Free-range thoughts The Mechanics of Opportunity: Can We Design Objects Not to Be Stealable?

Back then, theft of side mirrors and car radios was extremely common, at least in Poland. But it made me wonder: why do certain objects end up in the category of ‘worth stealing’ at all, despite their low or average value per item? And what defines that category?

Car radios are a good example. You can’t just grab one and walk off. You have to break into the car, damage something, and take risks. And for what? Even in the 90s, buying a radio wasn’t that hard or that expensive, so there must be something else, besides just the value, at play here.

So what makes objects land in that ‘stealable’ zone? And how to design objects in the first place so they don’t end up in that category?

Criminology made an attempt to define ‘stealability’ by using the CRAVED model. An object worth stealing is Concealable (easy to hide), Removable (easy to take away), Available, Valuable, Enjoyable and Disposable (easy to liquidate). But this is a very one-sided approach, which doesn’t take into account human behaviour aspects. Car radios or side mirrors were not enjoyable, nor were they valuable. They were easily liquidable, though, as there was a huge market for those with no one asking about the artefact’s origin. But this is not all there is.

Just to compare side mirrors and jewellery, the latter is much more enjoyable, equally easy to conceal, and much more valuable, which would make it fit the CRAVED model better. It is harder to dispose of in some cases, though, that’s true. But more importantly, it is more difficult to obtain, given that it is usually attached to the actual victim, while the car radio is available in a car parked in a public area with the owner not being around, which increases anonymity and lowers the cognitive load of the theft which the CRAVED model ignores.

Also, if CRAVED were the only thing there was, people would not steal big objects, just the concealable ones. And they would not bother with those that require a lot of effort. The solution then would be to design all objects big enough or fixed in place – and that would be it. As such, CRAVED does not really help with the design even if it defines ‘stealability’ in a shop’s context well.

So where’s the threshold? Is it a value vs time and effort vs size to obtain the object? Then again, where does the payoff peak? And more interestingly, what else should we take under account, and can we manipulate the other variables to design things so they don’t end up on the very hearth of the CRAVED model?

I’m no mathematician. Just recently, someone reminded me how good it felt when I was still thinking that way those 20 years ago. So let’s try again. Let’s do a thought experiment and attempt to build something that might be more usable during the design process, even before we place an object on the market and involve the law enforcement in protecting it.

To create something like an equation of ‘stealability attractiveness’, I took into account the value of the object, the physical cost of obtaining it and the cognitive cost of the theft.

So first, there is the actual value that is a combination of the perceived value and how easy it is to sell the object. Here, we have the following variables:

• $V_P$ – perceived value: The value that is immediately visible to the thief, which can be the black market value. Separating the front panel made this variable drop to near-zero. The object was worth nothing second-hand without the panel.

• $U$ – universality / fungibility: This is the degree of standardisation where ($U \in [0, 1]$). Does this object fit into any standard slot or environment without modification (or with almost effortless modification)? Old car radios had standard ISO connectors ($U \approx 1$). Modern infotainment systems are physically fused into a dashboard and are car model-specific ($U \to 0$). Universality is a physical design property that directly dictates how easily the object can be absorbed by any end-user.

Second, we have the mechanical obstacles:

• $E$ – effort and time to extract: This is the time required to detach the object from its environment. Smashing a window and pulling a radio takes seconds. Removing certain car parts requires specific tools and dismantling interior panels (let’s assume that $E$ increases exponentially and use $E = e^{\text{z}}$) where $z = \text{the number of elements to disassemble or break}$

• $H$ – hazards & noise: The above process can be quiet and safe or… it can be very wild. Does the plastic snap with a loud crack? Does the removal trigger a loud alarm that cannot be bypassed/disarmed? This is the sensory part of the theft.

• $S_W$ – size & weight: It is the physical footprint of the object. Large, heavy, or awkwardly shaped objects cannot be slipped under a jacket or carried while running.

Last but not least, we have the psychological cost:

• $A$ – ambiguity: The lack of certainty on whether the effort is worth anything. This is exactly what the loose wires achieved. The thief could look through the window but encountered a glitch: Is the radio actually there? Is it broken? Is this a trap? Is it wired to a makeshift alarm? Ambiguity forces the brain to stop and evaluate.

• $R$ – cognitive risk of apprehension: This is not about the actual law, but about the emotional price of stealing. When non-existent, say you find a $50 note lying there in the street with no one around and no CCTV, the risk is zero or close to it.

• $P_D$ – proximity / victim anonymity: It is more cognitively heavy to steal a wallet from someone’s hand than to steal a car radio from a car parked by an unfrequented road with the owner not being around. Plus, such a car radio is not personalised to the victim in any way – it’s completely anonymous. So $P_D$ is actually the $y/x$ where $y$ is the personalisation factor ($y \in [0, 1]$) while $x$ is the proximity to the victim ($x \in (0, 1]$) where almost 0 is the minimum (people don’t usualy steal from themselves, so I omitted 0 as a value here), and 1 means the victim is not around – so no contact is required – and is also unknown.

Overall, I put the actual value in the numerator and divided it by the cognitive load elements multiplied by the actual effort:

The stealability here is defined as the actual value that can be potentially lowered by how difficult the object is to reuse and sell, divided by the psychological cost and the physical cost multiplied together. The idea here is that if the psychological friction is zero (no victim, no ambiguity, and no risk), time and effort also do not matter – the denominator goes down to almost 0. It doesn’t matter how much time it takes to steal a car radio from an abandoned car in the middle of a corn field where your only witnesses are the crows there. It’s also true when reversed – if there is no effort, time is close to 0, and there is no noise – the cognitive effort does not matter, almost as if you were invisible while stealing. And when the denominator $\to 0$, the numerator $\to ∞$, so such objects will be highly stealabe.

When designing objects, there are obviously things outside of our control, like $U$ where the law dictates standardisation requirements, or $V_P$ since we do not control the actual second-hand value, and $S_W$ since the user dictates how big an object should be to be comfortable to use. But there is still stuff that we can manipulate during the design process:

• $A$ – if the object is designed in a way that it can be, for example, folded into looking like something else, it’s safer. The brain cannot scan for infinite options of what a thing can look like.

• $P_D$ – if we can make the object only visible when it’s close to the owner, we make the theft more cognitively heavy. And if we personalise objects in a way that it’s evident and the object cannot be turned into something else (personalised jewellery can still be melted down), we make the theft even less profitable by limiting the number of potential buyers.

• $H$ – if we can make the extraction more noisy or obvious (remember the clothes anti-theft protection that spilled paint allover?) we win.

• $E$ – and this is the one highly overlooked, I think – if we add a security measure to the object itself, i.e., do the IT thing called ‘security by design’ instead of relying on it being hidden by the user, we make theft more effortful.

So this is the entirely loose equation not based on research. But I wonder – what if by taking the theft possibility into account when designing objects, we can make theft less popular than by enforcing the law? Just an idea.

DISCLAIMER: This model was built for fun only. It is not based on anything else than my own thinking process, so if you want to use it, use it with caution.