 2/17
How robust is this number? Is it plausible that the AFU are losing 323 UAVs per day? Well, UAVs are being used for reconnaissance, strike and situational awareness across 1,200 km of front and dozens of kilometres of depth. 3/17
Each platoon tends to keep two up. Iām not going to go into tactics here, but both sides need them up for situational awareness. Then command posts put them up to see what is going on. Then artillery units do recce with them. SOF too. 4/17
Then UAVs are launched for the purpose of strikes. Jury rigged loitering munitions are used throughout operational depth, including by saboteurs. In those areas where there are significant ongoing operations it is not unusual to have 25 UAVs a side for each 10 km of frontage 5/17
A lot of these UAVs are non-hardened commercial UAVs. A lot of them have to stay up to provide situational awareness, irrespective of the EW threat. The loitering munitions donāt come back, irrespective of whether they destroy a target. So the consumption rate is high. 6/17
So, it is quite easy to get to 323 lost per day. But the numbers are also extremely variable. More skilled pilots lose far fewer UAVs. UAVs in artillery units that can plan their routes, can work down the seams in the electromagnetic spectrum where infantry units canāt. 7/17
Thus the head of Ukrainian artillery noted that while the infantry get through a UAV in an afternoon, in the artillery they survive much longer. wartranslated.com 8/17
Consumption does not just differ between units but also depends on the level of operational activity ongoing at a given time. Thus the actual number of losses will fluctuate massively, day by day 9/17
How many are being lost? If the AFU had a unified logistics system it could be simple to track. This is not the case though. Across Ukraine UAVs are being crowdfunded, gifted by Ukraineās partners, and are mass produced by Ukrainian private companies. 10/17
Then there is the fact that UAVs are being used by the Ukrainian Ground Forces, by the Territorial Defence Forces, by the National Guard, the Border Guards, and Ukraineās Special Services, and they donāt all draw from the same stores. 11/17
Thus the AFU donāt know how many UAVs they are operating. So how confident am I in the number 10,000 per month? As an exact number, Iām not. The actual total varies month by month and day by day and is not closely tracked. But consumption is definitely very hgih. 12/17
So if the figure isnāt exact - any neatly rounded figures are suspicious - why use it at all? Firstly because it is the number we got from the Ukrainian General Staff and from the part that has the best data. Other organisations concurred that the number was reaosonable 13/17
Putting a number to consumption is important in making the policy relevant point. The point is, unhardened UAVs are disposable tools like munitions and get consumed very rapidly. You need them in your force and you need them to be cheap. 14/17
If we werenāt confident in the number, why use a number at all? Because if weād written āyou need lots of UAVs because many are destroyedā this would do nothing to give a sense of the scale of consumption. It would not help bound the problem set. 15/17
Let us suppose that this month the actuall loss rate is 9,000ish, does that change the policy implication? No. If it were 5,000 then that would be a meaningful deviation from the number we gave. But that would just be a random number. 10,000 was the number given by the AFU. 16/17
The challenge here is that numbers can be necessary in getting the right message across while also being misleading in their exactitude. It is important that any numbers coming out of a war are read in this context. 17/17
Loading suggestions...