Halting bug

[dazkaz15]

The Halting bug is back I'm afraid.
Or at least a strain of it [;)]
I'll send the save to Panther@support or would you rather have it as Drop box to Miquel?

Also you might want to take a look at Intel reports.

Everything seems to be infantry or Engineers now, until they become "excellent" in status, and estimates seem very wild compared to what it used to be.
The most worrying thing about this is that artillery, even SP artillery is not identified until extremely close range, so the AI will stand no chance in being able to conduct any counter battery fire as the artillery is not identified as such.

Also what should we expect from infantry V unsupported armour at night?
At the moment the armour only positions don't seem to be getting over run/infiltrated, by infantry.
Just lots of halting instead.

[vandorenp]

ORIGINAL: dazkaz15

The Halting bug is back I'm afraid.
Or at least a strain of it [;)]

I have noticed something similar. But the problem has been repetitive through all releases. The FBB HQ, and the FGB HQ in the past, always locks up when Exit or move orders are given to it. Not until you give all subordinates seperate orders does it move on. But this time it was many hours later.

I plan to set up a scenario with several FBB's with the HQ as a Panzer RGT HQ, as a Panzer Div HQ and the brigade HQ and see if there is a difference.

[vandorenp]

ORIGINAL: dazkaz15

you might want to take a look at Intel reports.

Everything seems to be infantry or Engineers now, until they become "excellent" in status, and estimates seem very wild compared to what it used to be.
The most worrying thing about this is that artillery, even SP artillery is not identified until extremely close range, so the AI will stand no chance in being able to conduct any counter battery fire as the artillery is not identified as such.

Yes I noticed this. I was going to go back to scenmaker and reset the intel to see if it made a difference. We do remember that artillery is reliably spotted by impact analysis and sound detection, right?

[dazkaz15]

I believe it was flash spotting that was the most reliable on a fast moving front line, sound ranging taking quite a while to set up was used more for static, defensive positions, and impact analysis although useful would not give a very accurate fix on an enemy battery location, but would at least tell you what direction it was in.

I had a situation in the Hofen Scenario at the end, where I had no intel at all for a while on the 62 Armoured Field Atry Bn, yet it was actually direct firing at my infantry that was deployed in the woods. Even after it was identified it was still only ID as infantry Coy.
See slide 61 of my Hofen AAAR.

[dazkaz15]

Miquel

I have added the file to my dropbox its called "Halting bug D2 0013.cog"

Please let me know if you can't access it, or if I also need to sent the file to Support@panther.

Thanks
Daz

[Arjuna]

daz pls send it to support@panthergames

[BletchleyGeek]

ORIGINAL: dazkaz15

Miquel

I have added the file to my dropbox its called "Halting bug D2 0013.cog"

Please let me know if you can't access it, or if I also need to sent the file to Support@panther.

Hi Daz,

I didn't get any notification via Dropbox. What I'll do is to share a Dropbox folder to you, so you have a clear "channel" to funnel those saved games through

[dazkaz15]

Cool, thanks Miquel [:)]

I have sent a save to Dave as well via email.

This is a surrender image D2 00:13

From what I have read, A Coy 17 Tank Bn, should have no chance in the dark against 3 SS Coy 1 FB Bn in this situation.

The articles I have read indicate that Armour can not hold ground in the dark without infantry support.
When I get back from work Ill try to find the relevant links again and post them here.

The Mech Coy is well within the effective range for its Anti Armour weapons, and I assume the infantry component would now be leading the assault on foot, with the AFV's to the rear.

I may have misunderstood how the Mech Infantry is modelled in game though, so I'm very open to any advice you can give me about how they are best deployed, and used.

Something from the surrender image is confusing though.
The General tab indicates there are 3 AFV's, 1 non AFV, and 1 gun, but the E&S tab only indicates there are 2 Sherman's in the unit?

[BletchleyGeek]

We've been discussing a possible fix for this issue, we need to implement and test it. Thank you Daz for bringing this issue forward, it's very much appreciated.

[Phoenix100]

Which issue, Miguel? Daz raises a few, I think. Including that long list of 'halting'....

[BletchleyGeek]

ORIGINAL: phoenix
Which issue, Miguel? Daz raises a few, I think. Including that long list of 'halting'....

I was referring to the important one - the one about the "halting".

That thing about the intel reports is also interesting, both views should be in sync.

