Updated: Mar 5, 2022
Bar. Cage. Slat. Mesh. Net. Chain. It has many names, but all belong to a family of armour – statistical armours. Designed to defeat RPGs, it has become a normal sight on every class of vehicle from logistics trucks to main battle tanks. But how does it actually work, and why do you almost certainly think it does some things it doesn’t?
What is it?
As the name suggests, statistical armour is designed to defeat its target threats by use of statistical probability. Where conventional armour seeks to just cover a vehicle in protection from any axis, statistical armour plays a numbers game to provide good (everything is relative) levels of protection whilst keeping weight and volume down and visibility up.
Let’s get this one out the way up front and centre:
Statistical armour is exclusively for defeat of shaped charge munitions using a double-skinned nose as conductive path for a piezoelectric-based fuzing system.
That refers mainly to the RPG family, including systems in the RPG-7/PG-9/PG-15/SPG-9 families. Why have I written that in a patronisingly bold stand out manner? Because the pervasive misunderstanding of statistical armour is what it can and can’t do, and why it exists. Even professionals in the field of AFVs, weapons and armour commonly hold a flawed understanding of what this armour does.
First, how does it work?
To understand how statistical armour works you first need a basic overview of how the threat it is trying to defeat works. Below is a representative diagram of the cross section of a typical RPG. The explosive charge is the (6) yellow material, which will detonate and form a shaped charge jet from the (4) liner material forming a cone in front of it. The hollow area to the front of the RPG is just empty space needed to allow the weapon to initiate at a standoff from the target.
When the (1) nose of the RPG hits a target, it crushes piezoelectric crystals which create an electrical current that is transmitted to the fuze at the rear of the warhead via (2) an inner 'skin', resulting in detonation of the charge. Statistical armour targets this mechanism of action.
Going back to our statistical armour, it forms a big grid, where a majority of it is actually open space, with rigid elements dividing it into up. This means we now have a statistical likelihood that if an RPG is fired at the grid, its nose will go through a gap rather than hit something. The grid has been designed at such dimensions that it is slightly smaller than the anticipated warhead (generally 65-95 mm for something like a PG-7) and as a result after the sensitive tip has passed without incident, the main body of the warhead collides with the grid and is deformed, making the outer and inner skins of the RPG contact one another internally.
This has now short circuited the fuzing system, even if the tip does now impact something and send an electric charge, it will not reach the fuze at the rear and the charge cannot detonate.
Given the velocities involved the warhead is more often than not critically damaged, breaking it apart and destroying it, or catching it in the grid without impacting the vehicle itself. We are dealing with explosives, so whilst a detonation cannot occur and thus a shaped charge jet cannot be formed, the charge can deflagrate or burn (see my primer on Insensitive Munitions (IM) for some more detail on the varying grades of energetic reaction) and so a defeat is not always without some fanfare, but the key mechanism of attack via shaped charge jet has been thwarted.
Even in an ideal defeat where the RPG sticks in the bars and never impacts the hull of the vehicle at all, a degree of fragmentation and debris will still project from the RPG onto the vehicle. In a suboptimal defeat there will be deflagration with accompanying fragmentation. In any case the vehicle being defended by statistical armour needs some form of passive ballistic armour to protect it from the consequences of the defeat mechanism, and 'no effect' is an outcome that will never be recorded in a defeat scenario.
A secondary risk is that in the event of a defeat there may be a grenade trapped on the side of the vehicle that needs to be removed, though more often than not the warhead breaks up sufficiently from interaction with the bars/mesh to render the risk relatively low.
How effective is it?
Perhaps surprisingly ineffective, actually. The issue stems from the fact that the RPG needs to pass through a gap and have the side of its warhead strike a rigid element sufficiently to cause a short in the fuzing circuit. RPGs come in a range of sizes and shapes, and the grid can only be designed in a single size.
If the RPG impacts too close to a grid element it will initiate the charge and form a shaped charge jet as intended (see Things people commonly get wrong for more on this in a bit). If it hits too perfectly in the centre (and is of the appropriate dimensions), it may not be sufficiently damaged by the surrounding grid to neutralise it before the charge is triggered.
