I've seen several contradictory posts about whether auto-upgrade and auto-conquer work continuously during idle time (i.e. providing benefits that accrue during time idling) or whether they only upgrade/conquer all at once when the user signs back on. I decided to test this empirically, and it was easier to do this with auto-upgrade since the results are a little clearer.

I started Africa, played for a couple minutes, and began idling with a single L2 army camp producing 44.2 armies/sec (and about 4K armies in hand). I enabled auto-upgrade and disabled auto-conquer (so I could easily count armies). When I returned after 8 hours 20 minutes (= 500 minutes) idle time, the army camp upgraded to L5, producing 216 armies/sec, and I had 1.33M armies in hand. Within rounding errors, I had earned 1.33M armies = (44.2 armies/sec) * (500 min) * (60 sec/min). This indicates that none of the three upgrades took effect until the game restarted. (If the camp had upgraded during the course of the idle time, the first upgrade would have taken place about 10-20 minutes into the idling time, so total number of armies would have been significantly higher.)

I haven't explicitly tested auto-conquer (it's a little harder since conquering mines, army camps, etc. changes the rates of increase of things, and conquering caches gives instantaneous amounts of resources), but I would assume it would behave the same way. It would therefore appear that enabling auto-upgrade/auto-conquer have no effect during idle time (and only take effect all at once when the user reopens the level).

Thanks for testing that out. That's been my experience with auto craft as well- any time I'm away long enough to trigger auto crafting, it starts something at 0% when I get back. It makes auto advancements more or less useless in the normal game after the very earliest parts of the levels, which is really unfortunate.

The way to fix this without just running the game at high speed every time you come back would be to calculate the time until the next auto advancement would go off- the time you need to have the money for the next auto upgrade, the time you need for the next auto conquer to go off, the time you need for auto crafter to trigger. Run the calculation the game currently does to advance until that point in time, calculate the new time until an auto advancement, and repeat until the time calculated would be greater than your time away.

Auto-Anything does not operate while you are not in Idle. There will be an explosion of activity when you log in with what you have on hand. This is very easy to test if you have a Multi-Level Power up.

Go into one level, and have your Auto-Stuff active for a while. Then, disable them and switch to the other level. You will have money, armies, ores, etc. that would have been spent if Auto-Something was active.

I tested this myself out of curiosity when I was using a Multi-Level in the past.

They are not active when you are gone, because that would consume too many of your devices resources.

Instead, WZ simulates what should have happened in the past as soon as you return to WZI.

Instead, WZ simulates what should have happened in the past as soon as you return to WZI.

Yeah, but the autos are not part of such simulation. Sadly.

The autos do run during your idle period. However, they may not be as effective.

When you have the game open, it updates all your money/armies/resources ten times per second. However if you close the app and return, say, 5 hours later, it would be too time consuming to do ten updates per second for the entire 5 hours you were gone since that would be nearly 200,000 updates.

Instead, it divides the time you were gone into 50 chunks and does 50 updates. In this case, that would be 6 minute updates (5 hours divided by 50).

So it gives you all your money/armies/resources for 6 minutes of generation, then runs all autos (checks for upgrading army camps, conquering territories, etc). Then repeats this 49 more times, each time advancing 6 minutes.

The net effect of this is that your autos DO get a chance to run during the idle period, but they may lag behind a bit. Instead of activating the exact first second they were able to, they'd activate up to 6 minutes late in this example.

Edited 5/7/2021 19:49:55

Thanks for explaining this, Fizzer.

Thank you for the explanation Fizzer! This sounds like a good way to have automatic effects take place during idle time without excessive computation (and 50 subdivisions of idle time sounds like a good compromise here).

I'm still a little confused about how this works in practice, though. For the test I ran, the first army camp upgrade cost 244, and I had auto-upgrade at 5%, so it should have triggered when I had 4880 money. I started the test with 670 money and made 9.75/sec, so it should have taken me 432 seconds to reach the threshold. Since I have 500 minutes of idle time, I should have reached the threshold during the first subdivision (500/50 = 10 minutes each), which means that the upgrade would trigger at the 10 minute mark. If I understand correctly, I would make 44.2 armies/sec for the first 10 minutes and then at least 85.8 armies/second for the next 490 minutes. (Base army production is 17 for a L2 army camp and 33 for L3; I have +160% army camp boost from AP + artifact.) When the next upgrade triggered, I would make even more, but this should be a lower bound.

If this is correct, then the number of armies earned should be at least
(44.2 armies/sec) * (10 min) * (60 sec/min) + (85.8 armies/sec) * (490 min) * (60 sec/min) = 2,549,040 total armies
but instead I earned only 1.33M armies when I resumed, the same number that I would have earned without auto-upgrade (1.33M = 44.2 armies/sec * 500 min * 60 sec/min). Am I misunderstanding something here? Thank you again for your help figuring this out!

I'm not sure what happened in your test but if there's a bug I'd like to find it.

I just tested this. I did my experiment at the start of the Floating Rocks level.

The first army camp initially gives 11.9 armies per sec. It costs 2.17k to upgrade to level 2, at which point it will give 28.9 armies per sec.

We start at 0 money and 0 armies, and 1 money/sec and 11.9 armies/sec.

Experiment 1: Start the level and advance 6 hours with all autos disabled.

What I expect: 6 hours at 11.9 armies/sec should end up at approximately 21600 money and 257040 armies.

What I saw when I tested it: my money ended up at 21.6k and armies at 256.3k. Within margin of error.


Experiment 2: Start the level and advance 6 hours with "Auto upgrade army camps" enabled. I have it maxed at 100%, meaning it will purchase the army camp as soon as it can afford it.

