Dev Diary: False Intelligence

What's happening / TLDR: Developer diaries introduce details of Espiocracy - Cold War strategy game in which you play as an intelligence agency. You can catch up with the most important dev diary (The Vision) and find out more on Steam page.

Somewhere between 2007 and 2009, the Jordanian intelligence agency (GID) arrested a local doctor connected to al-Qaeda. After three days of interrogation, GID officers turned him against the jihadist organization. As a double agent, he was deployed to Pakistan, where he started providing intelligence and climbing the hierarchy. His case was partially taken over by the CIA. At the end of 2009, when he came back with a false claim of important intelligence about the leader of al-Qaeda, he was met for a debriefing by a team of top CIA experts on the region. During the meeting, the agent activated his suicide vest and killed nine people, including seven CIA officers. A week later, the Pakistani Taliban released a video confirming that he was turned by them - while further analyses suggest that he was also being handled by al-Qaeda's top leadership. In "a true wilderness of mirrors" (to quote James Jesus Angleton), a double agent of two intelligence agencies was turned into a triple agent of two more organizations.

As this case illustrates, counterintelligence does not end with hunting spies, nor even with countercounterintelligence. Deception can be followed by further deception, double agents can be tripled, intelligence may be false, and consequences can be fatal. While espionage is popularly understood as an art of reaching hidden information, in practice reaching hidden information is just a first step in a battle between intelligent opponents who try to exploit each other.

Espionage systems in strategy games usually miss this complexity and instead reduce spying to what I like to call a "knowledge tax". If you want to know how many tanks the other player has, you have to pay a tax in espionage points. It's the most boring implementation of espionage possible - so boring that Espiocracy does not feature it at all. You are an all-knowing intelligence community, and you simply know how many tanks other countries have. What you may actually not know is whether your agents are doubled (or even tripled), whether they feed you false intelligence about fundamental parts of the game world, and whether your next step will end in a disaster due to incorrect information.

Intelligence Materials

Espiocracy features a detailed economy of information: hidden/secret/covert events (eg. a meeting between leaders) leave traces in the world (eg. a memo from the meeting), which can be collected by intelligence agencies as intelligence materials (eg. a memo copied by an agent). These materials are then combined to produce actionable intelligence information (eg. the leaders are planning an invasion), which then fuels available actions and resources. This full cycle (and more) is contained in a single widget:

Now, going back through this chain, the success of available actions (and resources) depends on the quality of derived information, which depends on the quality of collected materials. You can, for instance, inspect each piece of derived information to see its quality grading:

This is NATO admiralty code. In the game, as inspired by the real-world usage, your operatives grade intelligence in terms of reliability (mostly the reliability of sources and intelligence collection methods, eg. good = from multiple trusted agents, poor = rumors) and credibility (mostly the credibility of the information itself in the context of the known world, eg. good = a logical expected invasion, poor = a backwater nation inventing a new technology). It's not objective, and instead it depends on the officers and their tradecraft (as well as their Cold War opinions). False intelligence may slip through and... genuine intelligence may be falsely graded as false. When information is judged poorly enough, it is rejected as false and not used in the further intelligence cycle (except for a potential counterintelligence investigation).

(By the way, players on the red side use "Markirovka" instead of the NATO code)

Going back further through the chain, false information is derived from false intelligence materials.

Every intelligence material has an inherent "falsehood" property (from 0 to 100). It is higher than 0 only when the material was purposefully prepared as a deception by another intelligence agency or a paramilitary organization - so it is not used for ordinary fabrications or for anything close to a knowledge tax. Instead, it is always a weapon in the battle for truth between intelligent opponents.