So the designers need to develop the grid geometry quite carefully to maximise defeat against the intended threat projectile. But this only works in a mathematically perfect scenario - a perpendicular engagement. In reality, RPGs come at you from all elevations and azimuths. And as the approach angle is tilted laterally or vertically the effective shape of the aperture that the RPG is being presented changes too, becoming smaller and thus more likely to see an impact on the grid rather than passing through it.
So you could open up the geometry to give better odds of an intercept at off-axis attacks, but then you just made the apertures bigger when the threat does come on-axis, and so it is now more likely those will fail to be neutralised. As with all AFV engineering, it is a case of trade-offs, modelling and developing the defence against the most likely threat, with mitigation or acceptance of the inherent drawbacks that you have taken in exchange.
So, saying how good statistical armour is in general terms can be something of a moving target. Against which RPG model? At what elevation and azimuth of attack? With what armour geometry?
A representative example is the below chart taken from a Polish study into mesh armour. The key takeaway is that in the ideal condition - 0° elevation/azimuth, perpendicular 'on-axis' attack with a design of mesh well aligned with the threat projectile dimensions - the probability maxed out at 42% chance of intercept, dropping to c.10% by a 40° attack axis, and then 0% by 45°. As the charts above show, statistical armour is more sensitive to elevation than it is azimuth generally, but suffers in any case.
Things people commonly get wrong.
There is one critical myth that is most routinely repeated about statistical armour, which is that its intent is to initiate an RPG on the armour and thus 'nullify the shaped charge jet before it can reach the vehicle'. This is completely wrong, despite how widespread its belief is. So pervasive is this myth, it even crops up in official engineering oriented documentation, such as the UK Defence Standard 23-10 Design Guidelines for Ministry of Defence Specialist Vehicles and Plant Equipment which, despite being totally wrong, says:
"Stand off armour or cage armour is designed to withstand an attack from anti-tank rocket propelled grenades (RPG). It functions by placing a rigid barrier around the vehicle which causes the shaped charge warhead, which uses a shaped explosion rather than kinetic energy, to explode at a relatively safe distance. " Defence Standard 23-10
As explained earlier, the specific method of defeat for statistical armour is to allow an RPG to pass through the holes in the grid and then damage the structure of the warhead, critically damaging its function as a fuze component. The intent of the armour is not, in any capacity, to present a spaced armour layer that initiates an RPG at a distance from the defended vehicle.
The reality of this myth is actually more significant - detonating the warhead on the bars can be at least as bad as if it had hit the defended vehicle and, in some cases, actually a worse result for the defended vehicle.
Shaped charge jet penetration is a well modelled relationship of standoff in cone/charge diameters (CD) resulting in a depth of penetration also measured in CD. The specific relationship can vary (quite significantly) depending on liner geometry, the material of the liner itself, the composition design of the charge and a range of other factors.
An exemplar chart is shown below, where one could calculate that for a given warhead, for example c.75 mm for a typical PG-7 series round, the penetration when at a in-built standoff of around 2 CD is in the region of 350 mm. Note that this chart is not showing the PG-7 curve specifically, the following numbers are just illustrative.
As always, things aren't that simple. As you can see, the penetration varies considerably as standoff changes, with the best penetration tending to be be in the 6 to 10 CD range. For a PG-7 that's a standoff of 450 mm to 750 mm, vastly more than can be achieved with the PG-7's warhead's built-in standoff of c.155 mm from nose to charge.
Hopefully the effect is becoming obvious - by impacting the statistical armour and detonating at that point, the armour has given the RPG a free c.300+ mm of free standoff extension, pushing from its inherent 2 CD standoff to something more 6 to 8 CD, and in turn upping penetration by a full CD or more. The armour just amplified the threat capability by a good 20%, making life worse for you inside that vehicle. The actual effect is heavily reliant on the specific CD standoff/penetration curve for the particular warhead design in question, but as a representative example it is indicative of the issue at hand.
These figures were hypothetical, the PG-7 does not conform to this specific chart, but the principle endures - RPG projectiles rarely have ideal built-in standoff in order to avoid large and unwieldy rounds that would be onerous to transport and handle, and so detonation on statistical armour is in essentially all cases a modest enhancement to their capability as a threat to you inside the vehicle.
In the rare cases where it is a degradation, for example with very large and unwieldy projectiles like the PG-7VR tandem warhead round, note that the penetration chart declines much more gradually than it rises - you need to get the round beyond 14 CD or more before there is a tangible degradation in penetration, and that reflects nearly a metre of standoff, which when doubled to reflect this armour being on both sides of a vehicle would have woeful implications to operating such a vehicle in anything but open country.
