The Beauty of the Mass Effect 2 Ending

The following essay was originally published on my Mass Effect Wiki blog and later reposted on TV Tropes. You may have also seen it in an academic paper form, which I published in December 2016.

Mass Effect 2 is the most universally adored part of the original Mass Effect trilogy, and despite its gameplay refinement and improved variety, Mass Effect 3 could never supplant its predecessor. Part two had something its sequel didn't: A mind-blowingly epic ending that left every player in awe. But what exactly made the ending of Mass Effect 2 so amazing? Although the magnificent cutscene direction and musical composition contributed a lot to it, the key to the awesomeness of Mass Effect 2's endgame segment lies in the gameplay design and the story scripting of the entire Suicide Mission.

Summary

The main appeal, driving force, and resource of Mass Effect 2 are the thirteen recruitable companions (squadmates) of Commander Shepard (which includes both the two DLC-only companions and the secret squad member Morinth). The entire game revolves around them, and the Suicide Mission is quite consciously a Final Exam Finale level, taking the form of a weird puzzle/deck-building game hybrid. The deck-building aspect comes from the fact that the outcome of the endgame depends largely on what "deck" (i.e. which squadmates) you bring with you to the Omega 4 relay. The puzzle aspect is owed to the complex but static set of rules that determine your success or failure in the Suicide Mission — rules that are pretty obvious if you pay attention to your squadmates beforehand.

A bulk of Mass Effect 2 concerns the composition and enhancement of your "deck". This largest part of the game is very options-driven, meaning that you can do any missions in any order, mostly without fear of immediate harsh consequences. This part ends with the first true major choice of the game that finalizes the state of the "deck", namely, the choice of when to enter the Omega 4 relay. The entire Suicide Mission boils down to (a mere) ten more choices that determine its outcome, as you are repeatedly forced to put your squadmates (including ones whom you might need later on) at risk of a Plotline Death in order to advance the mission. Your task is made easier by the fact that for every choice, there are at least three possible winning decisions — although, of course, you may not have all of them in your "deck" at that point. However, the game also refuses to hold your hand in solving this puzzle with obvious hints and instead makes you rely on common sense and prior knowledge of the individual squadmates' strengths and backgrounds to make ideal decisions.

The interchangeability of squadmates in certain story-relevant roles and, conversely, the dependency of said roles' outcomes (success or tragedy) on the suitability of the actors is what gives the strictly linear Suicide Mission its amazing variability. Even though it doesn't matter much, for instance, which loyal squadmate escorts the Normandy crew back to the ship, the option to choose one and the palpable consequence of that teammate being absent for the rest of the mission give the players an illusion of controlling their own story. Moreover, the multitude of winning scenarios guarantees that most players end up feeling like they have been told a personal and unique story, even though everyone starts the endgame with roughly the same "deck" and experiences exactly the same Narrative Beats along the way.

Pre-endgame variables

Your "deck" at the beginning of the Suicide Mission is effectively represented by three mathematical sets and two variables.

Recruited squadmates (R*)

This set is the base of all other pre-endgame options. There are 13 items available for this set (Garrus, Grunt, Jack, Jacob, Kasumi, Legion, Miranda, Mordin, Morinth, Samara, Tali, Thane, Zaeed) but it can only contain up to 12 at any time. Recruiting a squadmate generally takes place in the course of a single Dossier mission, but with several particularities:

• Two squadmates (Jacob, Miranda) don't have to be recruited, as you start the game with them.
• Three more (Garrus, Jack, Mordin) are mandatory, since it is not possible to enter the Horizon Plot Tunnel without recruiting them first.
• Two squadmates (Grunt, Legion) require manual activation after their respective recruitment mission before they can join as team members.
• Two more (Zaeed, Kasumi) require purchasing DLC and join immediately after a brief dialogue instead of a full-length recruitment mission.
• One squadmate (Morinth) can only be recruited by recruiting and then killing another during her loyalty mission (Samara).

