Razorback Ridge V3.11

[DVDBOATRIGHT]

Has Razorback Ridge been changed in 3.11? An MG now rakes the left hand side of the crossroads from across the river and an MG has appeared with the Panzerschreck on the right hand path. Also why can't US MGs not sight and fire through bocage when sitting behind it while German MGS can quite happily fire from there and set back from the bocage.

[AttorneyAtWar]

Has Razorback Ridge been changed in 3.11? An MG now rakes the left hand side of the crossroads from across the river and an MG has appeared with the Panzerschreck on the right hand path. Also why can't US MGs not sight and fire through bocage when sitting behind it while German MGS can quite happily fire from there and set back from the bocage.

1. Certain campaign/Normal scenarios have multiple AI plans so seeing those units in different places is common.

2. German MG's and American MG's don't have discrepancies for LOS, quite simply the spot that MG is in has bad LOS to whatever you are trying to look at. I know it can be frustrating when you move a unit into a position where you think it can see and it can't, unfortunately the best thing you can do is draw LOS from that position by selecting the movement point and using "target" to see what it can see.

I hope that helped and good luck!

[DVDBOATRIGHT]

1. Certain campaign/Normal scenarios have multiple AI plans so seeing those units in different places is common.

2. German MG's and American MG's don't have discrepancies for LOS, quite simply the spot that MG is in has bad LOS to whatever you are trying to look at. I know it can be frustrating when you move a unit into a position where you think it can see and it can't, unfortunately the best thing you can do is draw LOS from that position by selecting the movement point and using "target" to see what it can see.

I hope that helped and good luck!

Procedure goes something like this. Move to bocage, the light on grey simple appears above the gerrman mg. Then I try to trace los., no deal, pink line. Then the same german mg opens up and tears the US mg a new one.

[Duriel Krugaire]

If one MG fires on the other... the MG recieving fire physically must be able to see the other, unless the firing MG is using area fire.

There are a few spotting bugs. I was in a battle where I had to pull my tank destroyer up to within 10 ft of the rear of a panther before it could see it. The AI is very suspect.

[womble]

If one MG fires on the other... the MG recieving fire physically must be able to see the other, unless the firing MG is using area fire.

Not so. Your statement might be more correct if you put "MG Team" in place of "MG". LOF for an MG team (from the position it's currently in) for the purposes of the Targeting Tool is calculated from the precise and exact location of the heavy weapon in question. LOF for other members of the team can easily be different, and "spotting" by that team could mean that only one member of the team has eyes-on, and that might not be the gunner. The most common causes for this sort of apparent disconnect are:

1. the gunner is "Cowering" behind something that blocks his view while prone and foetal. If you have the SOP of setting a Face order (or short, directional Target Arc if you want them to hold their fire) at the "Deploy" waypoint, this is the biggest contributor by far to heavy weapons not being able to see what you expected them to be able to when they get where you sent them.
   
2. the team has set up the gun in a position that can't see past a linear obstacle. I don't think this is an AI problem, because AFAICT, the TacAI isn't involved; the team sets up in the same positions relative to the direction they're Facing, whatever direction they're Facing, and if that's not "through" the linear obstacle, there's a good chance the "standard team formation" doesn't put the MG near enough the hedge to see through it. Face command is your friend here (though it's not 100% reliable, it's pretty good - my SWAG would be it gets your MG pointing right 95% of the time).
   

Another disconnect can be generated by the fact that the LOF calculations from a waypoint where a heavy weapon is going to be are calculated from the centre of the Action Spot, rather than the location in the Action Spot where the heavy weapon ends up. Again, Facing will help make sure the prediction and actuality coincide, but might not be perfect for a number of reasons, including trees.

There are a few spotting bugs. I was in a battle where I had to pull my tank destroyer up to within 10 ft of the rear of a panther before it could see it. The AI is very suspect.

You realise that the AI has little or nothing to do with spotting calculations? I'm sure it has an influence on how frequently spotting is assessed, but outlier spotting problems aren't a sign of "suspect AI". And without some pretty solid evidence to the contrary (save games, for example), most people will be wondering what other factor that you've elided from your memory or never noticed in the first place (because CMx2 is an insanely complex environment, and it's real easy to miss things in the hear of the battle) made that Panther so hard to see.

