Sorry, long-term fan who's had enough

[Platoon_Michael]

You guys are playing other projects when the last 3 projects are incomplete?
Now that's playing games.
Whats even funnier is the map was codded right the 1st time around and someone decided to change it to the point where now the player can't do anything about it.

I'm done reporting your failures to you and for you.Go do it yourself.


Oh that's right...............
You can't.You have better things to do.

Well,
At least the next consumer hopes it's better.[:D]

[GaryChildress]

ORIGINAL: AdamRinkleff

ORIGINAL: Erik Rutins

Hi guys,

There have actually been a huge number of improvements from the original CC series to the present, including the area of AI. With that said, we made more AI improvements in the first update which have been well received so far and we're completely overhauling pathing for the second update (basically rewriting it instead of applying fixes on top of it) for the first time since the original series. We are as interested in a "better CC" as all of you, but I'm surprised by comments that things are the same as they have always been. I think you really have to look past a lot of positive changes to believe that.

Regards,

- Erik

We have noticed improvements, improvements in graphics, improvements in fire support, improvements in night fighting, all kinds of improvements. Ive even noticed improvements in the AI, it does engage in some more interesting tactics, it does use smoke more, and pathing has been improved.

But you have NOT addressed the fundamental problem in which the AI tends to mill about aimlessly. You have NOT addressed the fundamental problem in which the AI tends to just leave troops sitting hidden somewhere doing nothing (especially with AT guns), when the AI is supposed to be attacking. You have NOT addressed the fundamental problem in which the AI sometimes refuses to attack even when it has overwhelming strength and should be attacking. You have NOT addressed the problem in which the AI decides to attack a location, and then gets slaughtered, and keeps attacking taking outrageously high casualties. None of those problems are difficult to fix, but they do take some creativity and a willingness to do so.

The fact of the matter is I can mod the game to give the AI 150 soldiers (unfortunately you refuse to allow me to add more squads, which is another problem that should be fixed, because limiting us to 15 squads is unreasonable) -- and yet, despite being grossly outnumbered, and despite giving the AI extra armor bonuses, and extra weapons, and extra everything... I can still destroy the AI, because it fails to use all of its troops, it fails sometimes to attack, and when it does attack it fails to take into account its casualties and consider attacking at some other location where I must presumably be weaker.

Its not genuinely 'AI' until the computer decides to do something, and then changes its mind based on the fact that I've got two machine guns in that building, and I'm slaughtering everything coming down the street.

Talk about lousy AI, I've been playing Warhammer 40K DOW I series lately. Basically the AI simply attacks relentlessly, runs right straight into an inferno of guns and missles and doesn't seem to care how badly its troops get mauled. You can use the same tricks to destroy the AI base over and over and over again. It never learns anything from past mistakes. Changing the difficulty settings seems to simply change how easy or difficult it is to kill the AI's soldiers and how many of them it sends at you at a time and little more. But the game is fun until it gets boring. WH40K is a big commercial release but it doesn't have anything I would call "intelligence". Basically it just spams units and sends them headlong at your base. That's the only tactic it knows and once you figure out how to overcome that one tactic you basically will never lose again. The same thing is true of the whole Command & Conquer series and just about every other RTS series out there. The number of variables in your basic big commercial RTS game is relatively low compared to real life.

Why is it that CC's AI is "broken" when WH40K apparently isn't? Why is it that we expect CC to be so much more complex and intricate than a game put out by a HUGE commercial gaming company like THQ? If THQ can't figure out how to come up with realistic AI, then how do we expect Matrix to? Granted I want to see improvements also but maybe we should be a little more realistic in our expectations? Yes I would like to see an AI which mimicks reality as closely as possible.

Maybe Matrix should simply put out games like Command & Conquer or Wahammer and stop trying to do simulations. It would be a lot easier, a lot more profitable and there would be less criticism.

[GaryChildress]

Furthermore it seems that Matrix can't win. When the AI attacks people complain that it just sends troops to the slaughter. When it doesn't attack people complain that it isn't attacking.

