Matrix Games Forums

Armageddon invasion starts in 2014Command Rommels Panzers in Battle Academy!Servers UpdateThe Deal goes Fourth!Command Ops gets a Massive Update!Lost Battles AAR: Smolensk 41War in the East: Lost Battles is here!Conflict of Heroes: Awakening the Bear! Let's Play videoLost Battles Scenario Manual excerpts available!A sneak peek from Allied Corps!		 Forums  Register  Login  Photo Gallery  Member List  Search  Calendars  FAQ 

My Profile  Inbox  Address Book  My Subscription  My Forums  Log Out

RE: locked up?

 
View related threads: (in this forum | in all forums)

Logged in as: Guest
Users viewing this topic: none
  Printable Version
All Forums >> [New Releases from Matrix Games] >> Command Ops Series >> Tech Support >> RE: locked up? Page: <<   < prev  1 [2]
Login
Message << Older Topic   Newer Topic >>
RE: locked up? - 2/25/2012 9:25:00 AM   
phoenix

 

Posts: 951
Joined: 9/28/2010
Status: offline 		Thanks Dave. Much appreciated.

(in reply to Arjuna)
Post #: 31
RE: locked up? - 3/2/2012 6:18:58 AM   
Arjuna


Posts: 16450
Joined: 3/31/2003
From: Canberra, Australia
Status: offline 		phoenix,

Finally got a moment to look into this issue. Interesting. I can confirm that it is not a formation lockup. From here on it gets complicated though.

You ordered the 2Bn to use the quickest route to the bridge and you authorised it to bypass if necessary. The problem is that the vanguard, B Coy, gets to a point or starts at a point where the route will cause him to head towards a visible enemy threat in the woods NNW of its position. wAt this point the respective unit radius are very close to each other. They are within the 250m standOff range. This is new code I introduced into BFTB to prevent units from crashing into each other. So it halts. Unfortunately I had failed to include a line of code that I use in similar reaction code to force a reassessment by the unit. In other words, if you are stopped then try and do something else.

So I added this line of code. Sure enough B Coy decides to reassess. But he fails to bypass on his own initiative and calls some code to bring his boss's reassessment forward. All good. However, two subsequent issues arose. First one was that you had issued a Quickest route task parameter and this was overriding the bypass option. I don't believe that it should do so. So I have changed that now for human players.

The second issue is more profound and I am going to have to spend more time assessing how best to address it. Basically it's due to the way we manage the end time of human player tasks. This code goes back to the dark ages as far as the code base for this engine is concerned and it has a history of sorts. Originally we set the end time of a task to be whatever the start was plus the likely duration. However, we found that if this was unduly short then we ended up with a lot of processing to continually slip the end. So we decided to instead set the end of player tasks to be that of the scenario. After all there can never be more than one player task per unit. This situation will have to change when we introduce sequential tasking for the player anyway. I probably need to bite the bullet now and do so because it is having a profound effect on the bypass code for player tasks.

The reason this happens is that we use a test to see if the force is making progress and if it's not then we test to see if we should bypass. Alas the code to determine if we are making progress relies on a comparison between the "remainingDuration" and the "estimatedCompletedDuration". If the remainingDuration is greater then we're ahead of schedule. If less then we're behind. With player tasks using the scenario end they invariable end up ahead of schedule and hence don't bother to tyest for the bypass.

This will have to change. But I need to think about this some more. I hope to do so and work up a fix next week.

_____________________________

Dave "Arjuna" O'Connor
www.panthergames.com

(in reply to Arjuna)
Post #: 32
RE: locked up? - 3/2/2012 6:48:50 AM   
phoenix

 

Posts: 951
Joined: 9/28/2010
Status: offline 		Thanks so much, Dave. At least I have a diagnosos. Now I know I'm not mad, it's not all in my head. I don't understand much of what you said, of course. Is it that the lead units halt the formation on contact then get a bit stuck because they think they have until the end of the game to resolve the issue? Either way, I see it all over now - just long lines of 'halted' in the msg box - it really is driving me mad. It certainly looks like they think they have all the time in the world to get there!! . Until you do the fix can I avoid it by not using bypass? Maybe use avoidance for the planning instead and make do?

