View Issue Details
| ID | Project | Category | Date Submitted | Last Update | |
|---|---|---|---|---|---|
| 0011620 | AI War 1 / Classic | Bug - Gameplay | May 21, 2013 6:06 pm | Jun 8, 2013 10:57 am | |
| Reporter | Burnstreet | Assigned To | keith.lamothe | ||
| Status | resolved | Resolution | fixed | ||
| Product Version | 6.031 | ||||
| Fixed in Version | 6.042 | ||||
| Summary | 0011620: Ships not always aquiring targets although there some in range | ||||
| Description | In my current game I using space planes to great effect, but it seems some of my blob of 450 ships don't like to fight. While kiting around the enemies, some planes stay cloaked, although they got enough possible targets inside their range rings and are not paused. I tried playing around with performance profiles but it didn't help. | ||||
| Tags | No tags attached. | ||||
| Internal Weight | New | ||||
|
|
|
|
|
|
|
|
I loaded the save but by the time I paused and switched to your planes they had all fired. I pulled them back away until they cloaked and the pursuit broke off, then powered them down, brought them over to the AI ships and powered them up again. A few planes in the middle of the group did take about a second longer to fire but I didn't think that's what you meant. I tried it again with having powered them down earlier and noticed that the ones who took longer to fire were actually still reloading from earlier (reloading doesn't progress in stand-down mode). Is it possible this was an active factor in your experience? Anyway, at this point I can't reproduce the incorrect behavior, sorry. |
|
|
ok, seems my report was not detailed enough :/ I did not power down the planes, I kited them around the enemy ships (giving a move order about once per second), so that the enemies are always on the edge of the range of the planes, like you can see on the screenshot. Some of them did cloak although they had enemies in range. |
|
|
Ok, and would those cloaked ships eventually fire (and after how long?) as you continued to kite with stuff in range, or just stay cloaked? |
|
|
Eventually, yes, I think, but I can't say, after how long. A somewhat better to reproduce scenario: 1. load save, pause immediately 2. select all planes on palace fly 3. right click combat regent wormhole 4. x+shift+rightclick topmost harvester 5. unpause what I would expect now is that they move to the wormhole and then back up and as soon as the enemies get into range, they all stop and fight. what I see is that most of them stop right next to the enemies, and about 50-100 don't seem to fire at all, eventually cloak and move on. even when adding another attack move order back to the wormhole, some did not attack. it seems that about once per second some of the planes without targets do get some and stop to fight. |
|
|
|
|
|
I added another save(6.0.41) for reproduction. On tacit blue, if you wait for the enemy active ships to die and attack move 10-25 ships to a point behind the force field, sometimes some will fly around it and not attack. In the forum thread you wrote it may hang on the number of units getting targets assigned per tick, so I tried to reduce the number of units as far as possible. |
|
|
Thanks, that should help isolate things :) |
|
|
It does appear to be the target list throttling thing; here's the test I did: 1) Scrap the planes not near the wormhole on Tacit Blue. 2) Grab 19 planes from the wormhole group, power them down, and move them a fair bit out east of the AI command station. 3) Save. 4) Power those 19 planes up and give them an X+right click order to a point significantly west (but on roughly the same y-coordinate) of the AI command station. 5) Observe that several planes drop cloak and fire as they come into range, while several others do not (with 1 or so dopping cloak to fire every second during the pass) 6) Reload save from step 3 7) Move the still-on-the-wormhole group through the wormhole to Crew Bravo. 8) Power those 19 planes up and give them an X+right click order to a point significantly west (but on roughly the same y-coordinate) of the AI command station. 9) Observe that all 19 planes drop cloak and attack at effectively the same instant. Those planes by the wormhole don't intuitively know that all AI targets are out of range, they have to actually check, etc. I will investigate why the throttle is being so skittish, but it does appear to be that which is at work. |
|
|
Thanks very much for keeping at it, I think I finally got it :) Was throttle-related, but not exactly in the way I thought. For 6.042: * Improved the performance throttling logic for "should I let this ship request a new target list now?" to let ships with absolutely no target list check more than once every 2 seconds for a list if they are: ** An aggregate-targeting leader with followers. ** Or an aggregate targeting follower with a leader who does have a target list. ** This will particularly help ships in Attack-Move actually start firing on targets (and stop moving) as they come into range rather than waiting a while and in some cases simply sliding right past the target. ** This may also reduce performance in some cases, though generally in the bigger battles (where performance is most problematic) the ships in question have targets fairly soon anyway. ** Thanks to Burnstreet for the report and saves that demonstrated the need and provided a testbed for this change. Anyway, following the test steps I described in the previous comment I was able to do steps 4 and 5 with full caps of mkI/II/III planes and while they don't all stop and start firing in exactly the same instant they did do so before any planes actually entered firing range of the AI ships. And no planes just kept going. My guess is that if you got a ton more planes doing this you'd eventually have some slip by, and if the main performance throttles were slagged by a 2000vs2000 fight either on that planet or somewhere else then acquisition could get very sluggish again... but this is at least an improvement :) |
|
|
much better now |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| May 21, 2013 6:06 pm | Burnstreet | New Issue | |
| May 21, 2013 6:06 pm | Burnstreet | File Added: Screenshot_2013_05_21_23_38_07.png | |
| May 21, 2013 6:07 pm | Burnstreet | File Added: sp3 not firing.sav | |
| May 22, 2013 12:19 pm | keith.lamothe | Note Added: 0032566 | |
| May 22, 2013 12:19 pm | keith.lamothe | Assigned To | => keith.lamothe |
| May 22, 2013 12:19 pm | keith.lamothe | Status | new => feedback |
| May 24, 2013 4:17 pm | Burnstreet | Note Added: 0032610 | |
| May 24, 2013 4:17 pm | Burnstreet | Status | feedback => assigned |
| May 25, 2013 5:54 pm | keith.lamothe | Note Added: 0032625 | |
| May 25, 2013 6:18 pm | Burnstreet | Note Added: 0032626 | |
| Jun 4, 2013 2:04 pm | Burnstreet | File Added: target aquisition test.sav | |
| Jun 4, 2013 2:17 pm | Burnstreet | Note Added: 0032787 | |
| Jun 4, 2013 2:17 pm | Burnstreet | Note Edited: 0032787 | |
| Jun 4, 2013 2:45 pm | keith.lamothe | Note Added: 0032788 | |
| Jun 4, 2013 3:07 pm | keith.lamothe | Note Added: 0032789 | |
| Jun 4, 2013 4:53 pm | keith.lamothe | Internal Weight | => New |
| Jun 4, 2013 4:53 pm | keith.lamothe | Note Added: 0032790 | |
| Jun 4, 2013 4:53 pm | keith.lamothe | Status | assigned => resolved |
| Jun 4, 2013 4:53 pm | keith.lamothe | Fixed in Version | => 6.042 |
| Jun 4, 2013 4:53 pm | keith.lamothe | Resolution | open => fixed |
| Jun 8, 2013 10:57 am | Burnstreet | Note Added: 0032820 |
