Matrix Games Forums

As a past player of Second Front (on the amiga) amd WiR (on the PC) the AI in both cases was poor, and I would say the majority of players of this game would play against the AI as opposed to a human opponent, so when I purchase this game, I will be hoping and praying that I will be tested to the enth of my abilities by the AI ???

Will it make me break out in a cold sweet, and say that the game was hard fought out and a well deserved win for the AI ??? Hope So !!! As I would like the AI to win honestly in this new improved version.

I will confess to using a hex editor to hack the WiR game to improve the AI's odds against me, that the the rudimentry editor could not or would not do, but even with all this help I gave it, it was still easy to beat.. Stopped hacking and playing in the end, as there was no contest. Too difficult playing against a human opponent due to time availability, and time difference constraints etc. So it's the AI or nothing, so I truly hope it is good this time round..

Thankyou

ORIGINAL: shane_056

As a past player of Second Front (on the amiga) amd WiR (on the PC) the AI in both cases was poor, and I would say the majority of players of this game would play against the AI as opposed to a human opponent, so when I purchase this game, I will be hoping and praying that I will be tested to the enth of my abilities by the AI ???

Will it make me break out in a cold sweet, and say that the game was hard fought out and a well deserved win for the AI ??? Hope So !!! As I would like the AI to win honestly in this new improved version.

I will confess to using a hex editor to hack the WiR game to improve the AI's odds against me, that the the rudimentry editor could not or would not do, but even with all this help I gave it, it was still easy to beat.. Stopped hacking and playing in the end, as there was no contest. Too difficult playing against a human opponent due to time availability, and time difference constraints etc. So it's the AI or nothing, so I truly hope it is good this time round..

Thankyou

Let's be real here. You and I both know the AI will suck.There is not one game, of this scale, that has a good AI. I don't think this will be any different.

In all of these types of games, the enjoyment one gets out of the AI depends hugely on your expectations and your conception of what an artificial opponent is supposed to be. As a veteran of Grigsby games dating back to the early 1980s, across multiple platforms, I've always managed to enjoy the games solo even when (as was usually the case) an objective measure of the AI's tactical or strategic abilities would find them wanting. Maybe it's because I (mis)spent my youth playing board wargames solitaire, given the difficulty in finding reliable opponents, but I approach most historical computer wargames with a sort of solo player's mentality. It's the process, not the result, that matters.

So I don't go in thinking that the AI has to be a veritable Manstein or Zhukov; I'd settle for simply presenting me with a framework within which bad decisions on my part get punished and good ones are rewarded. I'm not concerned if I actually win all the time; I'm not looking for a purely competitive challenge in that regard. Rather, I'm looking for a structure within which to "play out" historical campaigns in a fashion that is intellectually satisfying. I therefore tend to refrain from using every nuance of the rules to crush the AI, but limit myself instead to things that I feel are more historically accurate or relevant. Should the AI be good enough to make such circumlocutions unnecessary? In a perfect world, sure. Should the AI be able to deliver a tense battle, with the outcome in doubt? Definitely, if we could write the script for the game industry. But I don't really use those desires as the bar for my expectations. As long as the AI doesn't do something really abysmal--routinely move every unit to the upper-left corner and pile them all into one hex, or something--I'm usually ok, in games with as much historical detail and "chrome" as this one. The simpler the game, the more apparent the AI quirks are, but the opposite is true as well. A very complex game can hide weaknesses in AI behind detailed systems, and often this helps make the experience more palatable.

None of that helps if your primary interest though is competitive play, naturally.

ORIGINAL: TheWombat

...As long as the AI doesn't do something really abysmal--routinely move every unit to the upper-left corner and pile them all into one hex, or something--I'm usually ok, in games with as much historical detail and "chrome" as this one...