This simple approach to false intelligence (fundamentally relying on one 0-100 variable) in the open-ended, simulationist intelligence economy organically creates many fascinating and complex scenarios. False materials may spread through the world and mix with genuine materials. Various intelligence materials may be in conflict about the same event, forcing analysts to make a judgement call. Players may unconsciously share false intelligence with allies who then realize they were duped. Good sources may be temporarily disabled due to picking up false intelligence or receiving poor grading, while double sources may gain trust by providing a lot of genuine intelligence. A mole hunt may use the "barium meal" method, in which specific false intelligence is released to a specific officer, and then counterintelligence services observe whether this false intelligence was acted upon by suspected handlers... and so on.

Feeding False Intelligence

As hinted above in many paragraphs, the main source of false intelligence in the game is double agents. A player taps into a source already trusted by another player and turns it against the original handler. Usually, agents can be turned after they are discovered and intercepted through a successful counterintelligence investigation (see previous dev diary). Turning is the goal of most investigations because it is much more valuable than an ordinary arrest or expulsion - it allows the player to feed false intelligence, which then wastes the opponent's resources, disrupts their other espionage activities (especially inside the country of the doubling agency), and sometimes even leads to disasters.

Generally, agents (including double agents) are autonomously handled by officers, balancing for instance the ratio of genuine to false intelligence to inflict the most damage. However, you may always intervene as a spymaster and, for instance, produce a specific false intelligence material to cause any intelligence avalanche you desire.

Turned agents may be turned again, into triple agents, who generally have much shorter careers and are much more damaging. (There are no quadruple or further agents. We prototyped an open-ended implementation, but it traded the distinct mechanics differentiating double and triple agents for mediocre, similar mechanics for nth-order agents who could rarely, if ever, appear in the game.) First- and third-layer handlers may be the same (as in the real-world case described at the beginning), bringing even more potential damage to the second-layer handler. In a case where handlers on all layers are different, the tripling handler comes out on top, while the first and doubled handlers are simultaneously damaged. In addition, triple agents have various distinctive characteristics, such as much higher mental pressure (which can even result in the agent running away from all three handlers, or even entirely from this world...).

Depending on the level of tradecraft, specializations, and other capabilities, players may pursue more advanced (and more delicate) methods of feeding false intelligence:

• Moles - officers of one intelligence agency who are recruited by another intelligence agency can internally spread false intelligence, especially when it comes to work against the handler's country (some places refer to this kind of agent as already "doubled" but in the game doubled/tripled consistently refer to the number of "betrayal events" so a usual mole is just an agent because they have betrayed only once)

• Dangles - false walk-ins, defectors, and even moles, who (if treated as real by the targeted player) are doubled from day zero

• False flags - broader deceptive actions that involve partially controlled state and non-state actors, for instance, a leader staging a holiday departure that generates false intelligence materials suggesting that an invasion is less likely while, in practice, an invasion is imminent

• Propaganda campaigns - most players automatically see through propaganda campaigns (since they also execute their own propaganda campaigns), but agencies poorly versed in the propaganda world can temporarily fall for false intelligence produced through disinformation propaganda campaigns

• False espionage activities and operational materials - players can establish false stations, organize false infiltrations and exfiltrations, launch false operations, manufacture false opportunities, invent false secrets, and so on; this is a standard tool of countercounterintelligence, usually launched autonomously by officers, with no groundbreaking results, and mostly intended to waste the opponents' counterintelligence resources

Seeing and Using False Intelligence

To avoid the aforementioned "knowledge tax", false intelligence can affect only a few critical places in the game. This possibility is always clearly marked:

(Note the calculator icon: nested tooltips explain calculations)

When the extent of falsehood is high enough (compared with inverse tradecraft; that is, higher tradecraft means quicker detection), false intelligence is marked as probable:

The main affected element is strategic intelligence (SI). It is extracted from derived intelligence information - if the information is false, the gained SI is also false. Since the recent dev diaries describing SI, it has been iterated upon: in its current shape, as you can see from the tooltip above, SI is both a parameter (yearly accumulation, used mostly by the government) and a resource (limited accumulation, used mostly by the intelligence community). As a resource, you can use SI to jump-start various strategic operations:

When you launch an operation based on false strategic intelligence, it will have a risk of failure proportional to the false information used. If the level of false intelligence is high enough (eg. it's possible to launch a coup d'etat based on completely false intelligence!) or if it's launched against the author of the false intelligence (authorship is usually tracked), it may end up as a huge disaster, with an ambush, folded networks, arrests, deaths, and damaged international relations, potentially even establishing a casus belli for the opponent.

