The aomstats website yes. The recorded game format is not designed for the kind of statistical analysis that we players would like to see, so there are going to be anomalies like that. Not to mention that you can easily drive your APM up like that by just spamming the same actions, and some actions count more towards the count I think.
I was confused by the fact that such anomalies only occur in some players, and not in everyone.
Yes, it is certainly curious, but I stand by my best guess that it’s not related to aomstats but rather the replay format or maybe the player behaviour.
It’s difficult to discern exactly what the cause is without parsing the specific replay file and inspecting what type of actions were being performed at that time. Most likely it was some sort of spamming shift click on unit icons, each of which may count for 5 actions (train 1 train 1 train 1 train 1 train 1) even though it’s only a single mouse click… so spamming 10 of those in 3 seconds may count for 50 actions in 3 seconds, not sure.
Also, it may be that that player is using some sort of mouse macro or keyboard macro or something, although probably not otherwise it would more likely be spread across the entire game.
I’m guessing it’s likely not aomstats but rather the replay format. For example, it doesn’t include actions such as adding/removing units to control groups, creating control groups, changing rally points, etc. So it’s already not really capturing the APM of the player. Not to mention it may be counting shift-click train commands as 5 actions rather than 1, and who knows what else.
Aomstats is doing the best it can with the replay format, but its APM counter isn’t that useful due to the replay format.
That’s my two cents on it
APM is general is not to be trusted, it can’t really distinguish “useful clicks” from just “spam clicks”. You can just spam “move to” right click 20 times or do it only 2-3 with the same result, and then it stops being a useful metric.