Idea for refining the "effective APM" metric (multiqueue eapm)#164

Only one suggestion per post, okay.

For multiqueue, I noticed that MrPlanner got credited with a very high eapm number in the lategame when using multiqueue, e.g. to make 3 halberdiers per barracks with 15 barracks. It seems like each unit created might be counted as one eapm?

A more practical way to count eapm would be to count a multiqueue input as just one action, i.e. 3 actions for queueing up 3 batches of halberdiers.

Implementation idea: Count at most 1 per 0.05 or 0.1 seconds from unit queueing.

@Iwer Sonsch geAPM doesn’t take into account the amount of units a player queues, but rather the queue commands themselves. Queueing 3 units and shift-queueing 15 units result in the same geAPM from my testing (which is higher than 3 geAPM if done in rapid succession because it’s an action frequency metric, not an action counter).
If you have an example game in which you feel the metric was misleading please feel free to share it with us :)

Oh, I don’t mean the shift modifier, I mean having multiple of the same building selected and queueing units in all of them using the “multiqueue” feature that became mainstream around the release of DE.

At 2:13:19 you can see MrPlanner (as MrSlow) making about 50 halberdiers in one minute, and his eapm jump from 20 to 67 even though during the game he didn’t really play much faster than before or after. A few minutes before that point we talk about the eapm counter mechanics and how they seem to work
https://www.twitch.tv/videos/2321400999
You can see him making the inputs at 1:57:17

Iwer Sonsch sorry, I wasn’t clear enough 😅
Shift queuing 25 units from 5 different buildings through 5 shift-queue commands across 5 seconds increases geAPM by the same amount as individually queuing 5 units from 1 building using 5 separate queue commands across 5 seconds. geAPM is a real-time frequency metric, not a retroactive aggregate. It’s similar to that old riddle “if it takes 5 elves 5 minutes to build 5 dolls, how many minutes does it take 100 elves to build 100 dolls?”
Hopefully it’s clearer now.

I’m not sure if we’re on the same page - queueing 15 units from 15 barracks only takes one single queue command, not 15 commands.

Shift-queueing 75 total units also only takes 1 shift-queue command, not 15 shift-queue commands.

This is because that singular command will be applied to every barracks in the selection, as long as there are enough resources.

You can see this in action at 1:57:22 in the vod I linked, here’s a clip for easier viewing: https://www.twitch.tv/mrplanneraoe/clip/SpunkyInterestingPlumagePoooound-tiPXsc_ThQ95tWeH As you can hear and see, each of his clicks queues up more than 5 units, and from more than 1 barracks each.

I believe this is called the “multiqueue” feature, and was made the standard with the release of AoE2:DE

Iwer Sonsch yes, we’re roughly on the same page. Trying to make my previous comment clearer now: Concurrently queuing 15 units from 5 selected buildings through 3 commands shouldn’t (and didn’t during my testing) increase our APM stat any more than queuing 3 units from a single selected building through 3 commands.

Interesting - maybe it works fine in singleplayer testing then but can run into issues in multiplayer?

Because in the twitch vod I shared, it really looks like MrSlow makes about 4 clicks to queue up about 40 halberdiers from 10 separate Barracks, but gets credited with about 40 effective actions for that.

Maybe something goes wrong when a player selects a combination of both finished and under-construction buildings at the same time and then gives a queue command?

Iwer Sonsch it should work the same on both modes. In my previous message, I said that either queuing 15 or 3 units should increase eAPM the same amount as long as 3 commands are used for each, this assumes that the speed at which these commands were issued is the same.

If the only actions a person makes is queuing 60 units with 60 commands 1 second apart of each other, they will have an eAPM of 60 (60 effective Actions Per Minute ), whilst instead queuing 150 units from 5 barracks with 30 multi-queue commands 2 seconds apart will have an eAPM of 30 (30 effective Actions Per Minute ). That is because the eAPM stat is a frequency metric, not an aggregate.

Let me know if it’s a clear-enough explanation :)

Thanks, yeah I think we agree on how the eAPM counter should be working. Could you check whether it actually did work that way in the game I shared, or whether there is a bug that can cause it to overcount in online games?

@Iwer Sonsch If it worked differently with online matches that’d be a game bug, however, from the clip you shared above, you can see towards the end 3 training commands issued within half a second or so.

While the playstyle in general might be slow, these types of command chains can temporarily spike one’s eAPM higher than normal since that’d be equivalent to 360 eAPM.

The source stream is no longer available and there’s no easy way for me to check from the rec, but let me know if you encounter more of these issues and I’ll take a look.

Thanks