Honestly How Competant Will the AI be this time ?

[Rasputitsa]

ORIGINAL: karlxii

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.

I am no expert on game development, but I would guess that most of the elements you list are easier provide and will enhance the game. However, super AI takes huge development and realistically this genre of game has a smaller, if enthusiastic group of followers. This will never be a mega-bucks part of the industry, so I will be thankful for what we get and from what the beta-testers are saying, it should be good. I enjoy WIR and like some of the quirks it has, as it brings some of the realistic chaos seen in the actual history of war. [:)]

[elmo3]

If you are expecting a Deep Blue quality AI that can beat the Garry Kasparov's of wargaming then you will probably be disappointed.  However we are working hard to make a challenging AI and I think most players will be happy with the results.

[Fred98]

Every AI is poor and this one will be poor too. (except SSG’s Kharkov AI who is excellent!)

We will never get a Deep Blue. To make the AI challenging do not allow it to cheat. Do not allow it to have extra bonuses.

All you do is change the scoring. As I have mentioned in other threads, any objective might be worth 10 points per turn against a human player but against the AI it might be worth 15 points per turn.

Then, that scenario is tested and balanced and will be challenging.




*Any player who finds that scenario challenging should never play a human player. My opponents can record a loss on the ladder now and save time.

[vonRocko]

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

That's some crowd you're playing with. The guys I game with would never even think of cheating, it is just a game after all. (unless money is on the game)[;)]

[Max 86]

I have always said that the ai is primarily for the casual gamer and like others have said it is unreal to expect an ai that will challenge the everyday player that logs a lot of hours. But for the guy that gets to play maybe 2-3 times a week for a couple of hours each time, the usual ai does just fine. And by the time it gets old there is a new game of some type just waiting to be bought.

I'm not sure but would wager that most wargamers fall into the casual gamer category and thats why Matrix always pushes for an ai component in most, if not all, of their releases.

[TheWombat_matrixforum]

While I've been wargaming for a long time, I'm pretty much a "casual" gamer in that my wargaming is exactly as mentioned, a few times a week for a few hours at a stretch. So AI play is important but I'm never going to put the hours in to "master" the game system. Therefore, it's probably a lot easier to challenge me than some of the pros on this board!

Also, I agree the AI in SSG's Kharkov is very good, as it has been in most of their games. But, and this to me is a big factor, they accomplish this IMO by tuning the AI to play the victory conditions, e.g., the game qua game, rather than to emulate historical strategies and operational parameters. Don't get me wrong, I love their games, but the challenge isn't so much in facing tactical or operational brilliance as it is in facing a computer opponent that is extremely good at using the game system to achieve victory objectives, whether or not the ensuing deployments and operations on the map bear much resemblance to anything historical. This tendency is mitigated somewhat in Kharkov, with the excellent AO system, but it's a trait I think all their games possess to some degree.

Which simply goes to show that AI in wargames is hard. No matter how you do it you're going to disappoint someone. I agree you need to try, and try hard, but the goals need to be reasonable and fit in with the overall development plan.

[elmo3]

For those of you wondering about the AI, I just got my butt kicked by the Soviet AI set to Challenging in the Typhoon scenario.&nbsp; Of course I was just <cough> testing for bugs rather than playing to win <cough>.&nbsp; The AI did a good job reforming lines after I broke through.&nbsp; It also counterattacked effectively once the Blizzards hit.&nbsp; And Gary is still working to make the AI better.&nbsp; Just what I need!&nbsp;[X(]

[DrewBlack]

Now thats what i call excellant news we just need a competant Ai that gives us the challange anything better than the origianl is a step up!!![:D]

[Obsolete]

For those of you wondering about the AI, I just got my butt kicked by the Soviet AI set to Challenging in the Typhoon scenario.&nbsp;

Well, that does not exactly mean that the AI is any good, it just has to be a little better than the human idiot it is playing against [:D]

[elmo3]

Hey fella, I resemble that remark.

[Josh]

Well if it can beat a playtester it most surely will kick my &ss for sure, and it will probably kick 90% of the rest of us as well.

[Captain B]

Hopefully we will soon all be able to test out the AI! If it is half as good as what ya'll are saying it is, then we have a winner!

[Neal_MLC]

Great, just what I need, another game that can kick my a$$... When can we buy it?

[sven6345789]

can you switch sides midgame like in WITP?

If yes, since the AI is generally better on defense, you can still switch sides when the time has come for the soviet offense (sometime summer 1943).

[Helpless]

can you switch sides midgame like in WITP?

yes.

[critter]

ORIGINAL: Helpless

can you switch sides midgame like in WITP?

yes.

How does that work victory wise? It seems strange to lose the game as the German only to win as the Russian..?

[HMSWarspite]

The side that wins....&nbsp; wins

I like it in eastern front games (WIR) - play as German until it is obvious you cannot win (assuming you dont - house rules needed sometimes to avoid exploited AI), then switch, and see if you can win as Russian. . The AI generally completely juggles everything so that within a turn or two, you have little useful tactical intel, and strategically you probably dont know much more than the Allies did in RL . &nbsp;&nbsp;

[pompack]

ORIGINAL: critter

ORIGINAL: Helpless

can you switch sides midgame like in WITP?

yes.

How does that work victory wise? It seems strange to lose the game as the German only to win as the Russian..?

More like "as good as" win the game as the Germans then switch sides and win again as the Russians [:D]

[HHI]

I think the A/I will be as good as it can be, which is very poor. I just hope that it is not loaded with cheats, such as WitP/AE to offset the complete failure of the A/I. If we have a "historical" setting, then it should be historical.

May I suggest, however, that it would not be terribly difficult to have a satisfying game experience with HTH against yourself. So much is known about this campaign that, for example, if you are actually playing the Germans, you require that the Red Army implement the "Stalin Directive." You, as the Red Army, would implement a strategy of suicidal counterattack, leading to formation of the gigantic pockets that were actually formed. If you do not, an NKVD officer will show up at your door and shoot you. After the initial debacle, the Red Army would implement a more rational defense. Granted, you are playing with perfect intelligence, but it beats playing against a bungling A/I.

[Joel Billings]

At the normal level their are minimal cheats. The only cheating going on at that level is some movement cheats, but we are trying to be careful to make them small and reasonable. This is because the computer is not so smart at which units to move first in order to build a line, so it needs an ability to cheat move a little. Only at the harder play levels do other advantages kick in for the AI.