There is no reason a designer would seek to extend the standoff in this manner, and so it is a myth that this is the intent of the armour concept.
A quick note that the material that the RPG charge liner is made from has a significant impact on penetration. Newer and denser materials can radically increase the CD penetration ratio. Most RPGs utilise a copper liner owing to a high density - 8,900 kg/m² vs 2,800 kg/m² when comparing copper to aluminium, another liner option, both shown in the chart below.
Newer and even more dense materials like molybdenum can further increase penetration by 10% to 20% over copper.
A less common myth is that statistical armour is a continuation of the German Schürzen armour panels from the 1940s, based on a belief that these armour panels were developed in response to Bazooka/PIAT-type weapons proliferating through US and Commonwealth forces.
This isn't true either. Schürzen is not a statistical armour but a spaced armour designed in response to Soviet anti-tank rifles, not Allied HEAT weapons, which didn't even exist in any significant capacity at the time that Schürzen was first fielded. Given the US and UK did not field large calibre anti-tank rifles like the Soviets, it is likely they simply erroneously connected the armour to their own nascent AT weapons and this stuck as a narrative.
So, where do ATGM come into all of this?
Well, they don't. Whilst they could feasibly suffer damage and destruction prior to impact, in almost all cases ATGM are of a vastly larger size than an RPG - for example 152 mm charge diameter for a 9M133 Kornet vs 93 mm for a PG-7VL.
They do not use the piezoelectric fusing concept of an RPG, and so destruction of the nose is not going to cause a short and thus defeat the weapon. Most ATGM utilise one or more of RF proximity and impact fuzing, so the overwhelming likelihood is that they detonate on or at a standoff prior to reaching the statistical armour. In addition, a good portion of ATGM designs utilise top attack profiles, negating statistical armour entirely.
Forward attacking modern ATGM designs mount the warhead quite far back on the missile, such as in the MBDA MMP diagram above, in large part to optimise standoff. This does mean that detonating on or before the statistical armour array is likely to be detrimental to penetration, however ATGM boast radically higher penetration capabilities than an RPG owing to their much larger dimensions and that most designs are tandem warheads with precursor charges almost as large as most RPG main warheads.
This penetration can be enormous, in excess of 1,200 mm of RHA armour in the latest guises, and so radically overmatches almost anything it will encounter when it comes to statistical armour packages, which are generally placed on the sides and rear of vehicles where the armour is thinnest.
When it comes to modern ATGM the solution is to prevent the hit in the first place via soft and hard kill active protection systems, signature management and other protective technologies, not to utilise statistical armour which is entirely unsuited to the threat. The solution is very much in the outer layers of the survivability onion prior to an impact occurring. If an ATGM is going to hit, it is going to do some very serious damage.
Drawbacks of statistical armour.
The key drawbacks, relatively limited efficacy aside, are the size and weight of a full vehicle package of this armour type.
Something like an M1126 Infantry Carrier Vehicle (ICV) variant of the Stryker gains around 2.5 tonnes when it adds a full bar armour package, which is about 15% increase in GVW (a brief acknowledgement of how light a Stryker is compared to the contemporary crop of 8x8s like Boxer which sits around the 38 tonne mark, over two times heavier).
Basic steel bar armour is the heaviest of the statistical armour types, this can be reduced by up to 90% by using lighter design types like mesh armour, albeit with potential implications to performance as mentioned earlier. Everything is a trade-off with engineering and design.
Size is the other major issue. regardless of type, statistical armour typically sits at a standoff of around n mm by mounting it on brackets that project from the hull. As it is on both sides, the effect on vehicle dimensions is doubled - an additional 600 mm increase or more in overall width and length.
This has significant impact on logistics, as loading limits on road/rail/sea/air transporters are very limited, and can rarely accommodate such a large increase. This means the vehicles need to be torn down and refitted with armour before/after transport, and the armour packages need to be transported in addition to the vehicles.
It also has impacts on the tactical use of the vehicles. AFV are not small as-is, and are challenging to operate in complex terrain such as woodland or urban environments. Adding signficiant projections around the vehicle further hampers sight lines for the driver and crew, and makes the vehicles even harder to manoeuvre without striking obstacles.