[Duriel Krugaire]

Well we have CMx2 and we have real life.

What you say about CMx2 may be true, but in real life if you are getting fired on you can fire back. If a gunner sticks his head up to fire, his head is available to receive fire. Tell me when this is not the case... in real life.

As far as your other comments, blame me for CMx2 spotting errors. Ha. It was a clear day, I was controlling a able tank killer tank, the german panther was 40 yds in front be hind some burning armor, I drove to his rear @ 40 yds, turn directly towards him, there were no armor vehicles in front of me now, & it drove 20 yds from the rear and sat there for 20 seconds, the panther rotated his gun about 120 degrees to aim at the rear before the tank killer got his shot off. That was ridiculus and cannot be explained. I could not target the tank at the start of the turn because he was behind flames.

It is still suspicous AI.

[JonS]

cannot be explained.

Sure it can be explained, but I strongly suspect you wouldn't like the answer because it's clear you expect your units to act with robotic precision and reliability, rather than fuzzy best effort simulating human weaknesses. So there's really no point trying.

Good luck!

[Duriel Krugaire]

Sure it can be explained, but I strongly suspect you wouldn't like the answer because it's clear you expect your units to act with robotic precision and reliability, rather than fuzzy best effort simulating human weaknesses.

Well, when you drive up on the rear of a panther @ 10 yds, with your barrel pointing right at it, not under any suppression. Yes, I would expect you to get a round off before the other tank spots you. You are at it's rear and it is buttoned, so I would think it would never see you as it is taking rifle fire from the front and actually firing rounds to the front. But no... it spots you and gets it's turret around 120 degrees before you crew even fires. It seems a little buggy.

I have the game saved before that round....

I don't know how to save the video or ????

[womble]

Well we have CMx2 and we have real life.

What you say about CMx2 may be true, but in real life if you are getting fired on you can fire back. If a gunner sticks his head up to fire, his head is available to receive fire. Tell me when this is not the case... in real life.

Oh, so sorry that CM isn't a perfect photorealistic recreation with every cubic centimetre of battlespace monitored and crossreferenced the entire time.

Read again what I wrote. There are plenty of RL situations that CM points up where an MG team can hose down another MG team and that target MG team's MG cannot see to return fire. You don't like that, find another game that deals more precisely with tactical combat at the scale of CM. Oh look, there isn't one.

As far as your other comments, blame me for CMx2 spotting errors. Ha. It was a clear day, I was controlling a able tank killer tank, the german panther was 40 yds in front be hind some burning armor, I drove to his rear @ 40 yds, turn directly towards him, there were no armor vehicles in front of me now, & it drove 20 yds from the rear and sat there for 20 seconds, the panther rotated his gun about 120 degrees to aim at the rear before the tank killer got his shot off. That was ridiculus and cannot be explained. I could not target the tank at the start of the turn because he was behind flames.

Trees (turned off), smoke (turned off - extra potential here cos you're in an environment with burning tanks), movement screwing with the turret rotation, panic, confusion of battle. Plenty of possible reasons that you've not eliminated why it didn't work out exactly as you expected. That the end result (stalking TD nails distracted Panther) was correct isn't enough for you? Maybe BFC have included some "nail biting" code to make "gotcha" setups like that seem a bit more exciting. Note: sarcasm may have been included in that last sentence.

It is still suspicous AI.

No it's not. At the very most, it's an outlier spotting SNAFU that you're fixating on rather than the millions of other times that you haven't noticed the spotting algorithms work just fine. An outlier spotting SNAFU that changed the eventual outcome not at all (unless there happened to be another unaware Panther in close arc that needed killing that minute that the delay meant the TD didn't get to - since you didn't mention that, I'm going to assume that it wasn't your birthday). You do know that it's a game and the product of fallible humans, right?

Nothing

Whatsoever

To

Do

With

AI.

And it's not "suspicious" either. What are you trying to imply? That the AI cheats? It's not very good at it, then if it does, is it, since your TD geeked the kitty.