Nah, we fixed that bug already. [:'(]

I disagree... I think the game should not be released untill the AI measures up. There is way to much junk out already. It shouldn't be that hard to have a game cover objectives, Spend its resources each turn, and to program stuff like Hq's, air bases and other assets to keep back and out of harms way.
In an AI game I'd like to see.
1) garrison each city and river rail crossing with a unit that doesn't move. Form defensive lines around Lenningrad and in front of Moscow with these same units.
2) keep a continues line of units in front of the human player and keeping them at the highest level by stripping assets off the garrison units not within 25 hexes or so of an attacker. Check strengnt levels before each sides turn. Filling the units that are in front on the non AI players units. Have trigger hexes that make the line fall back and incorperate the garrison units where poss.
3) Force tha attacker to surround and reduce the pockets by assigning an extra 1/3 to 1/2 of all combat losses to be re assigned to the AI force pools unless surrounded.
4) Change the Air system in an AI Game to insure every plane flies and every effort is made to insure units stay in supply.
Most of this is for the def AI. But by changing unit strenght levels at certian points in the game and pouring the rest of the resources into offensive units. It may even be possible to have the AI take the offensive.

My .02 worth

ORIGINAL: TheWombat

So I don't go in thinking that the AI has to be a veritable Manstein or Zhukov; I'd settle for simply presenting me with a framework within which bad decisions on my part get punished and good ones are rewarded. I'm not concerned if I actually win all the time; I'm not looking for a purely competitive challenge in that regard. Rather, I'm looking for a structure within which to "play out" historical campaigns in a fashion that is intellectually satisfying. I therefore tend to refrain from using every nuance of the rules to crush the AI, but limit myself instead to things that I feel are more historically accurate or relevant. Should the AI be good enough to make such circumlocutions unnecessary? In a perfect world, sure. Should the AI be able to deliver a tense battle, with the outcome in doubt? Definitely, if we could write the script for the game industry. But I don't really use those desires as the bar for my expectations. As long as the AI doesn't do something really abysmal--routinely move every unit to the upper-left corner and pile them all into one hex, or something--I'm usually ok, in games with as much historical detail and "chrome" as this one.

I agree with all of the above and however poor the AI in WIR might have been, I had many hours of good play. I am sure that WiTE will have a much more competent AI, so it can only be better. [:)]

ORIGINAL: critter

I disagree... I think the game should not be released untill the AI measures up. There is way to much junk out already. It shouldn't be that hard to have a game cover objectives, Spend its resources each turn, and to program stuff like Hq's, air bases and other assets to keep back and out of harms way.

My .02 worth

The AI has to be fallable, just like humans are, war is full of mistakes and lost opportunites. The skill is in exploiting your enemies' errors, no errors - no game. Not that the AI has to be a push-over, just a little 'human'. [:)]

The goal of a realistic and challenging AI is sort of a holy grail in computer wargaming, and like the Grail of legend I rather doubt we'll see it in its entirety in our lifetimes. Critter's criteria seem relatively realistic, but the problem with lists like that is that they are either a lot harder to actually implement than they seem, or when implemented they result in predictable and stilted gameplay. As students of history we're in a lot better position to manage these battles than the poor sods who actually had to do it in real time with real lives on the line, and with very imperfect information. What's more, we generally don't have any of the intangibles driving us--the emotion and fear and ambition and sense of real gambles with real consequences that inform so many military decisions.

Yes, I certainly hope that the computer opponent in this game makes the sorts of decisions Critter is highlighting, and I hope it behaves reasonably. But I also appreciate that one person's reasonable action can be another's rash decision, and that humans fighting wars are often, um, a bit odd about how they do things. So I'm not that worried. Besides, at the rate I learn games these days (sloooowly) the complexity will probably mean I won't get to the level of exploiting the AI's weaknesses until long after I've gotten my money's worth. That's not to say that an abysmal AI would be acceptable, but my definition of "abysmal" is fairly liberal.