(in reply to Arjuna)
Post #: 33
RE: locked up? - 3/2/2012 10:40:06 AM   
RangerX3X


Posts: 182
Joined: 11/21/2007
From: Jacksonville, FL USA
Status: offline 		Just one thing to note regarding 2 Para Bn HQ as implemented in this mission. It starts out at 75% of the estab with a staff quality rating of only 70%. The commander setting for John Frost has leadership only at a “good” rating of 64% and judgment only at a “good” rating of 72%.

The other sub units of the battalion fare worse in the ratings department for staff quality with the companies coming in at 58%, 59% and 65% respectively, while the mortar platoon is at 56% and the anti-tank battery is at 57%.

The point I am trying to make here is that in terms of staff quality, the battalion is not of “crack” or “elite” status as represented in this mission. Furthermore the entire battalion is only at 68-83% of their individual establishments for personnel.

I would tend to agree that adding additional units to the battalion or having them trudge through enemy territory with anything other than an attack command while lugging around AT and mortar units could bog down the AI subordinate commander.

And quite possibly there could be one unit out of all of them much more fatigued than the others which could slow the whole thing to a crawl while at the battalion level fatigue would appear not to be a concern.

Now that I read Dave's post above my response seems like rubbish

< Message edited by RangerX3X -- 3/2/2012 10:44:19 AM >

_____________________________



Marine Corps Mod - Download Latest Files Here

(in reply to Arjuna)
Post #: 34
RE: locked up? - 3/2/2012 12:51:01 PM   
phoenix

 

Posts: 951
Joined: 9/28/2010
Status: offline 		Well, not rubbish, as I did try Richard's suggestion and ran it all without the baggage - they went faster, but they still had this 'halted' 'halted' 'halted' problem.

(in reply to RangerX3X)
Post #: 35
RE: locked up? - 3/6/2012 8:27:47 AM   
phoenix

 

Posts: 951
Joined: 9/28/2010
Status: offline 		More on this issue, Dave. I really hope you can work up a fix, because it's driving me mad - so much so I'm reluctant to play any bigger scenarios now. I tried to avoid giving orders with the 'attack', 'bypass' 'ambush' boxes checked to see if that made any difference, but it made none. So, even if I just give an order to move, or secure a crossing, and pick only 'quickest' and fastest or fastest and avoidance etc, with no rest, then what I still see is huge lines of stalled troops when I try to play From the Meuse to the Rhine. When I check their logs I see long lines of 'halted'. I went to Joe's Bridge to see if it happened there and it did, though not as much. All the formations do eventually get to where they should be going, but there are hours and hours when they are just halted (sometimes as much as 12 hours, but more often between 3 and 5 hours), with no real explanation save 'halted'. I'm surprised nobody else is mentioning this. It's a new feature for me - that is, I'm only noticing it now. I might have missed it before, but it's hard to miss in the bigger scenarios, especially if you play at slow speed, because you end up sitting there with perfectly good orders but nothing happening for hours at a time. I can get you more saves if you want - just let me know - but I hope your 'fix' will address it. Thanks.

(in reply to phoenix)
Post #: 36
RE: locked up? - 3/6/2012 9:29:42 AM   
Lieste

 

Posts: 1608
Joined: 11/1/2008
Status: offline 		Sounds just like real world combat operations... hours of sitting about 'for no reason', a huge rush to get to the next location and more interminable waiting...

I think the phenomenon even has a name - 'hurry up and wait'. While it does sound like there is a fault which will be addressed, it is also possible that there are unreasonable expectations coming to the fore as well.