[Fred98]

People will recall that for Deep Blue to defeat a human at chess, deep Blue did not use any strategy nor tactics.

Rather, there are 32 pieces on the board and 64 squares.

Wherever the pieces happen to be on the board, there is always the perfect move.

The ere a zillion possible positions the pieces could occupy and always the perfect move.

Each combination and it’s perfect move was recorded in memory. Deep Blue simply accessed it’s memory.

This is not possible for Close Combat.

In theory the AI player might start a game with 15 units on the map. Let it begin with 30 units without the human player knowing.

This would make a AI player as tough as a human player if not as good as an human player.

-

[Peterk1]

The chess analogy as you presented it is not entirely correct here.
It didn't simply access it's memory. It's able to evaluate the value of each position due to a scoring algorithm programmed by experts. There is no "perfect move". There are moves with higher scores than others.

You can do the same thing in CC or any other wargame if you are inclined to. Most people are not inclined to do chess-like thinking in non-chess games because it's challenging.

But for CC....

Computer guys in open ground + human has unsupressed machine gun + lots of computer side's bodies in front of machine gun = really, really bad score. Recommendation: get guys to cover, start firing on MG.


Computer guys in cover + Computer  guys shooting at enemy positions + human guys are slowly breaking = good score. Recommendation: wait until break and then charge.

Why not?

In CC's defense, the AI is actually smarter here than in the original CMs, I find. CM didn't know how to use a mortar or machine gun in attack, didn't know how to find positions for those units in attack. Setups were perhaps better in CM but I haven't played it for a long while, so I could be wrong about that.

[GaryChildress]

ORIGINAL: Peterk1

The chess analogy as you presented it is not entirely correct here.
It didn't simply access it's memory. It's able to evaluate the value of each position due to a scoring algorithm programmed by experts. There is no "perfect move". There are moves with higher scores than others.

You can do the same thing in CC or any other wargame if you are inclined to. Most people are not inclined to do chess-like thinking in non-chess games because it's challenging.

But for CC....

Computer guys in open ground + human has unsupressed machine gun + lots of computer side's bodies in front of machine gun = really, really bad score. Recommendation: get guys to cover, start firing on MG.


Computer guys in cover + Computer  guys shooting at enemy positions + human guys are slowly breaking = good score. Recommendation: wait until break and then charge.

Why not?

In CC's defense, the AI is actually smarter here than in the original CMs, I find. CM didn't know how to use a mortar or machine gun in attack, didn't know how to find positions for those units in attack. Setups were perhaps better in CM but I haven't played it for a long while, so I could be wrong about that.

I don't know how realistic using a chess program to mimick all the confusion on a real battlefield would work. In reality commanders make mistakes as much as they make brilliant moves. If I were playing against an opponent who never made a mistake and made its decisions in real time a million times faster than I could make my decisions I don't think I would call that a "simulation" either. Ultimately CC is a simulation and it needs to act like a human, not a computer.

That does bring up an interesting military possibility though. If IRL a commander could simply program a computer to make his decisions instantly for him and never make a mistake based upon a given situation, think of what war in the future may be like? [X(] Just a side note...

[Peterk1]

I don't think even chess programs claim to find the perfect moves. But with the system that most of them use, they do have a great ability to filter out all of the patently bad moves. In most chess positions, most of the moves you =can= make are bad or don't do anything for you.

It would be great if CC could start doing doing that. I've played about 60-70 games of The Longest Day in the past few weeks and the main quibbles with the AI that I have found are:

1. Not setting up in cover at the beginning of a game. Even if you're planning on rushing a position, don't most players always set up in cover, just in case the opponent's set up is not what one hopes????

2. Keeping too many units out of play at the back at the beginning of a game. Might make sense for a long game but with 15 minute timers, pretty much every unit has to to put to use.

3. Starting off many games with a suicide banzai rush which almost never succeed and which almost always give me an easy win due to morale breakdown.
This one reminds me of a little 6 year old who starts off every chess game going for the scholar's mate. It's going to work maybe one game in a thousand. This is a "bad move" which should be removed from the AI's repertiore.

