Please submit corrections by e-mail.


  1. Page xv: "contribution of the" → "contribution of art to the"
  2. Page xx: "require more" → "requires more"
  3. Page 6: Pac-Man should not be italicized; it is the character and not the game referred to
  4. Page 117: "Adobe's allows" → "Adobe Illustrator"
  5. Page 140: "next chapter" → "chapter 9: Balance"
  6. Page 172: P(A or B) = P(A) + P(B) for mutually exclusive events, e.g., rolling a 1 or a 2 on 1d6. The correction is: Remove the "P(A or B)" rule from this location; change "all three" → "both". At the bottom of the page, add: "One way for events to be dependent is for them to be mutually exclusive. This means that at most one of them can occur. For example, a six-sided die roll may result in the events "x = 1" or "x = 2", but not both simulaneously. For mutually exclusive events, P(A or B) = P(A) + P(B)." [Tim Stellmach]
  7. Page 174: the expected value of 1d100 is 50.5 (it is listed as 55.5 in the text) [Rehaan Vij]
  8. Page 206: "mixed strategy of choose" → "mixed strategy of choosing" [Brian Hatleberg]
  9. Page 234: The second equation on the page should be "P(S) = max(min(P(H) - X, 19/20), 1/20)"
  10. Page 234: The second table should be:
    WeaponE[D]Damage per Attack (E[N])
    at P(S) = 15/20
    Cost per Damage
    crossbow5.5  4.680.52
    Beretta 9mm7.0  5.600.35
    M16A2 5.56mm9.0  7.200.45
    frag. grenade5.510.500.70
  11. Page 235: "P(Cxbow|S) = 2/5 = 0.40" → "P(Cxbow|S) = 2/15 ≈ 0.133" [Rori Brown]
  12. Page 235: "P(Cother|S) = 1/5 = 0.20" → "P(Cother|S) = 1/15 ≈ 0.067" [Rori Brown]
  13. Page 235: The 2nd and 3rd lines of the equations should be "= 0.75 * 5.5 * (0.133 + 1)" and "= 4.68 (hit points / attack)" [Rori Brown]
  14. Page 235: "is from 0.39 to 0.70, so they are" → "is from 0.35 to 0.70, so if they are each used once on a mission, then they are"
  15. Page 236: At the end of the first paragraph, add "Of course, a grenade is used once to attack multiple opponents and the other weapons used many times on single opponents, so considering the full cost of each versus the expected damage from a single attack is questionable. In a campaign that tracks ammunition useage, we should count the incremental cost per attack and ignore the cost of the weapon. In a campaign with infinite ammunition, we should consider the expected or maximum number of attacks likely on a single mission and compare that to the cost of healing."
  16. Page 243: "like a buildings" → "like buildings" [Brian Hatleberg]
  17. Page 299: "diffuse color of black" → "emissive color of black"
  18. Page 487: "cannon lists" → "canon lists" [Eric Haines]
  19. Page 517: "moddability" → "modability"; "poor controls simply" → "poor controls"

Clarifications & Notes

Things we plan to consider or include for the next edition:
  1. Page 12: "English" should be "literature" to be more inclusive of other nationalities; "liberal arts" should be "humanities" [Olle Bälter]
  2. Page 140: As first observed by Jonathan Schaffer, assuming an optimal rational opponent is a stragegy for reaching the Nash equilibrium (which is a draw in a fair game), not a strategy for winning. When the opponent is actually imperfect, a stronger strategy may be predicting opponent errors correctly and factoring those into a non-deterministic decision tree. This is especially strong in betting (N.B., not bidding) games like Poker, where the player can voluntarily increase the payoff for a risky move.
  3. Page 142: Player 1 is maximixing and player 2 is minimizing; later pages reverse this convention. Since the strategy is called "minimax", consistently have player 1 minimize in the text.
  4. Page 144: "We use plot graphics": footnote to make clear that we mean plot in the literature sense, not plot as in a curve visualizing data.
  5. Page 197: Expand the concept of Nash Equilibrium and explain it in more detail, with examples [Brian Hatleberg]
  6. Revisit relevance of figures 2.5, 6.6, 11.3, 11.7, 13.4, 15.1, 18.2, 18.11, 18.13, 18.14; either extend the captions and explanation or consider replacements better matched [Brian Hatleberg]
  7. Page 245: Game developers use the term "Entropy" to describe the removal of harnessable "energy" from a game, drawing it towards completion. The term is confusing because the players are applying work to increase the order of their pieces; disorder isn't increasing in the system. Propose another term, like "countdown" and/or explain how the scientific concept of entropy applies here. [Brian Hatleberg]
  8. Page 514: The GBA versions were heavily influenced by Jet Grind Radio, which was also made by developer Vicarious Visions. The PC and console versions by Neversoft were instead heaviliy influenced by the Top Skater arcade game and were intended to be an improvement over Street Boarders (specifically, the prerelease version of the port marketed by EA as Street Sk8ter). [Personal communication w/ Christer Ericson May 2, 2010]