(in reply to phoenix)
Post #: 37
RE: locked up? - 3/6/2012 9:49:29 AM   
phoenix

 

Posts: 951
Joined: 9/28/2010
Status: offline 		Could be, Lieste. But in which case - if they are actually doing something - then a message in the log would be best, I think - to make that clear. Since I started playing this series - way back with HTTR - formations have always behaved like this to a certain extent, I think - and I have always just put it down to real world modelling or whatever, but the delays are so long now, sometimes, that I begin to think that's no longer feasible. What could they be doing? Well, the most likely candidate, usually (and as Dave said had happened with 2 para in my original example) is that the lead elements run into something. Then what you see is activity with the lead element and everyone else stopped, strung out along the road behind, say. We must assume (and, again, Dave suggested this did happen) that the lead element thinks about it, solves the issue or sends back to HQ for instructions. HQ ponders, and everyone waits. I'm alright with all that. But then when nothing happens for another four hours I begin to wonder. Sometimes then the original engagement is finished and nothing is happening anywhere. My realworldometer - which is intuitive and not based on anything real at all, I admit - says that if a battallion is strung out along a road with, say, orders to secure a crossing five kilometres away by the fastest means, with a planing option to avoid contact, have no rest on the way, and leave stragglers behind, then if the lead company makes hostile contact and sends back for instructions a distance of only a kilometer to its HQ, I would hope to see something happening indicative of a plan by that HQ within a couple of hours. If nothing is happening five hours later then I could, of course, assume that the wires got cut by arty, all the runners were killed etc etc, repeatedly, but if this kind of thing happens often enough then I begin to think the norm is that five hour delay, and then I begin to wonder if it's a bug.

(in reply to Lieste)
Post #: 38
RE: locked up? - 3/6/2012 1:34:05 PM   
Lieste

 

Posts: 1608
Joined: 11/1/2008
Status: offline 		One thing you can do/try is to detach the lead element(s) if they are hung up, when the trail of the column will continue attempting to engage/bypass etc. using the contacting element as a screen or fixing force.

The risk is a requirement to regroup and re-org the formation, or the potential for loss of the separated element.
The advantage is that the now 'defending' element is less vulnerable to fires, and will develop more firepower itself.

(in reply to phoenix)
Post #: 39
RE: locked up? - 3/6/2012 1:50:30 PM   
phoenix

 

Posts: 951
Joined: 9/28/2010
Status: offline 		Will try that. Thanks.

Started playing FTMTTR through again this morning and so far I've taken all the major objectives, except the Arnhem Road Bridge (and, anyway, I always thought that was a bridge too far...) and haven't had as many of the standstill problems. So who knows. Maybe it isn't a bug, though I'm doing nothing different to all the other times, except making sure I uncheck the stragglers box for absolutely everything (in case that's what's slowing it all).

Much of the behaviour would be understandable, in a way. EG; I tell tell 1.505 PIR to march 8 kilometres and secure the Honinghutje bridge full speed. They march, make contact, start engaging, the lead element crosses the bridge, the others pause on the near bank. I assume they're re-orging to cross and secure the far bank with the lead element (I placed the objective on the far bank), but no, they just stand still for two hours. They've just marched 8 kilometres and had a little action, so I could understand them wanting to rest and regroup. But all the indicators tell me they're moving, when they're not. Except the log, which tells me they're halted. For two hours. Why? What are they doing? If I was told they were resting and regrouping I might accept it, but I'm not told that. I'm told they're getting on with the task, moving, in one indicator, but in the log that they're halted. It's irritating. Maybe they're taking a long late lunch, a petite dejeuner sur l'herb. The log should say that then - 'sunbathing', perhaps, or 'chatting, come back later'.....