You want to help identify whether something is broken? Post a replay phase save game where this happened to a dropbox account and put the link here. I suspect that since you mention you've got a pre-turn save you've probably not got a replay save too. But you might edumicate yourself some by having the game calculate that turn a few times and see whether the situation repeats itself. If it does, and you're sure you've got smoke and trees turned on, then take a save of that replay and post it to dropbox, and share the link here. Need to know the exact game version you're using. Replicatable things like this will attract the attention of those who can bring it to the attention of those who can do something about it.

[Duriel Krugaire]

The only thing you didn't explain:

What is a 'replay phase save game' and how do I do it? Never seen that button!

Sorry to train wreck the OP post, but guess this is a little like the gamey game.

[JyriErik]

You're going to hate yourself when you find out. A "replay phase save game" is a game saved during the replay phase instead of the orders phase.

Jyri

[Duriel Krugaire]

I'm going to hate myself... I don't think so.

I was thinking it was something you do so you can make a you tube or something with it.

I figured out I had the tank set to 'fast', changed the last bit to hunt and it worked better. Although I have not had such good luck with hunt, as it may target another inf team or something else, it will fire on them, then quit the 'hunt', and the main object can fire on it, but it can't see the object, yada yada yada. I don't know how to get it to restart the hunt after a target is found in the 1 min interval.

I had the smoke & tree's on, this happened in a road, without mortar/vehicle smoke so it was clear air. The only smoke was straight up out of the trashed vehicles.

The file is about 2800 kb (about 2700kb more than the allowed forum limit). I guess if there is real interest, I can post on a drop box, but that is more work, since we kinda resolved that issue.

I guess I learned another insight.

[RedBarron]

I figured out I had the tank set to 'fast', changed the last bit to hunt and it worked better. Although I have not had such good luck with hunt, as it may target another inf team or something else, it will fire on them, then quit the 'hunt', and the main object can fire on it, but it can't see the object, yada yada yada. I don't know how to get it to restart the hunt after a target is found in the 1 min interval.

When teams are given a "fast" order, they focus on getting to their way point, and try to get there at all costs (unless they get pinned). That comes at the sacrifice of situational awareness.

Conversely, the "hunt" order is just about the complete opposite. They focus on maximizing their situational awareness on the way to their way point at the sacrifice of speed. And as a counterpoint to the "fast" order's "getting there at all costs", a "hunt" order cause a unit to halt once it spots a threat (be it actually seeing something, or seeing bullets whiz overhead, or into a buddy in the team). Once it has halted, it will try to preserve their life, which usually includes firing on the threat, if they can, and taking cover in the square they are in, which could as simple as hitting the deck. Since vehicles don't hit the deck like infantry, you usually just see them stop.

Those are two extreme ends of the movement order spectrum. For vehicles, "quick" and "slow" are the middle grounds between the two.

My suggestion in your scenario is to replace the "hunt" order with "slow". Your tank will continue on to the their way point with the "slow" order, but will have an increased situational awareness (thus a higher chance to spot the tank sooner). Also, since it is moving slow, it will have a much better chance of putting its rounds on target once it spots a threat.

Combine a "slow" order with a covered "target armor arc" order set at the end of the previous "move" order waypoint, and it will ignore any infantry threats. That should ensure a higher chance that turret will be pointing at the armored threat as soon as the threat is spotted, but it might bite you in the ass when that infantry contact your vehicle ignored turned out to have some AT asset that was able to give your tank a little love tap, because no machine guns were discouraging the infantry's "unwanted advances".

On a side note, I doubt that a "target armor arc" order placed before a "hunt" order would mean a vehicle would continue on despite infantry contacts. Based on what I remember from the manual that shouldn't work. That's based purely on my understanding of the manual. I haven't tried it out myself.

[womble]

Combine a "slow" order with a covered "target armor arc" order set at the end of the previous "move" order waypoint, and it will ignore any infantry threats....

On a side note, I doubt that a "target armor arc" order placed before a "hunt" order would mean a vehicle would continue on despite infantry contacts. Based on what I remember from the manual that shouldn't work. That's based purely on my understanding of the manual. I haven't tried it out myself.