Loyal squadmates (L*)

The set of loyal squadmates is always a subset of R*, therefore it likewise has 13 potential items but can only hold 12 at any time. Its membership is determined mainly by completing the so-called Loyalty Missions of the respective squadmates, with following particularities:

• Four loyalties (Samara, Tali, Thane, Zaeed) can be forfeited by making certain decisions in the respective mission.
• Two squadmates (Samara, Zaeed) can die during their respective loyalty missions, although in Zaeed's case, it is only possible after the Suicide Mission and is thus irrelevant to the endgame.
• One squadmate (Morinth) is loyal immediately upon recruitment.

Additionally, two pairs of squadmates (Jack and Miranda, Tali and Legion) are scripted to get into a confrontation after the loyalty of both characters in a pair is earned. Unless you are able to persuade them to calm down, you are forced to support one side in the argument, losing the other's loyalty.

Ship upgrades (U*)

This set is a mix of story-relevant upgrades, gameplay-relevant upgrades, and even purely cosmetic upgrades (Med-Bay Upgrade). Only the three story-relevant upgrades (Heavy Ship Armor, Multicore Shielding, Thanix Cannon) have impact on the course of the Suicide Mission. The availability of all upgrades is directly derived from R*, since they can only be unlocked for purchase by talking with a specific squadmate.

Romanced squadmate (F)

The romanced squadmate is a singular member of the conjunction of L* and the fixed set of romanceable squadmates determined by Shepard's gender (Garrus, Jacob, Thane for female; Jack, Miranda, Tali for male). Of these, only one romance for each gender (Thane, Jack) has any effect on the course of the endgame, and that effect is purely cosmetic.

Delay between the attack on the Normandy and the Suicide Mission (D)

Since there is no In-Universe Game Clock in the game, the time is measured by the number of completed missions between two events, specifically, the kidnapping of the Normandy crew by the Collectors and the start of the Suicide Mission. In practical terms, this delay can have only three values: short (at most one story mission or assignment), middle (one to three assignments and no story missions), and long (anything above that).

Pseudocode of the Suicide Mission

Before passing through the Omega 4 relay, two variable sets are initialized: the squad survivors set (S*) is initialized with the current contents of R* and the crew survivors set (CS*) is preset according to D (if short, all kidnapped crew members; if middle, half of the crew; if long, only Dr. Chakwas). The following pseudocode is a more-or-less formal representation of the endgame puzzle:

1. CUTSCENE: Passing through the Omega 4 relay:
   1. Victim1 = "Jack"
   2. IF "Heavy Ship Armor" not in U* THEN remove Victim1 from S*
2. PLAYER CHOICE #1: Two party members for the fight with Oculus:
   1. SELECT P1* from S*
3. CUTSCENE: Eezo core overload:
   1. Victim2 = FIRST( ( { "Kasumi", "Legion", "Tali", "Thane", "Garrus", "Zaeed", "Grunt" } in S* ) not in P1* )
   2. IF "Multicore Shielding" not in U* THEN remove Victim2 from S*
4. CUTSCENE: Taking on the Collector Cruiser:
   1. Victim3 = FIRST( { "Thane", "Garrus", "Zaeed", "Grunt", "Jack", "Samara", "Morinth" } in S* )
   2. IF "Thanix Cannon" not in U* THEN remove Victim3 from S*
5. PLAYER CHOICE #2: Specialist Technician in the vents:
   1. SELECT ST from S*
   2. IdealST* = { "Tali", "Legion", "Kasumi" }
6. PLAYER CHOICE #3: Fireteam Leader during Infiltration:
   1. SELECT FL1 from (S* minus ST)
   2. IdealFL* = { "Garrus", "Jacob", "Miranda" }
7. PLAYER CHOICE #4: Two party members during Infiltration:
   1. SELECT P2* from (S* minus { ST, FL1 })
8. CUTSCENE: Sealing the doors after Infiltration:
   1. IF (ST not in L*) or (ST not in IdealST*) or (FL1 not in L*) or (FL1 not in IdealFL*) THEN remove ST from S*
9. PLAYER CHOICE #5: Biotic Specialist to hold the bubble:
   1. Biotics* = { "Jack", "Jacob", "Miranda", "Morinth", "Samara", "Thane" }
   2. SELECT BS from (S* in Biotics*)
   3. IdealBS* = { "Samara", "Morinth", "Jack" }
10. PLAYER CHOICE #6: Fireteam Leader during the Long Walk:
    1. SELECT FL2 from (S* minus BS)
11. PLAYER CHOICE #7: Escort for the Normandy crew:
    1. SELECT ESC from (S* minus { BS, FL2 }) OR none
12. PLAYER CHOICE #8: Two party members for the Long Walk:
    1. SELECT P3* from (S* minus { BS, FL2, ESC })
13. CUTSCENE: Showdown with the Collector Swarm:
    1. Victim4 = FIRST( { "Thane", "Jack", "Garrus", "Legion", "Grunt", "Samara", "Jacob", "Mordin", "Tali", "Kasumi", "Zaeed", "Morinth" } in P3* )
    2. IF (BS not in L*) or (BS not in IdealBS*) THEN remove Victim4 from S*
14. CUTSCENE: Opening the doors for the fireteam:
    1. IF (FL2 <> "Miranda") and ( (FL2 not in L*) or (FL2 not in IdealFL*) ) THEN remove FL2 from S*
    2. IF (FL2 = F) and (F in { "Jack", "Thane" }) THEN use different last words
15. CUTSCENE: Checking up on the Normandy crew:
    1. IF ESC = none THEN CS* = none
    2. IF ESC not in L* THEN remove ESC from S*
16. PLAYER CHOICE #9: Two party members for the final battle:
    1. SELECT P4* from (S* minus ESC)
    2. Implicit choice (squadmates holding the line): HTL* = S* minus P4* minus ESC
17. PLAYER CHOICE #10: Destroying or purging the Collector Base
18. CUTSCENE: Escaping the Collector Base:
    1. Remove (P4* not in L*) from S*
    2. Remove HTL_Death_Calculus(S*, L*, P4*, ESC) from S* — see Appendix A for details
    3. IF COUNT(S*) < 2 THEN Commander Shepard dies
19. CUTSCENE: Dialogue with the Illusive Man (changes depending on whether Shepard survives and whether Collector Base was destroyed or purged)
20. ENDING:
    1. IF S* = R* THEN "No One Left Behind" AND "Against All Odds" AND "Mission Accomplished"
    2. IF (S* <> R*) and (Shepard lives) THEN "Against All Odds" AND "Mission Accomplished"
    3. IF Shepard dead THEN "Mission Accomplished"

Analysis of the endgame puzzle

As already mentioned, the endgame puzzle (the Suicide Mission) is all about intelligent application of the resources the player had a chance to collect throughout the game. These resources can be abstracted into two types: Primary (squadmates and their loyalties) and Secondary (Normandy upgrades). With that in mind, the puzzle itself can be then abstracted into three phases:

• Phase I (beats 1 through 4) consists mostly of cutscenes and evaluates the player's initiative and readiness to go beyond the explicit instruction in preparation for the Suicide Mission—namely, to discover and to acquire Secondary Resources without an obvious indication to their usefulness. It punishes the insufficiently prepared players by handicapping them in phase II via the loss of some Primary Resources, but, ingeniously, makes this seem like a Plotline Death by denying the player any immediate agency in it, as well as giving only vague hints at how it could have been prevented. To further throw the player off-balance, loyalty, while essential everywhere else, provides no protection to crucial Primary Resources in this phase.
• Phase II (beats 5 through 16) is a Final Exam Finale that repeatedly tests the player's common sense, as well as their knowledge of the setting, of the squadmates' backgrounds and personalities, and of the basic game rules. Failure in these tests leads to further losses of the Primary Resource, negatively impacting the player's capacity to overcome later tests, and ultimately, to an inferior ending.
• Phase III (beats 17 through 20) once again consists mostly of cutscenes that rate the player's performance—only this time, they evaluate both the pre-endgame progress and the phase II decisions. The game then awards the player with one of three possible endings.