4. Not appearing to ever use the basic tactic of suppressing a target location with many units before moving in to take the location.

Whether some chess-like algorithm is used to score strategies or not, you should still have some basic way for the computer to decide whether something it is about to try is good or not. And definitely it should have most of the common tactics used by human players in its repertoire. 

[GaryChildress]

ORIGINAL: Peterk1

3. Starting off many games with a suicide banzai rush which almost never succeed and which almost always give me an easy win due to morale breakdown.
This one reminds me of a little 6 year old who starts off every chess game going for the scholar's mate. It's going to work maybe one game in a thousand. This is a "bad move" which should be removed from the AI's repertiore.

Sometimes suicide banzai rushes work. See my AAR in the AAR forum. Basically it is the ONLY way to take some of the bridges in the game intact BEFORE they are blown. Part of the problem seems to be that the AI doesn't put all its weight into the attack. It will rush two or three teams at once and you simply mow them down. If it put ALL its units into the attack that might work better.

EDIT: BTW morale seems to be modeled very well in the game. Whenever I did a banzai rush I had to constantly keep ordering my soldiers to move toward the objective. The soldiers would break down and retreat when the opposing fire was too heavy. I have to say kudos to the makers of CC for modeling morale in the game. It does appear to make it VERY realistic.

[7A_Woulf]

ORIGINAL: Gary Childress
EDIT: BTW morale seems to be modeled very well in the game. Whenever I did a banzai rush I had to constantly keep ordering my soldiers to move toward the objective. The soldiers would break down and retreat when the opposing fire was too heavy. I have to say kudos to the makers of CC for modeling morale in the game. It does appear to make it VERY realistic.

Second that!

Some of my general observations on the game after the patch:

1) With a few exceptions, the tactical AI is more likely to act aggressive if they already have 'combat-contact', i.e. if you play a pure defensive battle it's more likely that you have to sit and wait for 15-20 mins.

2) The strategic AI is generally more aggressive, sadly sometimes to a point where it's pure madness to attack and I personally would retreat/wait and 'rest' that BG. (Guess it's hard to find a nice balance on this behaviour?)

3) AI controlled KG Jungwirth at Dinther + attack on Veghel = slaughter! Sadly, it's a duck-shoot when they attack. Due to their deployment zone in this case all I have to do is deploy a L-shaped defence south of the canal and west of the bridge, then I can let loose and win the battle in 4-5 mins. (Once I used the mortar barrage, but that felt like cheating! [;)] )

4) In general I like the mechanism that makes elements of disbanded BG's joining other friendly BG's of the same parent unit, but in my recent GC it has created some strange effects:
-The survivors of 107th PZ. Brigade and KG Rink have managed to cross the Vilhelmina Canal, the rivers Dommel and Aa before joining KG Jungwirth, all without controlling bridges and while 'retreating' from battles where their BG's have been disbanded! [8|]

[RD Oddball]

ORIGINAL: Platoon_Michael

You guys are playing other projects when the last 3 projects are incomplete?
Now that's playing games.
Whats even funnier is the map was codded right the 1st time around and someone decided to change it to the point where now the player can't do anything about it.

I'm done reporting your failures to you and for you.Go do it yourself.


Oh that's right...............
You can't.You have better things to do.

Well,
At least the next consumer hopes it's better.[:D]

Actually you didn't take Neil's comments how he'd meant them. His comments were in reference to WaR. It would've been more accurate to his meaning if he'd said: "Although testing of WaR continues in anticipation of it's next update the main testing focus is on LSA at the moment so we can get the update out the door and into customers hands as soon as possible. If you happen to find any bugs or areas you see as improvements on WaR or any other CC build to please file a bug report through the Matrix bug reporting system as that will be the most sure way something doesn't get overlooked when we make the next update for WaR or TLD." If you carefully look at what he'd said that is what he'd intended.

Rest assured we're not going to take on any new projects without addressing any reported issues on WaR, TLD and LSA first. You can best help realize your desires for WaR by reporting anything you feel needs to be looked at. We'll be glad to take a look at it for the next WaR update.