(in reply to Lieste)
Post #: 40
RE: locked up? - 4/15/2012 11:43:40 PM   
Deathtreader


Posts: 651
Joined: 4/22/2003
From: Vancouver, Canada.
Status: offline
quote:

ORIGINAL: Arjuna

phoenix,

Finally got a moment to look into this issue. Interesting. I can confirm that it is not a formation lockup. From here on it gets complicated though.

You ordered the 2Bn to use the quickest route to the bridge and you authorised it to bypass if necessary. The problem is that the vanguard, B Coy, gets to a point or starts at a point where the route will cause him to head towards a visible enemy threat in the woods NNW of its position. wAt this point the respective unit radius are very close to each other. They are within the 250m standOff range. This is new code I introduced into BFTB to prevent units from crashing into each other. So it halts. Unfortunately I had failed to include a line of code that I use in similar reaction code to force a reassessment by the unit. In other words, if you are stopped then try and do something else.

So I added this line of code. Sure enough B Coy decides to reassess. But he fails to bypass on his own initiative and calls some code to bring his boss's reassessment forward. All good. However, two subsequent issues arose. First one was that you had issued a Quickest route task parameter and this was overriding the bypass option. I don't believe that it should do so. So I have changed that now for human players.

The second issue is more profound and I am going to have to spend more time assessing how best to address it. Basically it's due to the way we manage the end time of human player tasks. This code goes back to the dark ages as far as the code base for this engine is concerned and it has a history of sorts. Originally we set the end time of a task to be whatever the start was plus the likely duration. However, we found that if this was unduly short then we ended up with a lot of processing to continually slip the end. So we decided to instead set the end of player tasks to be that of the scenario. After all there can never be more than one player task per unit. This situation will have to change when we introduce sequential tasking for the player anyway. I probably need to bite the bullet now and do so because it is having a profound effect on the bypass code for player tasks.

The reason this happens is that we use a test to see if the force is making progress and if it's not then we test to see if we should bypass. Alas the code to determine if we are making progress relies on a comparison between the "remainingDuration" and the "estimatedCompletedDuration". If the remainingDuration is greater then we're ahead of schedule. If less then we're behind. With player tasks using the scenario end they invariable end up ahead of schedule and hence don't bother to tyest for the bypass.

This will have to change. But I need to think about this some more. I hope to do so and work up a fix next week.


Just curious as to how this might be panning out...............
Thanks!

Rob.

_____________________________

So we're at war with the Russkies eh?? I suppose we really ought to invade or something. (Lonnnng pause while studying the map)
Hmmmm... big place ain't it??
- Sir Harry Flashman (1854)

(in reply to Arjuna)
Post #: 41
RE: locked up? - 4/16/2012 8:08:18 AM   
phoenix

 

Posts: 951
Joined: 9/28/2010
Status: offline 		+1

(in reply to Deathtreader)
Post #: 42
RE: locked up? - 4/23/2012 9:35:15 AM   
Arjuna


Posts: 16450
Joined: 3/31/2003
From: Canberra, Australia
Status: offline 		Hi guys. We've had a bit of problem with getting our new Source control software working properly. After a lot of hassles that looks sorted. So today I started back on this issue of timings.

As I mentioned above there is an underlying issue relating to the determination of a force being ahead or behind schedule. this was affecting the Vypass reasessment code. I have made changes to now calc a route and determine the duration of the task based as soon as you the Player issue a Move order on the map. I then set the end of the player task based on the estimated StartAt time ( factoring in orders delay down the line ) plus the estimated duration. I have just stepped through the code and now the bypass works as it was originally intended. I have also removed the test for route type that was preventing bypass if the route type was set to anything other than unspecified. It now only depends on the Bypass option setting.

However, I discovered that the settingf for these options ( allowBypass and allowAttacks ) defaults to false for all tasks. Now that is fine for the human player because he can override that with the optioon checkbox. But this is not good for the AI. Effectively it denies him the option. I had put in place special provisions within the code to override this for the AI controlled forces. but I think it best to simplify the matter and rely solely on the task option setting. This will require me to convert the scenarios. I'll start on that tomorrow.

