Actions per Minute

From Liquipedia StarCraft Brood War Wiki
Jump to: navigation, search

"Actions per Minute" (APM) is a measurement for the speed at which a user is able to interact with the game.

General Information[edit]

Even though APM is a known basic term for almost all Real Time Strategy Games, there is no standardized algorithm. APM can be measured in various ways, depending on which kind of actions are recorded to calculate the ratio. Generally speaking, APM measures any kind of command performed within the game. This includes selection of units or buildings, movement commands, selection and assignment of hotkeys and related commands. Furthermore, the higher a player's APM is, the higher his potential to fulfill all necessary and important tasks.

Since the APM is not standardized the meaning and interpretation of APM numbers is one of the most discussed topics within the Brood War community. Other algorithms like the so-called "effective APM" (eAPM) are often used as alternatives. More importantly, the link between a player's "skill" and the APM ratio are subject to interpretation as well.


Measuring APM[edit]

The Actions per Minute can be calculate post- or ingame. While calculations post-game (via replays) was popular in the earlier days of Brood War, the latter options, calculating the APM Ratio during a game only received more attention once Launchers became a standard tool for competitive play.


Replay Analysis[edit]

There are different programs available to measure APM. The two most important ones will be discussed. All of the programs calculate the APM Ratio with the game speed fastest.

BWChart[edit]

The standalone programme "BWChart" was one of the first tools available to manage replays. It does not only allow players to read out interesting figures of a replay, such as Build Order, Upgrades researched and Hotkey configuration, but also introduced the APM to the Brood War community. The programme also shows the APM graphs of players if selected. In addition, it offers more details of the APM - it categorizes actions as either "Macro-APM" (commands performed in order to "macro manage") and "Micro-APM" (commands performed in order to "micro manage").

bwRepInfo[edit]

Progression of APM measured for two players (red and green traces) over the course of a ~21 minute game

The Russian tool was widely unknown to the players until the later days of Brood War. It is the most popular programme used to determine a player's eAPM Ratio in addition to the APM ratio.


APM vs. eAPM[edit]

EAPM measured over the same game as above. Note some increases in activity during the later stage of the game, peaking for the eventual victor (red trace), likely due to execution of a final push or attack of the opponent (green trace)

The algorithms used by bwRepInfo and BWChart share some characteristics. For instance, both tools calculate the APM ratio based on the game speed fastest and rule out the initial 150 seconds of a game. Commands performed in these initial minutes will not be used to calculate either APM or eAPM.

The eAPM calculation is based on the assumption that some commands are "redundant". This is a strong assumption, as it postulates there would be an universal effective way to handle the game and supposedly "messy" spam commands are only performed in order to inflate a player's APM. However, it should be noted that some redundancy is needed for online games, especially when small lag spikes occur; performing the same command repeatedly decreases the chances of commands not being actually performed. Moreover, the description and notation of eAPM is only vaguely given by bwRepInfo's homepage. As consequence some strategies support a higher eAPM ratio, such as Mutalisk Harassment, as the constant movement done in order to force the Mutalisks to stack would not be counted as "redundant". In contrast to this, moving a larger ground force in a lagging game might be very well considered as redundancy.

The bwRepInfo's developers claim to have created an overall "better" algorithm therefore should be interpreted with great care, as it is only one out of many options to measure the APM.

BWChart itself offers an alternative to the eAPM in its own programme, the so-called "Valid APM" (VAPM). This leaves out "impossible" commands, such as constructing a building when not enough resources are available. This number usually does not vary greatly and is very close to the APM. Therefore the VAPM never beat the more popular eAPM.

Differences in Calculation
BWChart's APM[1]


Recorded, General APM
  • Selection of Unit/Building
  • Moving of Units/lifted Buildings
  • Attack, Stop, Hold Position, Patrol, Harvest commands
  • Casting of a spell
  • Training of units
  • Researching technologies or upgrades
  • Summoning a (Dark) Archon
  • Morphing a Lurker
  • Building an add-on
  • Morphing a Creep Colony into a Sunken/Spore Colony
  • Hotkey Selections


Not Recorded, General APM
  • Clicks on the minimap
  • Clicks on the map not selecting a unit, building or resource
  • F2, F3, F4 keys
  • Pressing keys not used by the game
Recorded, Micro APM[2]
  • Selection of Units via mouse or hotkeys (including shift select)
  • Attack, Move, Hold Position, Patrol, Return Cargo commands
  • Set and clear rally commands
  • Stimpack activation, lifting of buildings, (un)siege, lay Spider Mines, (de)cloack, ComSat scans
  • (un)burrow
  • all spells


Recorded, Macro APM[2]
  • Select Building via hotkey or mouse
  • Build, morph, warp, hatch and evolve commands
  • (cancel) train, research or upgrade
  • merge (Dark) Archon)
bwRepInfo's eAPM[3]
Categorized as redundant actions
  • Training, building or researching units, structures or upgrades, when not enough resources are available
  • Creation of unused hotkey groups
  • Selection of units or buildings without giving commands
  • Selection of an already used group with a hotkey
  • Using move, attack or rallypoint repeatedly on the same spot
  • Very fast reptition or duplication of commands


Live APM[edit]

APM can also be calculated during an ongoing game. To display this information a player needs a launcher and special plug-ins. One of the most popular plug-ins offering this feature is APMAlert. Most other plug-ins are based on it.

It should be noted that APM calculated during a game will almost always differ greatly from the APM ratio calculated via replays. For instance lag spikes will be shown directly. The more important reason for the alternating numbers are found in the intervals used for the APM. While replay analysis tools calculate the APM based on the entire game duration, live plug-ins calculate mostly in sixty second intervals.


APM and Skill[edit]

Most players agree that APM and skill are linked. However, the causality and the quality of the correlation between both is open for interpretation. The majority agrees that a certain threshold of APM is required in order to perform the needed commands. Yet, there are different play styles, which can be followed. Several of them need a higher APM ratio, others do not.

Furthermore, skill depends on other variables as well, most notably experience. At some point the APM most likely does not matter, if the initial threshold was met. Hence, skill does most likely not need a very high APM.


References[edit]