Slitherine

quote:

ORIGINAL: sabre1

Bo,

For me personally, most games and the AI can kick my proverbial you know what. For me what makes a great game is one that is intuitive and quickly picked up with the UI easy to remember, if I can't play every single day. Oh and I am a sucker for eye candy, not super fancy, but pretty dang pretty. Kharkov by SSG is as far down in eye candy as I personally want to go.

One of the reasons I gave up on flight sims like Falcon was that I couldn't remeber all the key commands. I gave up on COTA because the tutorial took just to long for "me" to get through and then remember what the heck was going on next time I sat down to continue.

Grigsby games have that issue also, but I appreciate his talent and the masterpieces he creates.

I LOVED PG. Good strategy wargame, attractive graphics, and just plain fun. CEAW is another one I liked, but for some reason just doesn't have that I want to come back to it months later.

Guys like you who can lick the AI consistently, I am always in awe of.

I guess I'm saying these are OK, not great in the gaming world, because only every once in a blue moon does a game come out that hooks me to the point of I can't wait to get back and play it. Maybe I'm just getting old and jaded (56 ).

I just bought two more games by small developers, and I'm hoping they are "easy" and enjoyable. Life is tough enough without dealing with unapproachable games.

All IMHO.

quote:

ORIGINAL: Erik Rutins

Bo,

Please do not make political analogies - the rest of your post was fine, but we try to keep politics (and religion) out of our threads here.

Regards,

- Erik

quote:

ORIGINAL: Arctic Blast

Well, I guess the first question is...what makes for a good AI? Should it be aggressive? Okay, but then what about when it should be playing a defensive game. Okay, so it should be defensive. Oh...wait...no, that means it will just sit there and not take advantage of opportunities when they present themselves. Should it be more on the devious side, trying to use alliances and trade agreements to gain an advantage on you that way? Fine, but what about when it gets in to a situation where it absolutely must fight.

There are a lot of variables, and the bigger and more complicated the game, the worse the situation becomes simply because there are more variables. To truly program a REALLY good AI takes a LOT of time...you basically need a programmer or two working on absolutely nothing else during the entire development time of the game. And if you want a truly multiple strategy capable AI...well, how long are you prepared to wait? And how long can a small software company with limited resources AFFORD to wait for it to be done, paying the salaries out to their employees the entire time? What if they have the publisher they signed on with breathing down their necks (which happens a lot)? And as for the big developers who CAN afford to spend their time working on an AI, why would they? How many games sell based on an impressive AI vs. top end graphics and performance? For every person who looks for the former I absolutely guarantee you that there are at least 100 others looking only at the latter. So why spend the time and resources on developing for the few when you can spend that same amount of time and money developing for the many?

quote:

Now maybe the programmers of TOW do not feel that N. Africa is not that important to the overall scope of the game and probably it is not but its just another thing that annoys me.

quote:

ORIGINAL: Arctic Blast

Well, I guess the first question is...what makes for a good AI? Should it be aggressive? Okay, but then what about when it should be playing a defensive game. Okay, so it should be defensive. Oh...wait...no, that means it will just sit there and not take advantage of opportunities when they present themselves. Should it be more on the devious side, trying to use alliances and trade agreements to gain an advantage on you that way? Fine, but what about when it gets in to a situation where it absolutely must fight.

There are a lot of variables, and the bigger and more complicated the game, the worse the situation becomes simply because there are more variables. To truly program a REALLY good AI takes a LOT of time...you basically need a programmer or two working on absolutely nothing else during the entire development time of the game. And if you want a truly multiple strategy capable AI...well, how long are you prepared to wait? And how long can a small software company with limited resources AFFORD to wait for it to be done, paying the salaries out to their employees the entire time? What if they have the publisher they signed on with breathing down their necks (which happens a lot)? And as for the big developers who CAN afford to spend their time working on an AI, why would they? How many games sell based on an impressive AI vs. top end graphics and performance? For every person who looks for the former I absolutely guarantee you that there are at least 100 others looking only at the latter. So why spend the time and resources on developing for the few when you can spend that same amount of time and money developing for the many?

quote:

ORIGINAL: gwgardner

Are you going pedal to the metal when you play? I mean, strike directly from taking Warsaw, to attacking France in October '39? Crushing Yugo and Greece in Spring '40? Attacking the USSR a year early?

I think some players are crushing the AI so easily because they rush the game.

Look at my solo AAR, trying to follow an historical timeline. I haven't found it so overwhelmingly easy so far.

In North Africa, with decent leaders and better morale, perhaps the Italians could have overrun the British, even without German help. You're supplying the 'decent' leader. So you have an advantage over the historical Italians right there. What level of fog of war are you using?

I'm just saying that if you want a better AI, then do like a famous radio announcer does, and tie half your brain behind your back, just to make it fair. Give the AI the advantage of more production. Don't launch attacks in the middle of the mud season and the winter. Don't let your human controlled countries give all their PPs to their allies, etc.

Lots of ways to get more bang for the buck in this game.

As for the difficulty programming an AI in a war game: it's all a matter of resources. Put a team of programmers on the ToW AI for a year, and you'll get your super AI. But what development company has the resources to do that? Maybe Blizzard? [later] ditto what Artic Blast said!

quote:

ORIGINAL: Bleck

Arctic Blast, thanks for writing down all those aspects of developing strategic game with AI. :)

quote:

Now maybe the programmers of TOW do not feel that N. Africa is not that important to the overall scope of the game and probably it is not but its just another thing that annoys me.

Programmers have to write AI code that will work everywhere, on any map, not only historical ones. We have to predict every possible combination of many variables (firs define those variables), then designers have to define (and this is the REAL hard work, programmers are only implementing those things) rules, conditions, modificators, possible responses, calculations etc. for each combination, then someone else have to set all modificators values, then test and balance AI. And as Arctic Blast wrote - there is no way to test all possible scenarios.