False intelligence also disrupts large governmental actions, such as managing nuclear brinkmanship, which rely on SI and on the difference in SI between the countries involved (e.g., the USSR with SI higher than that of the USA will have more time to covertly deploy nuclear missiles to Cuba). While such actions are launched based on the perceived SI, final calculations do not include the false fraction (e.g., the USSR may have actual SI much lower than its perceived level due to false intelligence, and then the deployment to Cuba will be quickly spotted before it's finished).

From the more tactical operational side, false intelligence can also affect standard intelligence operations - either by inflating or lowering the 0-10 score of the operation, both for the attacking and the defending intelligence agency.

The level of corruption of an operation is determined by a series of calculations that take into account double and triple agents, moles, false intelligence materials, and false operational intelligence (vulnerabilities, opportunities, secrets). When the operation is resolved in the final steps, usually through a dedicated simulation, it takes into account only the real 0-10 score. The discrepancy may be extreme in either direction: perfect operations misperceived as 0/10 (and therefore not launched) and disastrous operations seen as 10/10 (and therefore ending in a full ambush). The latter can happen in cases where everything aligns against the active party, for instance, when MI6 launches an operation against the KGB from a false opportunity manufactured by the KGB, which also has a mole inside MI6, and one of the MI6 agents in the USSR who was supposed to support the operation is doubled by the KGB.

The third category of potentially false values in the game is critical country-level risk: coup d'etat, invasion, and fall of the government.

On the one hand, the level of these risks gates and aids various actions - so, for instance, a coup d'etat becomes easily launchable once its risk reaches 80-90%, and its success chance increases with every percentage point... but the final coup-day simulation takes into account only the real value, so a coup launched based on false intelligence inflating its risk will probably fail. On the other hand, a real value reaching 100% still can launch (nudge) relevant events, so false intelligence lowering the perception of an invasion is an organic way to achieve surprise invasions in the game. Moreover, here self-deception is also possible - for instance, through an equivalent of Operation RYAN, the USSR may convince itself that an invasion by the West is imminent. The opposite case is also possible: everyone in the world is deceived except for the homeland, which in the game usually recreates the surprise fall of the USSR.

For now, this is the extent of the effects of false intelligence. We are prototyping other approaches (such as direct wartime false intelligence), but it's a powerful tool that significantly affects gameplay, so we are very careful with its implementation. It's no coincidence that false intelligence is the topic of the 74th dev diary.

Counterintelligence

False intelligence is generally temporary, decays over time, and continuous deception requires continuous feeding. However, waiting it out is the least valuable approach.

Counterintelligence investigations can detect false intelligence at all stages (collected materials, derived information, false strategic/operational/risk values). Naturally, if successful, they remove false effects - but far more important is further attribution. By tracing false intelligence back to its sources, your counterintelligence officers can discover double and triple agents, unreliable sources of intelligence collection, moles, and all the other tools used to feed false intelligence. This is the major tool for detecting turncoats (the second-most important tool, after defectors pointing to turncoats). Unlike standard "espionage operations" in various strategy games, where a failure is just a failure, in Espiocracy a failed operation is often a step forward in a counterintelligence game that may dismantle very valuable assets of the opponent.

Final Remarks

Next dev diary will be published on July 3rd.

If you're not already wishlisting Espiocracy, consider doing it

https://store.steampowered.com/app/1670650/Espiocracy/

There is also a small community around Espiocracy:

--- Photo credit: Stasi encryption machine from Deutsches Spionagemuseum Berlin