This year, 16 members of the International Security Assistance Force serving in Afghanistan have been killed by Afghan soldiers or policemen turning their guns on their Western partners. On Saturday two British servicemen, an RAF man and a soldier from the 1at Battallion Welsh Guards, joined this list, being shot by men wearing Afghan police uniform in Lashkar Gar. The incident challenged assertions, made in March by local commanders after a British soldier and a Royal Marine were killed by an Afghan sergeant, that it had been an one-off incident.

Such incidents are clearly far from rare. But forming a judgement of what they mean is hampered by a shortage of evidence. We lack the ability to assess relative risk - Afghan on ISAF; ISAF on Afghan – because that requires knowledge of the relative sizes of the ISAF and of ISAF-trained Afghan National Army (ANA) personnel deployed alongside them; and of the numbers of military fatalities of  ANA troops, including how many of their fatalities (if any) were internecine (ISAF on ANA).

For months now, Patrick Mercer OBE MP has asked parliamentary questions to try to elicit these figures, but he has been rebuffed by the pusillanimous response that ISAF-trained ANA deployments and fatalities are a matter for the Afghan government. Not so, when UK lives are at risk from an enemy-within.

Over the recent past internecine fatalities by Afghan treachery have accounted for many more ISAF military lives than friendly fire. In the recent 60 weeks from 29 November 2010 to 22 January 2012, when there were 638 ISAF-fatalities in Afghanistan, treachery by Afghan personnel or trainees claimed at least 29 ISAF lives (all by small arms fire). Indeed, internecine deaths caused 4.5 per cent of all ISAF deaths; and accounted for 29/121 (24 per cent) of ISAF-deaths by small arms fire.

Are these incidents the result of deliberate infiltration of ANA troops and police by Taleban supporters, or of an extreme clash of cultures that causes exceptional breaches of military discipline? If the cause is the clash of cultures, one might expect such incidents to occur roughly in proportion to ISAF & ANA co-deployments.

The number of ISAF-troops is around 150,000 and the number of ISAF-trained Afghan army or police deployed alongside them, though not routinely reported, may be around 300,000. Internecine attacks on ISAF troops have amounted to about one per week in 2012. Given the numbers deployed, we might expect twice as many casualties among ISAF personnel as among Afghan personnel, if the breaches of discipline were equally shared.  In fact, the ratio has been closer to 20 to one (if any). So the breaches appear to be ten times as common among Afghan military and police than among  ISAF-personnel. 

Better data would allow us to infer publicly whether a simple form of proportionality fits the observed pattern of internecine deaths, or whether the non-proportionality is such as to point to deliberate infiltration.

The UK’s Ministry of Defence should have conducted such analyses. We cannot provide them, other than in a crude form, because the UK government has not been willing to pressure the Afghan government to make public the number of ISAF-trained ANAs deployed, and their fatalities.

Notice that friendly fire accounted for 7/638 (1.1 per cent) of ISAF-fatalities in the 60 weeks from 29 November 2010 to 22 January 2012, but internecine deaths for four times as many: worryingly, 29/638 (4.5 per cent).