Having held the same belief, I was persuaded some time ago that I needed to try it, and it does work. A unit on Hunt will not react to things it sees that it's not supposed to shoot at. It will still react (by stopping) to incoming that it perceives as a threat, just not for spotting or tentative spotting. This works for all Target Arcs, so if you're concerned about infantry ahead, say, but not to the (cleared, presumably) flanks, you could use a "Hunt-plus-normal TA in the front quadrant" combination to stop the movement if anything pops up to the front. In the TD's case here, a Target Armour arc would have been efficacious in keeping the M10 focused on their quarry and ignoring any pesky lurking Landser.

It can be worth considering using a non-circular TA to orient the main gun more precisely on the intended target at the intended stopping point, making the delay until engagement once stopped shorter, but constraining the arc too tightly (and "too tightly" is heavily situational) can result in potential targets of opportunity or threat being excluded.

[Duriel Krugaire]

In reference to the TA situation, I did have that on during one turn. I learned during that turn that it is a little complicated. As this turn involve the panther tank was to the north about 10 yds (sitting behind some burning HTracks) and east about 20yds. The TKiller had to go south 40 yds, go east 20 yds, then turn north 30 yds to arrive at the rear of the panther.

I had the target arc to the front of the tank, but when the TK moved the arc locating points stay stationary, so the panther was no longer in the arc.

I am going to mess around a bit with this last turn of the battle.

I was a little pissed a few turns earlier when I had 3 TK's sitting in overwatch on the road, the panther came into area from behind a bldg about 100 yds away & took out 2 TK's before either got a round off. So I was getting more skeptical about the in game AI.

[DVDBOATRIGHT]

Having held the same belief, I was persuaded some time ago that I needed to try it, and it does work. A unit on Hunt will not react to things it sees that it's not supposed to shoot at. It will still react (by stopping) to incoming that it perceives as a threat, just not for spotting or tentative spotting. This works for all Target Arcs, so if you're concerned about infantry ahead, say, but not to the (cleared, presumably) flanks, you could use a "Hunt-plus-normal TA in the front quadrant" combination to stop the movement if anything pops up to the front. In the TD's case here, a Target Armour arc would have been efficacious in keeping the M10 focused on their quarry and ignoring any pesky lurking Landser.

It can be worth considering using a non-circular TA to orient the main gun more precisely on the intended target at the intended stopping point, making the delay until engagement once stopped shorter, but constraining the arc too tightly (and "too tightly" is heavily situational) can result in potential targets of opportunity or threat being excluded.

Problem is the AI controlled units do this in 360 arc out to any range. They can spray through a small hole in a hedge and then spin around and spray any other small area 270 degrees away. It is easy for an algorithm to 'pixel bitch' every possible los, a human not so much. It is funny that individual team members of computer control units seem to place them self perfectly, yet the same can not be said for team members of player controlled units.

[*]the gunner is "Cowering" behind something that blocks his view while prone and foetal. If you have the SOP of setting a Face order (or short, directional Target Arc if you want them to hold their fire) at the "Deploy" waypoint, this is the biggest contributor by far to heavy weapons not being able to see what you expected them to be able to when they get where you sent them.

[*]the team has set up the gun in a position that can't see past a linear obstacle. I don't think this is an AI problem, because AFAICT, the TacAI isn't involved; the team sets up in the same positions relative to the direction they're Facing, whatever direction they're Facing, and if that's not "through" the linear obstacle, there's a good chance the "standard team formation" doesn't put the MG near enough the hedge to see through it. Face command is your friend here (though it's not 100% reliable, it's pretty good - my SWAG would be it gets your MG pointing right 95% of the time).

Another disconnect can be generated by the fact that the LOF calculations from a waypoint where a heavy weapon is going to be are calculated from the centre of the Action Spot, rather than the location in the Action Spot where the heavy weapon ends up. Again, Facing will help make sure the prediction and actuality coincide, but might not be perfect for a number of reasons, including trees.

The unit was not being fired at so it wasn't cowering and I had set the facing directly at the german mg and no team members had their faces pressed against a tree trunk. When I started the faffing around the german mg was busy tearing up another unit through a gap in the leaves of the bocage. When it finished with that unit it did a 180 and in less than second spotted and fired one burst killing half the US mg team.