[Erik Rutins]

I'll echo what Jim said. Please give us a little benefit of the doubt folks.

[Andrew Williams]

4) In general I like the mechanism that makes elements of disbanded BG's joining other friendly BG's of the same parent unit, but in my recent GC it has created some strange effects:
-The survivors of 107th PZ. Brigade and KG Rink have managed to cross the Vilhelmina Canal, the rivers Dommel and Aa before joining KG Jungwirth, all without controlling bridges and while 'retreating' from battles where their BG's have been disbanded!

that's being looked at as it seems it masy be to your advantage to have you BG disbanded under certain circumstances.

[Platoon_Michael]

It's hard for me to have any patience with you guys because.......

1) I'm more than willing to bet you guys don't look at anything when testing,other than stuff that would crash the game.And the severly MASSIVE graphic issues that even Stevie Wonder could find.
I have found so many errors with WAR it isn't even funny anymore,and god knows what else I will find.

2) Butgenbach:Have you guys even looked at the elevation on that map,and the lack of shading? (The most extreme error I can point out)
and yet you will look into the littlest comments that xe5 has pointed out for LSA
(no mean intentions but come on, WAR has been out for 3 years and no one goes back and says whoa this isn't right?)

3) I'm more than willing to bet you guys are 2 to 3 patch's ahead of us who are actually playing just by viewing your posts.

4) Martelange: View #3 (you guys aren't even on the same page when it comes to maps,as pointed out since 2008)


Bug report?
That's a new one.
Why don't you show me the bug report for WAR since it's release by the beta testers?
This way I can show you who to fire as you still don't seem to be able to figure that out after 3 years of testing and a free $50 game.
And if you your not, and your going to stay with the tried and true of using customers to find errors why not give them some sort or kick-back?

[RD Oddball]

Please do report any errors you find that's the best way to ensure they get taken care of. We look at any and all issues that are reported. Massive graphics errors? Please illustrate them in the WaR forums. We have not seen any graphics errors. Also please illustrate the problems you are perceiving with the Butgengach map there as well. I just looked and don't see any reports about graphical errors or anything about Butgenbach. If I'm missing your posts please post a link. Same for the issues you're referencing on Martelange.

There has been testing on WaR since the release of the last patch of 9 weeks ago as a continued effort to continually improve the releases we've made available. It's not reasonable to think any program is ever going to be perfect. Even with the massive resources of a company like Microsoft it's still not possible for them to release an OS that's anywhere near perfect even after years of polishing. Not saying it's an excuse to give up or not try to release a perfect build. We do the absolute best we can. The point being that even with our best effort there are bound to be issues that get past us and as customers you can benefit your investment by reporting any issues that did get past us so we can take care of them as we've promised to do.

[TheReal_Pak40]

I have to agree with Michael. My patience is just about gone. I've bought every original CC game and two of the re-releases. NOTHING has been improved in the tactical engine since CC2:ABTF:

TacAI: just as stupid as it always was. Blindly sends it's troops to straight to the nearest VL only to be slaughtered.

Deployment AI: It's the exact same as it always was - HORRIBLE. AI always deploys around VLs with no consideration to cover or fields of fire.

Vehicle Pathing: WORSE than in CC2. I'm not just talking about the "can't go there" remarks, I mean the actual path that the vehicle takes once given a move order. Tanks on a road are given an order to move 40 meters straight up said road, but the driver decides that a scenic ride over a wall and through the woods is the path of least resistance. Honestly, Miss Daisy could drive better than that.

Instant Artillery: Wow, offboard artillery falls so quickly that there isn't enough time to realistically allow for the flight of the round, let alone the whole process of communication. And it's always perfectly accurate.

Instant On Board Mortars: The main reason why AT guns don't last more than one or two rounds after opening up. Mortar rounds land 2-3 seconds after the fire order is placed. The is completely unrealistic. The flight path of a mortar round takes at least 10 seconds alone, not to mention the whole process of adjusting the mortar to get the correct direction and distance.

LSA MGs: As if MGs in CC weren't powerful enough. I kind of liked it better when a stone wall could actually stop a MG42 bullet.