The AI is becoming quite good on defence. It's pluging all the holes, retreating from forming pockets, and attacking weak exposed units. Theres still room for improvment. As far as the offence theres alot of work still to be done, and is much harder to program to become effective. But as I said before, Gary is spending alot of time on it...I do agree the AI on WIR was very poor...

this game begs to be played h2h, but as we may want to see a great ai, without fudging that will be nearly impossible...
WE gettoooooo much info oin almost every game that really during war no one even dreamed of having... things just are not as neet and tidy as our games...
Then if one models restricted reactions, poor hq or communications players gripe about a games straightjacketing them, I know I do, it s not a simulation its a game, and to ask for an ai that will challenge a grog, straight up no fudging, is silly imo
on the other hand it is good to have out there goals, otherwise we would never venture into the land of what if and why not, those twin cities near Detroit!

I visit a number of game forums and read the comments, and one of the most discussed topics is the AI, or lack of AI, as any game delivered without a strong working AI is severely restricted in it's appeal, and consequently how much money it is really worth and sells for, so any development time and resources spent in developing a strong AI is virtually money in the bank, as it will help to sell and more importantly keep selling this game, which I hope it will sell in large numbers so as to justify further future development.

Probably right in that we wont get a Zukhov or Lewinski (Manstein) both equal 8.5 out of 10, but should be aiming to get a Konev or Model (both 7.5 out of 10) out of the AI
Thanks

A lot of effort is going into the AI.  It will challenge many but perhaps not everyone.  There are several different levels of difficulty.

And there is always fudging, if the ai is not tough enough for you.. add power to it... that is not so hard.... and fow makes for really intersting ai, as one gets the added depth of , what does the enemy have over there feel..
again this looks great guys and I in no way meant to say it was a poor ai, just that IA will never be a person

I agree on the importance of not neglecting the AI only becuase it will never be good enough. Most players will play solo and so requires an adequate AI to have any enjoyment of the game. I remember buying Commander - Europe at war and stopped playing after the first try because of the bad AI. I don´t want that to happen to this game.

Just because it is hard to make a good AI or that no AI is good enough is not reason not to put huge resources into trying to build the best AI to date. Nothing is more satisfying than to be impressed and outmanouvered by an AI that is not cheating.

To let the AI cheat by giving it more strength or resources is ridiculous. I never play against anyone who is cheating. The computers drawback is the lack of a deep human thinking but it can compensate that many times by its huge calculating power to compare and plan where and when to attack or defend or by calculating odds.

To improve the AI to give it the feel of playing against a Human, I did suggest to the developers to digitise little Munchkins and then add them to the game.

to be honest, to make it play like a human, you may as well let it cheat, most humans I have ever played, cheated there butts off any time they thought they could get away it

ORIGINAL: Hard Sarge

to be honest, to make it play like a human, you may as well let it cheat, most humans I have ever played, cheated there butts off any time they thought they could get away it

Well, true enough but still.....as long as I believe they don´t [:)]

ORIGINAL: karlxii

To let the AI cheat by giving it more strength or resources is ridiculous.

I agree, no bonuses or combat modifiers for the AI which would inflict higher losses to the human player's troops. What can be put in place are presets like "normal AI" and "strong AI", if one wants to put modifiers in.
Lower FOW for the AI, plus the ability to detect enemy units and attempting to plug holes in the front line are needed in order to provide for a decent challenge

ORIGINAL: karlxii

Just because it is hard to make a good AI or that no AI is good enough is not reason not to put huge resources into trying to build the best AI to date. Nothing is more satisfying than to be impressed and outmanouvered by an AI that is not cheating.

The basic problem is that at this level of gaming there are not huge resources to put into anything, including the AI. It's the feel of the game that counts and however primitive the AI is in WIR, it still gave, and is still giving, a good gaming experience. I am looking forward to the next level of gaming experience, but I am not expecting a perfect AI, nothing in war is perfect. [:)]

I am not expecting a perfect A.I. but I expect that it is very high up in the list of priorities. I expect the A.I. to supercede 8-bit computer games A.I from the 1980:s like Kampfgruppe, Typhoon of Steel, Fire Brigade, Second Front (1990) or you name it.

I don´t care about fancy map graphics or good sound effects, or perfectly researched historical name for every single brigade if the A.I. don´t deliver. Everything comes down to how to prioritze the resources at hand. The graphics might draw people to it but the A.I and a friendly user-interface will keep them happy with it.