What I expect: We start with 1 money per second, which means it takes 36 minutes to afford the 2.17k upgrade. So we should get 36 minutes at 11.9 armies/sec and the remaining 324 minutes at 28.9. 36*60*11.9 + 324*60*28.9 = 587520

What I saw when I tested it: My money ended up at 19.4k and armies at 587.7k. Money was still 1 per second and armies was 28.9/sec. Also within margin of error.


So it seems like it's working correctly to me. I realize I can test this easier than you since I can click one button to advance 6 hours, so if you see a problem with my testing methodology, or if you can think of a way my test may have been different from yours, or a way I can test it better, please let me know.

For the test I ran, the first army camp upgrade cost 244, and I had auto-upgrade at 5%, so it should have triggered when I had 4880 money.

Did you have auto-conquer on when you did the test? If Fizzer's test is giving him roughly expected results, one possibility I can see for your results to seem incorrect is you forgot to turn off auto-conquer before you started your test. The number you got coming back seemed to resemble what you would have with AQ at 50% on during the test.

He said in his first post that he disabled auto-conquer

Thanks for testing this on your end. Your tests definitely look like they're working as expected.

In case it's useful, I do have before and after screenshots of the level:
https://drive.google.com/file/d/1eEMWeYyIF1mVfvN2HEuU5TT5gYKh566I/view?usp=sharing (before)
https://drive.google.com/file/d/1eI7rQ7aG0FE1a3TDZ4dKdKxNdEaFoS5L/view?usp=sharing (after)
They do confirm that auto-conquer was disabled during the test, so that isn't the problem.

I wouldn't think that the behavior would be platform-dependent, but I ran the test on Firefox on Linux (Ubuntu 18) in case that's relevant. Is it possible that the "advance six hours" button does something differently from just waiting six hours (you would know better than I would how that button works)? Alternatively, do you see something going wrong on my test?

Just a shot in the dark, but perhaps the problem is happening with the level of Auto-Upgrade you have? Perhaps the game is doing the math with the 5% wrong? I mean, this was a difference between you two trying effectively the same. Perhaps while idle, the game doesn't use 5% but say .005% or ,25% (the squared value) or something similar ridiculous (and therefore waits way to long), whereas while actively playing, the game uses the correct 5% (and therefore produce the impression that it happened right before login, because only while actively playing the threshold is exceeded). With 100% Auto-update the error might not happen for whatever reason (because 100%=1 squared is again 100%=1?!?).

Has anyone seen the Auto-upgrade activating with exactly 5% of your money savings while actively playing?

Experiment 3: I reset my AP and put "Auto upgrade army camps" at 20%, no other advancements. Army camp level 1 now gives 7/sec, level 2 gives 17/sec. Still costs 2.17k to upgrade. I will start the level and advance 8.5 hours.

What I expect: The army camp won't be upgraded until we reach 10850 money, or 3 hours in. So 3 hours will be at 7/sec and 5.5 hours at 17/sec. 3*60*60*7 + 5.5*60*60*17 = 412200

What I saw when I tested it: 28.5k money and 410.1k armies

Okay, strange.
Z's answer already suggests (and I remember such conversations somewhere, too), this wasn't the first post in this forum saying that Auto-advancements seem to not activate during idle. While, afaik no-one yet tested it as systematically as Anthony, this was the current consensus. Meaning that Anthony's observations aren't new and let me believe that this actually is how idle functions currently.
Then, the only conclusion would be that, as Anthony suggested, idling and your Advance-button work differently. Right now, I can't think of another variable we haven't ruled out yet.

I looked back in the forums, and the previous discussion of autoconquer was in this thread: https://www.warzone.com/Forum/520480-autoconquerupgrade-executed. Maybe there's another thread as well that I didn't see. Autoconquer is a bit more difficult to test (in the linked thread, ThijsD conquered a nickel mine and resumed with 0 ore, but this doesn't prove anything definitively because they *might* have conquered it in the last timestep that Fizzer described above). In that thread there was disagreement about what the intended effect was, so thanks again to Fizzer for describing what should be happening here.

Auto-smelt/craft is probably a more definitive test, and Sefer's above post indicates that isn't operating during idle time either.

I tried testing auto-smelt in a similar fashion. I went idle with 2 copper wire and ~1000 copper bars (+ smelter making more so this def wasn't limiting) on a level where barbed wire costs 1 copper wire + 5 copper bar, and copper wire costs 3 copper bar. I had my crafter making barbed wire, and went idle for ~4 hrs 20 minutes. When I got back, a copper wire just started at 0% and I had no copper wire (but a lot of copper bar) in my items. It seems I'm also experiencing the bug that anthony is, where the auto-smelter triggers right when I start playing again, rather than durng my idle time.

You guys are right -- there is a bug that makes autos not fire during idle time. I have figured out why this was broken and will have it fixed in the next update. Thank you for reporting it and for helping debug it, and sorry for the problem.

So it seems like it's working correctly to me. I realize I can test this easier than you since I can click one button to advance 6 hours, so if you see a problem with my testing methodology, or if you can think of a way my test may have been different from yours, or a way I can test it better, please let me know.

Question about your advance button. When you click it to advance idle further, does it count the time skipped as time away from the game, when it's closed, or does it count the time skipped as you using time warp powers? From what it sounds like, your advance button is supposed to act as time away from the game, but since you said you were seeing intended results when testing using your advance button, and you just confirmed that there was a bug with idle not running auto advancements when the game is closed, this makes me believe that the game is seeing you using your advance button as you using a bunch of time warps.

Edited 5/9/2021 04:39:46

On that note, has anyone analyzed the behavior of Autos in regards to Time Warps yet? I guess that if the issue exists there, too then the fix will also cover it. But you never know...