With exception of the last one, these are all legitimate complaints from your consumers for the past several re-releases and the original Atomic releases. However, like Atomic, Matrix has solely focused on re-releasing new games with new maps and not actually fixing their product. The only thing that has improved is the strategic campaign engine. This a damned shame since Close Combat's primary focus is, after all, about close combat.

Honestly, should I have any more patience after ten releases of Close Combat?

[RD Oddball]

Have you tried the first update Pak40? There are quite a few customers who would disagree with your assessment of the AI. We're working on vehicle pathing and an update will be out soon. We want to make sure it's properly put through it's paces before releasing it so we won't be rushed into making it public before it's time. Continue to be patient.

[e_barkmann]

Just to voice another opinion here, I have bought all CC's including the CoI and WaR Matrix releases but was not sufficiently convinced that things had improved enough in the single player tactical game to purchase TLD and LSA. 

My main gripes are uninspired AI and 'infantry survivability' (or lack of it!)

I look forward to seeing feedback on the next LSA patch and am crossing my fingers it's going to be positive.

cheers Chris

[Dundradal]

The AI is improved in the first patch. It's definitely an improvement over the release version.

I'm hoping the path finding in the second patch will be an improvement.

I'm with Chris. I've got my fingers crossed we'll get some good improvements.

[Peterk1]

ORIGINAL: Gary Childress
Sometimes suicide banzai rushes work. See my AAR in the AAR forum. Basically it is the ONLY way to take some of the bridges in the game intact BEFORE they are blown. Part of the problem seems to be that the AI doesn't put all its weight into the attack. It will rush two or three teams at once and you simply mow them down. If it put ALL its units into the attack that might work better.

Hi Gary,

But an AI should not do something frequently just because it is the thing to do in some very special situations (ie. something has to be done very quickly). A chess player will sacrifice a queen when it sees that it can win the game by doing it. But it's not going to sacrifice a queen just because it MIGHT be able to win. In your case, the "score" of the catching the bridge intact in the next few minutes would outweigh the loss of some "pieces" so it makes sense to take risks. Or in The Longest Day, maybe it makes sense for the first wave on the beaches to push hard for an exit fast. But it's another special case.

I don't think the rushes make sense in 99.9% of the cases. Maybe to model Japansese or Russian charges...but even then the CC AI is way more suicidal than even the most hardened Russian commissar would approve of. Most of the time, from what I've seen, the AI would do better by playing a more patient game. A campaign/map defender should only start attacking seriously once it has the advantage or its position in the game is lost...it should not attack itself into oblivion as a normal case.

Just a couple of ideas to bump up the percentage of satisfying match-ups against the AI which wouldn't require any super-sophisticated evaluation routines:

The AI could cheat a bit to check whether a charge on a flag is likely to work before the game starts just by checking the set-ups. Something like...

AI has 20 men heading towards position covered by 15 rifles and an MG in buildings, and two mortars waiting....UMMMM....probably not going to work. Don't do it. 30 versus 5 rifles and one mortar? That might work. Go for it. Some little rules of thumb could be used to cancel out the massacres before they even start. Yes, the AI would then be cheating, but if it results in a more satisfying game for the player, I would say go for it.

How about the AI doing it's set-up AFTER it sees the player's setup? Or at least just shifting a few units around to better counter what the player has indicated he intends to do via its setup? I have no idea if it's already doing that, but my gut feeling is no.

[mooxe]

ORIGINAL: RD_Oddball
There are quite a few customers who would disagree with your assessment of the AI.

Those quite a few customers voiced their satisfaction within 48hrs of downloading the patch and we havent heard much since. I suppose they wanted things to be better so badly that when they spotted the AI doing something smart, that instantly meant it was better. Theres no way the changes that were implemented were tested throughout all the different BG commanders attributes.

ORIGINAL: RD_Oddball
We're working on vehicle pathing and an update will be out soon. We want to make sure it's properly put through it's paces before releasing it so we won't be rushed into making it public before it's time.

...does this mean another beta patch for us to test? haha..