[AttorneyAtWar]

Problem is the AI controlled units do this in 360 arc out to any range. They can spray through a small hole in a hedge and then spin around and spray any other small area 270 degrees away. It is easy for an algorithm to 'pixel bitch' every possible los, a human not so much. It is funny that individual team members of computer control units seem to place them self perfectly, yet the same can not be said for team members of player controlled units.

Your crusade to demonize the AI is a failed one I am afraid, if its a QB the ai unit placement/movement plan was made by a designer and placed "perfectly", if its in a campaign or scenario it was placed there "perfectly" all by a human. AI units are afforded no advantages, all of these problems you are bringing up besides LOS issues which can happen are all baseless, I am pretty bad at the game too its ok, but I still love it!

[womble]

Problem is the AI controlled units do this in 360 arc out to any range. They can spray through a small hole in a hedge and then spin around and spray any other small area 270 degrees away.

Well, first off, you might want to quote something that's relevant to the point you're trying to make. What has your kvetching got to do with Target Arcs making Hunting units less easily-spooked?

It is easy for an algorithm to 'pixel bitch' every possible los, a human not so much. It is funny that individual team members of computer control units seem to place them self perfectly, yet the same can not be said for team members of player controlled units.

What you're missing here is that the same algorithm is working for the player as works for the AI opponent. At the level of spotting and target acquisition, the units behave exactly the same for both sides. The executive AI doesn't troll its units along checking LOS every decimeter until it finds that needle-thin way through the foliage. There is no "pixel-bitching": it'd be far too CPU-intensive in CM's environment anyway; spotting is calculated on a periodic basis, the period of which for any given unit is determined by the same rules for both sides.

I believe that there are reasons above and beyond "confirmation bias" that mean we don't see as many reports of player-driven "miracle shots" through apparently impenetrable foliage for example as we do AI-piloted ones. Chief among these is that we players are better at siting our long range assets so they aren't required to shoot through impenetrable canopy. Second is that we players tend to be a lot more proactive with our assets than the AI is, which means we don't as often have guns sitting still "trying to roll a 1 on a d-million" to see through heavy concealment. Our units are rolling around, possibly getting fewer spotting chances at worse odds, so they don't get to see those outlier canopy "gaps". Confirmation bias remains the largest reason people whinge about this kind of thing though: they don't remember the times the AI has worked in their favour, because "of course I put that vehicle there so it could take that shot" (even if only subconsciously), but they are extremely put out when they're on the receiveing end.

[DVDBOATRIGHT]

Your crusade to demonize the AI is a failed one I am afraid, if its a <b>QB</b> the ai unit placement/movement plan was made by a designer and placed "perfectly", if its in a campaign or scenario it was placed there "perfectly" all by a human. AI units are afforded no advantages, all of these problems you are bringing up besides LOS issues which can happen are all baseless, I am pretty bad at the game too its ok, but I still love it!

Quite clearly you never played Razorback Ridge

[AttorneyAtWar]

Quite clearly you never played Razorback Ridge

I have, it was one of the first scenarios I ever played and I still haven't beaten it, however plenty of others have. Accept that you aren't very good at the game yet like I do and many others and keep playing and learning instead of complaining about something that doesn't exist. Infact, if you ask around people will be happy to help with things like unit placement and tactics.

[Blazing 88's]

I have, it was one of the first scenarios I ever played and I still haven't beaten it, however plenty of others have. Accept that you aren't very good at the game yet like I do and many others and keep playing and learning instead of complaining about something that doesn't exist. Infact, if you ask around people will be happy to help with things like unit placement and tactics.

+1..........

[DVDBOATRIGHT]

I have, it was one of the first scenarios I ever played and I still haven't beaten it, however plenty of others have. Accept that you aren't very good at the game yet like I do and many others and keep playing and learning instead of complaining about something that doesn't exist. Infact, if you ask around people will be happy to help with things like unit placement and tactics.

