Game Development Reference
4. Either as a percentage or as X occurrences out of Y attempts.
5. This is play testing. The testers are playing the game, not testing the game.
6. Run the following series of tests in two-player mode. The tester should control only one
character, so the other character just stands there and doesn't fight back.
a) Count the number of punches it takes for the suspect character to kill the default
b) Count the number of kicks it takes for the suspect character to kill the default character.
c) Count the number of punches it takes for the default character to kill the suspect
d) Count the number of kicks it takes for the default character to kill the suspect character.
Repeat these tests several times, using three or four different characters against the suspect
character. If the suspect character consistently needs more punches or kicks to kill an oppo-
nent that the average of the other characters, the suspect character may be underpowered.
1. The answer is specific to the reader.
2. Incorporating the different control possibilities into a single table minimizes the total
number of tests. While separating the controller functions from the option parameter tests
would produce more test cases, it may cost less in terms of test execution. If the team is
incrementally adding features, the separate tables would be useful to establish the function-
ality of the game parameters and then when the controller functions are expanded, you
could start running the controller tests. If these two concepts are blended into a single table
from the start of the project, many tests might be blocked and you would be unable to pro-
vide valuable feedback for certain pairs in the test table.
3. Representing the “snap back�? behavior on the TFD requires a state to represent you
avatar at the starting location and another state to represent you standing at a gun or
ammo location. A MoveToGun flow would take you from the PreMatch location to the
“standing�? location. A flow with a PrematchTimerExpires event would take you from
your standing location to the NoGunNoAmmo state, accompanied by an action describing the
“snap back�? to the starting position. For the case where you don't move from the initial
spawning location, add a PrematchTimerExpires flow from the PreMatch location to
NoGunNoAmmo but without the snap back action.
4. Different firing methods would be represented by additional flows in parallel with the
existing ShootGun and ShootAllAmmo flows. They would simply need distinct names and flow
numbers. With more than one shooting control connected at the same time, it would be
possible for a player to use both simultaneously. If this is blocked by the game code, then
simultaneous shooting flows would loop back to their present states. If it is permitted, then
another parallel flow is required wherever shooting flows already exist. The possibility of
different controls provides Configuration trigger coverage. The ability to fire simultane-
ously from multiple controls provides a Stress situation.