Assassinations in Megagames

So my ambitious plans for weekly posts did not come to pass. Here is the long delayed post on Assassinations in Megagames.

First, by assassination I mean “murder by sudden or secret attack often for political reasons : the act or an instance of assassinating someone (such as a prominent political leader)”. This is the Merriam-Webster dictionary definition. Death from natural causes, in battle, or by execution are all quite different.

Second, the problem. Assassinations are challenging because they directly target a player’s character and/or role in the megagame. Character death and/or role change is a disruptive experience, and has the potential to make the experience of play an unhappy one for the affected player. In chatting about death mechanics in megagame design forums I have definitely run into people who think character death is a bad experience to include in a game. Part of the joy of Megagames, however, is the agency granted to players to try special actions in the game. So how can we reconcile this?

Mechanical Implementation

Factors I think need to be considered: (1) An assassination attempt follows a player decision, rather than being a Control inject into the game. (2) An assassination attempt is a secret action, requiring Control adjudication (unless your game is big enough to have a player in the role of assassin for hire, or some kind of assassin’s guild). (3) Some form of mechanical resolution is required, its not a game of rocket tag.

A short digression on bad megagame mechanics:

• Pinatas – where players get a free swing and might get lucky
• Vending machines – where players put in resources and get what they want
• Roulette wheels – where players put in resources and might get lucky.

The reason I think these forms of mechanics are bad in a megagame, is that they do not require interaction with other players. If the best way to get something done in a megagame is to ignore other players, I think the overall game experience is likely to be poor. Combine these with the surprise of an assassination for the target, and players are potentially being told “you’re dead, no saving throw”. Which is not fun.

Now some examples of what I think are good mechanics from a couple of boardgames I like:

• Junta – a key interaction in this game is that each player chooses a hiding location each turn (e.g. Home, Bank, Nightclub, Headquarters). An assassination can only succeed if another player successfully predicts where you are hiding. Some locations allow special actions, e.g. being at your HQ allows you to start a coup.
• Dune – in battles, subordinate leaders in a faction can be killed. Each leader can play one card for defence (e.g. a poison snooper) and one card for attack (e.g. a projectile weapon), or commit an expendable “cheap hero” to lead in their place. Having a leader killed can cause your side to lose the battle.

Both of these games have elements of bluffing and trying to predict what the other players are doing. The mechanics are all player facing and do not need Control adjudication.

Simulating modern era assassinations

The key insight from the article I read in writing this blog post is that assassination attempts are rarely successful in the way intended by the assassins (or desired by gamers).

From 1875 to 2004, there were 251 serious assassination attempts on national leaders, with 59 successful assassinations. This excludes leaders murdered during coups, or plots that were uncovered before the assassination was attempted. Rounding up, you could call it a 25% success rate. If you include less serious assassination attempts, the success rate drops to around 20%. Assassination attempts are more likely in large nations, autocracies, and states at war, but are still a relatively rare event in modern history (about one every two years).

Firearms were used in 55% of attempts, and explosives in 31% of attempts. Firearms had a 30% success rate, while explosives had a 7% success rate – and explosives tended to produce more collateral damage to other people.

The actual impact of an assassination is relatively small:

• Assassinating autocrats has a 13% chance of accelerating a transition to democracy, and a 19% chance of increasing future leadership transitions by institutional means
• Assassinating democratic leaders causes no institutional change or change in leadership transition
• Successful assassinations have a 25% chance of ending a high-intensity war early, a 33% chance of extending the duration of a mid-intensity war, and have no statistically significant effect on triggering new wars. The assassination of Archduke Franz Ferdinand is a significant historical outlier.
• Failed assassination attempts can allow autocratic leaders to impose repressive measures in their country.

Simulating pre-modern assassinations

I do not have a handy article for success/failure rates of pre-modern assassination attempts. When government is dominated by mainly hereditary dynasties, there is an inherent fragility potential in dynastic succession. Assassinations are usually done by family members or subordinate officials, rather than foreign professionals. An assassination can end a dynasty, install a regent or council of wicked advisors, or replace a weak leader with a strong leader (and vice versa). The stakes are definitely greater than for modern governments.

Considerations for a Megagame

First, is this appropriate for your game? Not all games will require assassins, or perhaps you could make assassination the key focus of the game – a John Wick megagame? In the Survivor: Dark Lord megagame, I started most of the players in service to the Dark Lord, and every 20 minutes the Dark Lord executed one of their minions, and Control drew one random name out of the Assassin ballot box (which players could pay resources to stuff with the names of targets). So by the end of the game nearly all the players had new roles as rebels fighting the Dark Lord. Another way to include assassination in a game is simply to have it occur as a major event in the first Act of the game, e.g. Caesar gets murdered in the Senate, starting a civil war.

Written and oral briefings to the players should be clear about the distinctions between role (e.g. President), character (e.g. Abraham Lincoln), and outline what happens if a character death occurs. I think the main options following an assassination are:

• replace with natural heir after a short break, keeping the player in the same role with little or no change to their goals, resources and abilities. This returns the game to the status quo.
• replace with new character, keeping the player in the same role, but with different goals, resources and abilities. This can change the game a bit.
• replace by shuffling team assignments, e.g. a Vice President becomes President, and one or more other players on the team shuffle into new roles. This is definitely likely to change the game.

Replacements should be made quickly. This means having prepared alternate roles and characters before the game. These roles/characters should be potentially as fun to play as the original, not third tier operatives acting above their pay grade.