Conclusions and discussion

The following section was not part of my original blog post but was added a few years later when I rewrote it for a game studies paper.

Having broken down the Suicide Mission into its abstract constituents of resources, beats, and phases, we can now attempt to extract general recommendations for the gameful narrative design pioneered by the creators of Mass Effect 2.

• Endgame structure. The Suicide Mission clearly follows a cinematic three-act structure, with the first act/phase mostly introducing complications outside of the protagonist's (immediate) control, the second, longest act consisting of confrontations over continuously rising stakes, and the final act containing the climactic resolution wherein the protagonist either lives or dies (beat 18), depending on the player skill. Each act contains a number of story beats that alternate between player decisions that advance the narrative in meaningful ways and cutscenes that provide regular feedback on the efficacy of said decisions, creating a ludic feedback loop between the player and the game within the constraints of a cinematic narrative.
• Transfer of learning. To reach the best ending, the player is expected to repeatedly go beyond explicit instructions given by the game. The three endgame-relevant Normandy upgrades cost valuable in-game resources, but the player is never explicitly informed of their true purpose beyond vague hints that every investment may count during the endgame. Likewise, in the choice of the Specialist Technician, the Fireteam Leaders, and the Biotic Technician, the player is expected to deduce, despite the lack of an explicit skill ranking in the game, which squadmates are inherently better at certain tasks for reasons that are only explained in the "flavor text" and do not impact the core combat gameplay.
• Failing forward. Apart from the "game over" scenarios in the core combat gameplay, it is impossible to fail the Suicide Mission, as the game contains no cutscenes or dialogue pertaining to Commander Shepard’s failure to defeat the Collectors. The characters thus always succeed at both their individual tasks and the overarching objective. Dramatic tension is maintained, however, by the question of the price they will pay for their success, since every wrong choice quickly results in at least one fatality. While the success of the mission is predetermined by the game, its magnitude is entirely up to the player. This allows the game to pick an ending based on a simple tally of the survivors, representing one of the most ludic approach to concluding game narratives.
• Static roles, variable actors. The key to combining gameful experience with tight narrative scripting in Mass Effect 2 is the decoupling of its key roles from the specific characters (with the notable exception of Commander Shepard). This approach has several far-reaching implications:
  • The narrative theme determines the specific roles. In Mass Effect 2, the theme is infiltration and sabotage, so the roles represent areas of expertise needed on such a mission (technology, leadership, biotics, and combat). Another major theme is dedication to the cause, so loyalty is a consistent factor after phase I and even exclusively determines the characters' fate in beats 15 and 18.
  • In order to make role assignments meaningful, a larger actor pool (i.e. many more squadmates) is required than is strictly necessary to fill each role. Doing so, however, also allows for a highly modular narrative built around a static underlying structure, especially if the ideal actors for each role number more than one but much less than the total.
  • Despite this, all actors are involved in the narrative at all times, even if only off-screen. Particularly the "hold the line" sequence in beats 16 through 18 gives narrative significance to those squadmates who have not been assigned a particular key role up to that point.
  • Fully defined non-randomized priority queues that determine who dies and who lives in each cutscene allow the developers to manage the high dimensionality of the possibility space without reducing it to a "choose your own adventure"-style branching tree.
• Design variety. Mass Effect 2 is built upon a handful of design patterns that underlie most of its content. The pattern for squadmates' character arcs, for example, is "Shepard receives a dossier, recruits the squadmate in one mission, then gains their loyalty in another", while the pattern for story beats during the endgame is "Shepard assigns a squadmate to a task, they complete it and either survive or die". While learning these patterns allows the player to gain confidence in their skills faster, they can also make the game too predictable, so the developers go out of their way to deviate from the formulas in almost every instance. To name a few specific examples:
  