_____________________________

Dave "Arjuna" O'Connor
www.panthergames.com

(in reply to phoenix)
Post #: 43
RE: locked up? - 4/23/2012 10:48:39 AM   
phoenix

 

Posts: 951
Joined: 9/28/2010
Status: offline 		Thanks Dave. None of that means anything to me. :) But if it means that in the end they'll stop all that senseless halting (or the sense of it will be explained) then I'll be happy. I had an interesting experience playing through the huge 15 day Spearhead v Reich scenario yesterday - I wonder if it's connected to this issue. This is what happened. I pulled back everything I had to a defensive line running roughly from the Erezee exit, through Grandmenil, Manhay, Vaux-Chavanne and then up the road to Chene-al-Pierre. I concentrated all reinforecments up to Day 7 ish on holoding this line and gathered all the arty into a big block I could control. The AI attacked in 2 places. Up the big red road from Baraque Fraiture to Manhay, and from Oster to Grandmenil. By day 5 there were massive concentrations of axis forces in Oster (looking to attack Grandmenil) and stuck on the road leading up to Manhay. I bombarded these concentrations mercilessly. What happened? The AI seemed to get stuck. By day 7 the attacks all stopped. I was sittting there in my defensive line waiting for reinforcements so I could come round the roads on the eastern flank and re-take Baraque Fraiture. but I wasn't getting attacked at all!!! I surrendered and had a look. The AI was very far from finished. There wwere still massive concentrations of Axis forces on the map, they had retreated and gathered near Lamormenil and Malampre. And when I continued play this didn't change much. 8 days of the scenario left and the AI just seemed to have siezed up. There were a few attacks, but mostly it just seemed shattered, stuck in disorganised ocncentrations, unable to come up with further attacking plans, but with ample material to do so. I started attacking and the shot below is from Day 11, showing theat the AI has now retreated into the area I've ringed. There have been no attacks since day 7, I think. So I had the feeling the AI might be 'halting, halting, halting' if only I could see his stats properly?




Attachment (1)

(in reply to Arjuna)
Post #: 44
RE: locked up? - 5/7/2012 4:34:51 PM   
phoenix

 

Posts: 951
Joined: 9/28/2010
Status: offline 		Hey Dave. Any progress you can mention on this piece of work?

(in reply to phoenix)
Post #: 45
RE: locked up? - 5/7/2012 11:57:14 PM   
Arjuna


Posts: 16450
Joined: 3/31/2003
From: Canberra, Australia
Status: offline 		I've done the change at my end. We need to test this though and that requires a new debug build and testing round. However, we opted first to add new code that would convert user made data content to the latest version. That way we will better support and encourage user content. What started out as a simple undertaking has turned a lot more complex. What we uncovered were some fundamental issues with our database engine and Microsoft's Visual Studio ( which is the development software we use ). We are working our way through these. As soon as I can we will get a new patch out.

_____________________________

Dave "Arjuna" O'Connor
www.panthergames.com

(in reply to phoenix)
Post #: 46
RE: locked up? - 5/8/2012 6:34:38 AM   
phoenix

 

Posts: 951
Joined: 9/28/2010
Status: offline 		Many thanks Dave. Appreciate the update. Look forward to that patch, then.

(in reply to Arjuna)
Post #: 47
Page:   <<   < prev  1 [2]
All Forums >> [New Releases from Matrix Games] >> Command Ops Series >> Tech Support >> RE: locked up? Page: <<   < prev  1 [2]
Jump to:





New Messages No New Messages
Hot Topic w/ New Messages Hot Topic w/o New Messages
Locked w/ New Messages Locked w/o New Messages
 Post New Thread
 Reply to Message
 Post New Poll
 Submit Vote
 Delete My Own Post
 Delete My Own Thread
 Rate Posts


Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI

0.268