Case Study 1: XCOM --- The 95% Miss and the Perception of Fairness

The Shot Heard Around the Internet

There is a screenshot that every game designer has seen. It has been posted to Reddit, Twitter, gaming forums, and design talks hundreds of times. The screenshot shows an XCOM soldier standing directly next to an alien --- close enough to touch it --- with a displayed hit chance of 95%. The player took the shot.

The soldier missed.

The player lost the mission, lost the soldier (permanently, because XCOM has permadeath), and lost their faith in random number generators. The post went viral. The comments were unanimous: "This game is broken." "The RNG is rigged." "95% should be a guaranteed hit."

Mathematically, there is nothing wrong. A 95% hit chance means a 5% miss chance. That is one in twenty. If a million players take a 95% shot, fifty thousand of them will miss. Some of those fifty thousand will miss at the worst possible moment --- when the mission is on the line, when the target is one hit from death, when the soldier who missed was the player's favorite. Those are the players who post the screenshot.

But here is the fascinating part: the players are right that something is wrong. Not with the math. With the design.

XCOM is one of the most important case studies in randomness design because it demonstrates, more clearly than any other game, the gap between mathematical correctness and player experience --- and because the developers at Firaxis spent years engineering solutions to close that gap.

Why 95% Feels Like 100%

Human beings are catastrophically bad at intuiting probability. This is not a game design problem. It is a cognitive architecture problem. Our brains did not evolve to process probabilities. They evolved to process narratives: "I aimed at the alien. I was right next to it. I should hit it."

The psychological research is consistent: people treat probabilities above 90% as certainties and probabilities below 10% as impossibilities. A 95% chance feels like "this will definitely happen." A 5% chance feels like "this will never happen." When the unlikely outcome occurs, it does not feel like the normal operation of probability. It feels like a violation of the rules.

This is compounded by two cognitive biases:

Negativity bias: Bad outcomes are psychologically weighted more heavily than good outcomes. A player who hits twenty consecutive 75% shots does not think "I have been lucky." They think "the game is working correctly." When they miss the twenty-first, they think "the game just screwed me." The hits are expected; the miss is an event. Twenty positive experiences are invisible. One negative experience defines the entire perception.

Availability heuristic: We judge the frequency of events by how easily we can recall examples. A dramatic miss at a critical moment is highly memorable. The hundreds of shots that hit as expected are forgettable. Over time, the player's memory of XCOM becomes a highlight reel of devastating misses, even if the actual miss rate was exactly as advertised.

💡 Intuition: Imagine a perfectly fair coin. You flip it a hundred times. It comes up heads 50 times and tails 50 times. You would not remember any specific flip. Now imagine one of those tails flips caused your favorite soldier to die permanently. That single flip is now the defining experience of the coin. Your emotional memory says the coin is unfair, even though your rational mind knows it is not. This is the designer's dilemma: the math is right, but the experience is wrong. And the player lives in the experience, not the math.

XCOM Cheats --- For You

Here is the design secret that Firaxis eventually acknowledged publicly: on lower difficulty settings, XCOM lies. It lies in the player's favor. Extensively.

On Easy and Normal difficulty in XCOM: Enemy Unknown (2012), the game applies hidden modifiers that increase the player's effective accuracy beyond the displayed number:

• Aim bonus after misses: Each consecutive miss by the player's soldiers adds a hidden +10% bonus to subsequent shots. Miss twice in a row, and your next shot is secretly 20% more accurate than displayed. This bonus resets when a shot hits.
• Damage reduction after hits taken: When a soldier is wounded, subsequent enemy attacks against that soldier have a hidden accuracy penalty. The game protects injured soldiers from being eliminated by a second hit.
• Critical hit suppression for enemies: Enemy critical hit chances are silently reduced on lower difficulties. The displayed stats show one number; the actual calculation uses a lower one.

On Classic and Impossible difficulty, these hidden bonuses are removed. The displayed percentages are accurate. This is why veteran XCOM players who move from Normal to Classic often report that the game "suddenly feels unfair" --- they were unknowingly benefiting from invisible safety nets that have been removed.

🎮 Design Spotlight: Firaxis designer Jake Solomon discussed this openly in post-launch interviews. His reasoning was straightforward: the game's fantasy is that you are a tactical genius commanding an elite military unit. A tactical genius does not miss point-blank shots. When the random number generator produces that outcome, it breaks the fantasy. The hidden bonuses maintain the fantasy by quietly adjusting probabilities to match the player's expectation. The player feels like a competent commander. The math cooperates. The experience is preserved.