  • Some characters need additional interactions after their recruitment mission before they join as squadmates. One of them is "recruited" in the course of a main story mission.
  • Some loyalty missions may not secure squadmates' loyalties, some loyalties may be lost again in post-mission arguments, and one loyalty mission even doubles as both recruitment and loyalty mission for a secret squadmate who replaces a previously recruited one.
  • A small subset of Secondary Resources decides the Primary Resources' fate in the endgame phase I, regardless of loyalties, introducing the initial complications to the plot.
  • In phase II, player choices are clustered in groups of three and four at the start of the Infiltration and the Long Walk sequences, respectively, and their ramifications are displayed in a sequence of cutscenes punctuated by core combat gameplay.
  • Even an ideal Specialist Technician dies if paired with a non-ideal Fireteam Leader; likewise, even loyal party members can perish defending a non-ideal Biotic Specialist.
  • The Escort role does not have to be assigned, unlike all others: leaving it unassigned preserves the squad's strength, but results in the death of all crew survivors.
  • If the second Fireteam Leader perishes, they exchange characterization-laden final words with Shepard; Miranda, however, cannot die at any point of the Long Walk sequence.
  • Mordin and Thane have no purpose during the Suicide Mission: they are neither ideal choices for any particular roles (except the Escort, if loyal), nor confer any significant advantage during the "hold the line" sequence, even if loyal.

Mass Effect 2 was, perhaps, the most innovative video game ever produced by BioWare, and many of the design patterns it pioneered or improved upon have not been revisited since, by them or by other developers. It is the author’s sincere hope that this analysis will inspire more video game developers to produce gameful narratives in the manner of Mass Effect 2, armed with the understanding that its success was not a random fluke but a result of skillful and deliberate iteration on a few relatively simple design and content patterns.

Appendix A: Hold the Line death calculus

The following addendum was not part of my original blog post but is a summary of the research subsequently carried out by me and my fellow tropers.

The method for determining which squadmates die during the Hold the Line (HTL) section of the Suicide Mission (step 18 above) is an example of Algorithmic Story Branching, wherein the course of the story (who lives and who dies) is determined according to a more or less complicated mathematical formula, rather than writer's discretion or simple true/false event flags. In case of the HTL section, the developers assign each squadmate a numeric stat that represents their effectiveness in holding a fortified position against continuous assault. This "HTL score" is hidden from the players, but can be easily inferred from the individual squadmates' characterization and dialogue. Each squadmate starts the game with an HTL score ranging from 0 to 3, which is increased by 1 if their loyalty is secured and not subsequently forfeited (see above):

The computation in step 18 takes into account both the average HTL score of all living squadmates left behind in step 16 and their total number, both of which can be highly variable at this point. In the following, the former number (average score) will be referred simply as "HTL", and the latter (number of fighters), as "Strength".

First, the game computes the total number of casualties among the defending squadmates according to the following table. Generally speaking, if the average HTL score is 2 or higher, then everyone survives; otherwise, there is at least one casualty. Note that because only four squadmates can have an HTL score of 0, an average score of 0.0 is impossible with 5 or more defenders.

Second, once the number of casualties is determined, the game uses a hidden priority queue to select the specific squadmates from among those holding the line who are killed in action, which goes as follows:

any non-loyal squad members > Mordin > Tali > Kasumi > Jack > Miranda > Jacob > Garrus > Samara/Morinth > Legion > Thane > Zaeed > Grunt

Note that for the purpose of this calculation, all non-loyal squadmates count as a single casualty, so it is entirely possible for more characters to die than indicated in the table above. For a discussion of this system and of individual squadmates' HTL scores and placement in the KIA priority queue, refer to the Mass Effect 2 analysis page on TV Tropes.