There are solutions to mitigate these issues. AmSafe Bridport's Tarian mesh armour is lighter and lower profile than simpler bar armour, and can be mounted on folding brackets. The brackets collapse when fouled, preventing the armour panels from being ripped off or causing excessive damage to the fouled object, and can be rapidly folded flat against the vehicle for transport rather than having to unbolt and remove everything.
Bar armour also has human factor implications to fit it. The UK Mastiff version of the US Cougar 6x6 doesn't have doors for driver/passenger at the front like the US version does as passive and bar armour packages block them entirely, the crew have to crawl in (and out) from the rear door.
Whilst you can hinge bar armour panels, they risk fouling or prevent full movement of hatches and doors owing to their substantial size and the standoff brackets. Bar armour arrays on the sides and rear of tank turrets will block engine bay access and have to be removed to allow simple maintenance tasks to take place.
The implications of fitting a statistical armour package can be a range of blind spots, fouled weapon and optics arcs and significant implications for ingress/egress of personnel and access to key areas of the vehicle.
Statistical armour types.
There are three broad types of statistical armour. They’re all doing the same thing described above, just in subtly different ways and with varying pros and cons attached.
The first are the most common, referred to as bar armour (a more British term) or cage/slat armour (a more American term), they are big rigid metal assemblies that look akin to a metal railing or cage at the zoo.
They're the heaviest statistical armour type by some margin and very bulky, but also reflect the most basic and straightforward implementation of the concept. Being made of steel (though aluminium-based examples are often used too), they can be cut and welded in the field to repair incidental and combat damage.
Mesh and net armour are, as the names suggests, a more mesh net or chain-link fence-like construction and generally lighter, more flexible and less vision blocking than their more rigid cousins.
The mesh is usually a metal core with fabric or polymer coating on it, and some designs use a rigid block that is bonded onto each joint of the fence material to ensure strength when hit with the significant forces of an RPG, and to give some more substantial elements to aid in breaking up projectiles as they collide with the mesh.
Chain armour is perhaps the most simplistic of the family, with literal chains hanging in a curtain arrangement, generally with some kind of heavy ball at the end to maintain some degree of tension to the chains.
You don't see this configuration all that often, with the main example being the underside of the bustle of the Israeli Merkava turret. The advantage is that it is very flexible and so can readily brush against obstacles and the vehicle itself without incident, where bar or mesh is likely to be damaged. Its efficacy, however, is not the same as the more consistently designed versions above, reflected in its limited utilisation.
Home grown armours
A final group are the entirely home made examples that are increasingly seen around the world, with Syria having been a particularly bounteous source of examples. In general they have used the bar armour approach, though the design of the grid itself appears to be more based on what happened to be lying around than a deliberate design choice.
One aspect of these designs that can be very significant is that the users have a tendency to fill the space behind the bar armour with solid material - rocks or sand bags - in the mistaken belief this is providing some additional physical protection. In reality, this is blocking the gaps in the grid and meaning a vastly higher likelihood that the nose will impact this material before the warhead is deformed by the grid, and thus the charge will detonate.
With shaped charge jets rather unimpeded by sandbags or rocks of the dimensions involved, these configurations do little to limit the effect of the weapon and, as mentioned earlier, may actually be making it worse for the defended vehicle by improving the optimal standoff conditions and decreasing the odds of a successful intercept.
Is bar armour worth it? It depends, but broadly yes, it is cheap, simple, and adds a layer of increased protection against a common and widely proliferated threat that would otherwise be very likely to do significant damage to a vehicle, particularly lighter classes of vehicle like protected patrol and utility vehicles (MRAPs in old money) that are vastly overmatched by the threat.
Are they an assured defeat mechanism - absolutely not, the reality is that they are relatively low probability of defeat and even then only in ideal engagement geometries. And, if you take nothing else from this little primer, remember the reality of the greatest myth of this armour - it is not in any way meant to detonate an RPG at a standoff from the vehicle!
Disclaimer: These primers are intended to be short and sweet, and very introductory. There are nuances and depth to the true engineering picture that can see specific exceptions to rules and pedantry (which I entirely approve of) will be able to pick holes in the explanations. But as a short, rough and ready summary of the tech, it hopefully does the job. If you want the full depth on statistical armour there are a legion of books and papers out there to dig into the specificities, many in the public domain.