Thank you for your assessment of my skills, as a matter of fact I have beaten it ,all be it a minor victory only. And the mg unit placement was based on advice from this forum. Have you read the original post? I have progressed from V2.12 to V3.11 while playing this scenario and the difficulty has increased quite a bit.

for example

<b>Riding the Pig: or, how I learned to stop worrying and love Razorback Ridge</b>

So why does the crossroads work? Firstly, you can set up your mortars in the field to the left of the road on your start point - move one or two up to the end of the field where they can get LoS to covered areas beyond the crossroads.

Seemed to work in 2.12 now units get shot to pieces quickly by the eagle eyed mg or one shot from a mortar, I dread to think if it was the Panzershreck

No surprise that the crossroads is a prime target for german mortars, to don't stay stationary there for long - keep guys moving either in to the road to the farm, or down the left branch where you can go through a gap in the bocage into the field between the two roads

Doable in 2.12 after that an mg from across the river can hit the left hand edge of the cross roads. Also in the right hand lane in 2.12 an mg from across the river could get you when you cross the gap after 2.12 lining up on the left of the lane further back from the gap can get you hit too.

All my tries have been from a saved game made under 2.12 so the plan has not changed.

[womble]

Quite clearly you never played Razorback Ridge

You do realise how ridiculous it is to assert that the flawed scenario is a demonstration of how the AI is "suspect"? There are a bunch of problems with Razorback Ridge that gamey reading of the VC descriptions solves, but it remains a scenario whose difficulty is heavily dependent on precisely where/if you took previous losses, even once you've realised that the requirements to get a victory are pretty easy (so long as you haven't lost the wrong half-platoon).

The AI doesn't cheat. It has no advantages over you that you do not give it. The designer may "cheat", since he knows what you've got, but that's his perogative, as the story teller in a given scenario. It's part of his job to make things interesting.

Your crusade to persuade people that the AI cheats is doomed from the start, because it doesn't (or if it does, it cheats the same for both sides, if the circumstances are the same), and that's blatantly obvious if you play the game with any kind of critical faculty turned on.

[DVDBOATRIGHT]

Thank you private for saving a little bit of my sanity.

[General Jack Ripper]

This is in reply to the original post.

 

DVDBOATRIGHT, on 05 Dec 2014 - 09:28 AM, said:

Also why can't US MGs not sight and fire through bocage when sitting behind it while German MGS can quite happily fire from there and set back from the bocage.

 

Short answer: because German MG's can be fired without being deployed, giving them much greater flexibility and reaction time.

 

Long answer: I noticed the machinegun problem a while back: (Caution: Harsh Written Language)

 

http://youtu.be/a99QjvlpiIM

 

When using the editor, or deploying troops in a battle, you might need to re-deploy your machineguns to the same action spot more than once.

 

I managed to fix that .50cal MG in the video by un-deploying the weapon, moving the team back one AS, then back to the AS they were in with a face command added, then deploying the machinegun again. After that, everything worked fine.

 

 

Reading ahead I find this:

 

womble, on 05 Dec 2014 - 2:31 PM, said:

Another disconnect can be generated by the fact that the LOF calculations from a waypoint where a heavy weapon is going to be are calculated from the centre of the Action Spot, rather than the location in the Action Spot where the heavy weapon ends up. Again, Facing will help make sure the prediction and actuality coincide, but might not be perfect for a number of reasons, including trees.

 

This is exactly true, and leads me to another of my rule of thumbs: "Never assume a machinegun or other heavy weapon will deploy exactly where you want it to. If things get messed up, check your LOS and try to deploy again with a face command or target arc active."

 

Using machineguns in bocage, or fighting in bocage in general is a major pain in the a**, and requires a good amount of practice and training to conduct effectively. Even long time Combat Mission veterans screw it up occasionally, as anyone who watches my youtube can attest.

 

To turn around and scream "I can't win! The game is broken! The A.I. cheats!" is just childish. Combat Mission is a serious game, and should be taken seriously. The first thing you need to do is unlearn everything you know about war games, and start all over again. Should be easy, right? Then you need to realize that like all simulation level games, sometimes things don't work perfectly. However, as far as the machinegun problem is concerned Womble and I have told you our workarounds for a bad deployment, but it is up to you to listen and learn.

 

Welcome to the forums, feel free to ask questions.

 

- SLIM