This approach is not unique to XCOM. Many games "cheat" in the player's favor:

• Fire Emblem (modern entries) uses a two-roll system: it rolls two random numbers and averages them. This compresses the distribution toward 50% --- high hit chances are slightly higher than displayed, low hit chances are slightly lower. A displayed 90% is actually ~97%. A displayed 30% is actually ~16%. The system makes likely outcomes more likely and unlikely outcomes less likely, matching player intuition without changing the displayed numbers.
• Left 4 Dead dynamically adjusts zombie spawns based on player performance (the "AI Director"). A struggling team faces fewer zombies. A dominant team faces more. The game is constantly cheating to maintain the desired tension level.
• Celeste uses coyote time and input buffering (discussed in Chapter 8's Case Study 1) to make jumps succeed that "should" have failed by strict physics.

The common principle: the displayed system and the actual system are different, and the difference always favors the player's experience.

The XCOM 2 Solution: Showing the Math

XCOM 2 (2016) took a different approach to the perception problem. Instead of only hiding bonuses, it increased transparency:

1. Hit chance breakdown: The tooltip now shows why the hit chance is what it is. "Base accuracy: 70%. Range bonus: +10%. Flanking bonus: +20%. Cover penalty: -25%. Total: 75%." The player can see the factors. This transforms a mysterious percentage into a system they can understand and influence.

2. Overwatch cone visualization: When a soldier enters overwatch (reaction fire), the game shows the cone of fire. The player can see which tiles are covered and estimate the interception probability. Information replaces guesswork.

3. Damage range display: Instead of showing a single damage number, the game shows the range: "4-6 damage." The player knows the outcome will vary and can plan for the worst case.

4. Percentage coloring: Hit chances are color-coded. Green for high probability, yellow for moderate, red for low. The color provides an immediate gut-check: "Is this shot worth taking?"

📝 Design Note: The transparency approach addresses a different part of the problem than hidden bonuses. Hidden bonuses fix the outcome. Transparency fixes the understanding. A player who sees "75%: Base 65% + Elevation 10%" understands that the number is not arbitrary. It is a calculation based on their tactical decisions. When they miss, they are more likely to think "I should have flanked to get +20%" than "the RNG is broken." Transparency converts frustration into learning.

Lessons for Your Design

The XCOM case study teaches four principles that apply to any game with output randomness:

1. Displayed Probability Must Match Perceived Fairness, Not Mathematical Accuracy

If 95% feels like a guaranteed hit, you have three options: never show 95% (cap the display at 90%), make 95% functionally guaranteed (hidden bonus), or remove the number entirely (show "Very High" instead of 95%). What you cannot do is show 95%, deliver a 5% miss, and expect the player to respond rationally. They will not. They are human.

2. Agency Mitigates Randomness Frustration

The player's frustration with an XCOM miss is proportional to their perceived helplessness. If the player positioned their soldier carefully, chose the right weapon, flanked the target, used a scope, and still missed --- the frustration is intense because they feel they did everything right. But if the miss happens and the player has a backup plan (another soldier who can take the shot, a grenade as a guaranteed-damage alternative, an ability that does not roll to hit), the miss becomes a setback, not a catastrophe. Give the player tools to manage randomness, and randomness becomes a challenge to overcome rather than a punishment to endure.

3. Stakes Determine Tolerance for Bad Luck

A 95% miss on a routine shot against a weak enemy in the middle of a mission? Mildly annoying, quickly forgotten. A 95% miss on the final shot of a final mission that causes a beloved soldier to die permanently? Rage-inducing, never forgotten.

The stakes of the moment determine how much randomness the player will tolerate. Low-stakes moments can absorb high variance. High-stakes moments demand either low variance (high probabilities, hidden bonuses) or guaranteed options (abilities that cannot miss, items that provide certainty).

4. Feedback Determines Whether Randomness Feels Fair

When an XCOM soldier misses, the game shows the bullet trajectory. The bullet almost hits the alien. It passes inches from the target's head. This visual feedback transforms "the game rolled a miss" into "the bullet went slightly wide." The miss feels like a near-miss rather than an arbitrary failure. The feedback converts a mathematical event into a narrative event, and narrative events are emotionally digestible in a way that mathematical events are not.

Compare this to a game that simply displays "MISS" in red text. Same outcome. Completely different emotional response.

🎯 Key Insight: XCOM proves that the randomness problem in game design is not a math problem. It is a psychology problem. The math works perfectly. The player's brain does not process math. It processes stories, expectations, and emotions. Your job as a designer is not to make randomness mathematically fair. It is to make randomness feel fair. Those are two different problems, and they require two different solutions.

Further Exploration

Play XCOM 2: War of the Chosen on Commander difficulty. Note every moment where you feel the randomness was "unfair." Then replay the same mission on Rookie difficulty. Take the same shots. Compare your experience. The difference is the hidden bonus system at work --- and it demonstrates that the "fairness" of randomness is a designed experience, not a mathematical property.

For design analysis, compare XCOM's output randomness to Into the Breach's deterministic approach. Into the Breach shows the player exactly what every enemy will do and exactly what every player action will accomplish. There are no dice rolls. Every decision is perfectly informed. Does the removal of randomness make Into the Breach better or worse than XCOM? The answer depends entirely on the experience each game is trying to create --- and understanding that answer is understanding randomness design.