[Arjuna]

The root cause of the original problem about why the German coy halts is because of an anomaly in the intel system. Basically what happens is that each minute when each unit processes its current event it calls a function called ScheduleNextEvent() which handles all the scheduling and reaction for the unit. At the top of this function is a call to SpotENemyForces() in which the unit determines the visProb of each enemy unit. Then throughout the reaction code it will call another function called IsVisible() on each of the enemy units. This IsVisible() includes a random roll and if this is less than the visProb then the enemy is visible.
 
The problem is that it is possible within the same minute for the same enemy to be both visible and not visible due to this use of random. So in this particular case they were first failing to see the enemy and hence did not fire or launch an attack or retreat or call off the attack. They had multiple chances but the visProb is extremely low. However, in the code that determines if they are within grenade range of the enemy they were deemed to be visible and so would halt.
 
The fix is to roll the random just once during the SpotEnemyForces() and store the result inside the EnemyThreats class. Then during thre reaction functions it will be a simple lookup rather than a determination. This same aspect would have affected other things like firing etc. So it needs to be fixed. I am working on this today. Good catch. Thanks.

[dazkaz15]

Thanks for the update Dave.

I've said it before I know, but the support you guys give this game is second to none.

That was a very informative reply, and makes the effort of reporting it worth while.

Thanks [:)]

Daz

[Arjuna]

Your welcome but the above post was really only the tip of the iceberg. It's been a day of pulling different strands of string. Needless to say that I have finally gotr your scenario to run and the 3 PG Coy duly overran and destroyed the US Tk Coy ( all two tanks o it anyway ).

[Phoenix100]

It's superb support. And still the best game out there (though I have been having increasing fun with Steam and Iron!) by a long way. There was a fallow time last year (when Dave was hors de combat with the hand) when I thought the game development was over, to be honest. But it's back with a vengeance now. This beta is looking great - a big improvement.

Are there only 2 extant bugs now, then - the halting and the support companies leading? Interesting that these are regulars, through different builds. Interesting for me (you know like Oscar Wilde's comment about how he loves work - he could sit and watch it all day long...) less so for Dave et al, I assume!

I'm really looking forward to the devs moving on now and getting out CO2 and the EF game, especially. But also the med pack.

Hope these 2 bugs are the end of it for this iteration. I had thought that the major change for this build - the change that might throw up the unintended consequences - had been the supply grid refinements (the 100m thingy) - but these recurrent bugs are nothing to do with that, right, Dave? So why do these two bugs (halting and support units leading) keep coming back? Was it to do with this new build, or have we just by chance discovered them now and they were always in there?

[dazkaz15]

ORIGINAL: phoenix
I had thought that the major change for this build - the change that might throw up the unintended consequences - had been the supply grid refinements (the 100m thingy) - but these recurrent bugs are nothing to do with that, right, Dave? So why do these two bugs (halting and support units leading) keep coming back? Was it to do with this new build, or have we just by chance discovered them now and they were always in there?

Don't want to put words into Dave's mouth but there were a lot of changes in the last patch that could introduce more problems Peter.

tm.asp?m=3527107

The one from the change list I suspect that has caused unintended consequences to the Intel is this:

"Keep enemy units spotted for longer"

The flashing in and out of Threat AKA the halting bug I think might have been with is a while now, and may have been something to do with the single M8 surviving so long in my Hofen AAR, that was before I updated to the latest patch?
(This is just an uninformed and uneducated guess)[:'(]

There will always be bugs, and things that can, and for a better player experience need to be done.
Highlighting them can only go to improve future releases of the engine.

We don't want the same bug in Command ops 2 right, but there will be bugs [;)]

[dazkaz15]

This is quite an interesting one as well.

"We introduced the possibility of target overkill (i.e. unless a vehicle brews up, it's still considered a valid target)."

Does this mean that the unit will halt in place and keep firing at a target that is destroyed, until it actually catches fire?

[BletchleyGeek]

ORIGINAL: dazkaz15

This is quite an interesting one as well.

"We introduced the possibility of target overkill (i.e. unless a vehicle brews up, it's still considered a valid target)."

Does this mean that the unit will halt in place and keep firing at a target that is destroyed, until it actually catches fire?

Not so sure about the halting bit, but yes, AFV's which have been destroyed but have not "brewed up" are still considered valid targets by the targeting routines.

[dazkaz15]

Thanks Miguel

What I really wanted to know though is how this will effect the game play?

[BletchleyGeek]

ORIGINAL: dazkaz15

Thanks Miguel

What I really wanted to know though is how this will effect the game play?

That's difficult to say - is a very tactical detail which might well go swept away by other outcomes at the level of detail we're portraying here. But as a general rule, the casualty rates of units with highly ronsonable vehicles will be higher than that of those with less ronsonable vehicles.