Making assassination an interesting experience

My own design assumption is that it can be taken as given that significant leaders have bodyguards, doubles, and secret service detachments, and that they are effective at their job. I do not see “hire more German bodyguards” as an especially interesting decision for players. I think that for assassination actions to be interesting, they need to involve the following elements:

1. Resources – the action needs a cost of some kind, otherwise players will keep trying it over and over again (Pinatas are bad mechanics). This can include the time taken to set the action up.
2. Secrecy – the action should be a surprise to the target. This is likely to require Control adjudication to resolve.
3. Conspiracy – the action should be more successful if more players support the action, and this takes time to organise, and allows betrayal.
4. Betrayal – a plotter might betray a conspiracy causing it to fail, but in the same vein having a “loyal” person on the target’s team support the conspiracy should increase the odds of success.
5. Risk – both plotters and target should have skin in the game, with potential consequences for both depending on how the uncertain outcome is resolved.

A simple mechanic

For a modern game like Watch the Skies, if you wanted an approach that simulates reality more than Hollywood movies, you could say assassination attempts have no effect on democratic states (at most you might ask the player to sit down for a minute as their replacement character is sworn in), but for an autocratic state roll a d20:

• Resources: A special action card, special agent, and/or a resource token.
• Secrecy: If ten minutes are spent preparing the attempt, then on an odd numbered roll, the target is informed who commissioned the assassination attempt, and on even number they are left to guess
• If less than ten minutes is spent on preparing the attempt, the target is always informed who commissioned the assassination attempt
• Conspiracy: Add +5 to the roll if you are supported by a conspiracy inside the target state (alternately, roll 2d20, use the best result)
• Betrayal: Change to the conspiracy bonus to -4 if the plot is betrayed before the action is resolved (alternately, roll 2d20, use the worst result).
• Risk: On a roll of 1-15, there is no effect
• On a roll of 16-19, the target is assassinated, and the player must sit out for a minute, and then resumes play in the same role. Repressive security crackdowns mean that future assassination attempts in that state later in the game only succeed on a roll of 20.
• On a roll of 20, the target is assassinated and the player must be given a new role in the game. Another player on their team must take up their old role.

You could also flip this around, and have the target roll the die, with 1 and other low numbers being the untimely demise of victim. This might allow them the use of any “luck points” or similar resources, and gives them more involvement in the resolution.

A more complex mechanic

For the Barracks Emperor megagame design that I am working on, the assassination mechanics need to reflect the history of that era of Roman history.

First, this is a period of history with a rapid turnover in Emperors, and many failed revolts. It averages out to about one new Emperor every two years between about 235 and 285 CE, with no one managing to create a stable dynastic succession until Diocletian set up the Tetrarchy. Not including character death is a disservice to the era.

Second, There are essentially two main ways that Emperors are murdered. The first way is a decision made just before a battle between rival armies. It was not uncommon for junior officers desiring promotion, or senior officers wanting to collect pensions, to collectively decide to murder an Emperor or Usurper if they thought their side’s chance of winning the battle was low. A bonus of preventing the battle with a quick act of murder, was preserving the Roman army for use against barbarians. The second way is a conspiracy among palace officials and aristocratic members of the Senate. These did not always succeed and often led to widespread purges of the Senatorial class. Which explains why part of the oaths a new Emperor swore before the Senate was an oath not to murder them.

Third, when a character is killed, the player will respawn with a new character somewhere else in the game. I did have people say I should just have an “heir” fill the role of a dead character – but its really clear in the source materials for this period that the heirs or relatives of deceased Emperors and Usurpers were either too young to succeed or were murdered as well. So having having the mutinous commander of a legion re-entering play as the commander of the same legion does not feel right for this megagame.

The KISS design principle suggests there should be one conspiracy mechanic to handle both forms of imperial murder. As part of the wider game system I plan to give each character their own mortality track. When a mortality check is required:

1. Roll 3d6.
2. If the roll is a 13, the character dies.
3. If the roll is 8-12, check that number. If it is rolled again in a future mortality check, the character dies.
4. If the roll is 3-7, or 14-18, the character survives.

So you pretty much have a 90% chance of surviving the first mortality check, but you are bucking the odds if you survive more than five such checks. Most characters will only test mortality if involved in battle, or in a province struck by disease, but Emperors will have to check at least once every game turn.

A conspiracy check is made (a) before any battle involving an Emperor/Usurper, or (b) if a conspiracy special action is submitted to control with any character named as a target (this will be a limited resource, budgeted to occur not more than once per game turn).

The conspiracy check is done like this:

1. For a battle, get all the characters with the army in one place. For a conspiracy special action, get all the Senator, Praetorian, and Official characters in Rome together.
2. Everyone closes their eyes.
3. Control asks everyone supporting the conspiracy to raise a hand. The Praetorian commander gets to raise both of their hands.
4. A majority supporting the conspiracy means the Emperor must make a mortality check.
5. A tie or minority supporting the conspiracy means no mortality check occurs, and the Emperor is told the name of one of the conspirators.

This is relatively straightforward, with a maximum of one die roll, but it should engage all the players involved. The main requirement is the time required to assemble all of the involved players, and some private space where the rest of the players cannot see what is happening. It will need rule clarity – should an absent player default to supporting or opposing a conspiracy, or not being counted at all?

Comments and criticism of the ideas in this post are welcome.