AGEOD Forums

Again, another complitely broken game AI wise.
Please advertise your games as PBEM only, because solo, it's hardly playable. 45k vs 55k battle results in a 75k+ casualties?
Same bugs, same retarded AI, same flaws that broke the WWI game.

You lost me as a customer. This level of quality, endearing at first, doesn't suit a 30+€ priced game.

I will just tell you one thing Ageod: get your shit together, and be straightforward toward your customers; be honest about the fact that they are, at first, your free-of-charge late beta testers.

This is a fruitful business practice, I know, but it doesn't pay off in the long term, and I hope you realize it.

Not a scam. There hasn't been a single PC based game that's come out that hasn't had multiple issues with it. Did they rush this to market? Yes, definitely.

BTW, the casualty issue has been found. Not verified (or denied) by the devs yet. If you would like to fix prior to patch, PM me.

Using the word 'scam' is harsh. Most issues can (and I'm sure will) be fixed in a couple of patches. The developers (a small team compared to the big companies) are always very responsive to the community and they do care about the quality of their games. AGEOD are one of the few developers remaining who are still creating in depth historical war games. I believe in their integrity and the issues highlighted by the community will be addressed as the game matures.

Hubu wrote:Again, another complitely broken game AI wise.
Please advertise your games as PBEM only, because solo, it's hardly playable. 45k vs 55k battle results in a 75k+ casualties?
Same bugs, same retarded AI, same flaws that broke the WWI game.

You lost me as a customer. This level of quality, endearing at first, doesn't suit a 30+€ priced game.

I will just tell you one thing Ageod: get your shit together, and be straightforward toward your customers; be honest about the fact that they are, at first, your free-of-charge late beta testers.

This is a fruitful business practice, I know, but it doesn't pay off in the long term, and I hope you realize it.

This is harsh. like the immense majority of strategy game studios, it is a little boutique doing the best it can and having to rush the product to meet deadlines. Give it a patch or two and it will be great. oh and by the way, as a mainly AI Player I can tell you this : However good the AI, it is always retarded in a computer game more complex than chess.... The important issue is finding the right balance between propping the AI a bit (some little cheats in force org for example) AND making the AI completely immune to the logics of the game.

People seem to forget that computer chess is as old as computers themselves. We talk about decades of research and millions of dollars. I would argue, that given that kind of an effort, any strategy game from civilization to wars of napoleon would have an AI that could regularly compete and beat any human opponent without any cheating. Chess naturally being a complex game, especially at the top level.

The last I heard AGEod doesn't have access to Deep Blue, nor a team of hyper-computing programmers, nor a budget of millions of dollars

Where did I say they had access or should have access to deep blue or a team of hyper-computing programmers or a budget of millions of dollars? I only made a comment about why chess AIs are so strong.

Chess Ais are not *that* strong. Play outside the book (Like the Mieses Opening.) It involves playing conservatively for a long-term advantage that the computer is unable to find in its game tree search. This will frequently involve selecting moves that are believed to be sub-optimal in order to exploit known weaknesses in the way computer players evaluate positions.

Now Go Ai, on the other hand, is pathetic if played on a full size board.

Chess has far, far fewer available moves than any war game does. Chess AI is dealing with a very basic system with hard rules.
The essence of war gaming is soft rules. In War in the West I can invade Denmark as the Allies in 1943. In WON I can as the Ottomans ally with Russia and Sweden and invade Britain.
In chess you can never use the 9th row.

No large company has made a good AI for strategy game yet; AGEOD is not a big company.
I wish these games could be marketed as PBEM only as that would give us a better game since they would not waste time making a semi competent AI, but I understand the financial reasons behind every computer strategy game saying they support solitaire play.

Personally, I only play these types of games against unsuspecting sheep, I mean honored opponents so i have never understood people playing these solitaire.

lycortas2 wrote:Personally, I only play these types of games against unsuspecting sheep, I mean honored opponents so i have never understood people playing these solitaire.

They work great playing solitaire if you give the AI all the available bonuses and also suspend your foreknowledge of what's to come as best as possible and don't do any min-maxing.

If you try to role-play as if you were the actual nation you're playing and don't go for a total conquest victory using min-maxing techniques, but simply play the game straight up, then the solo games are extremely fun and in fact quite challenging.

The AI in AGEO games is actually very good for a wargame. It's not perfect, and sure there are bugs and flaws especially early in release but the ai doesn't need any apologies, it's actually quite impressive.

As to the op, to call this game a a scam is basically a lie by the poster. Can't say any more than that.

It's a great game. It'll be even better after a few patches, but that's only to be expected. It has some issues sure, like all games, but it's eminently playable as is and really quite a blast to play.

Regularly beating the best humans sounds pretty strong to me. Playing out of book or using anti computer positions are no quarantee of a victory against a modern chess AI these days. If that would be the case, every top player would still be able to defeat them.

speakeasy wrote:Regularly beating the best humans sounds pretty strong to me. Playing out of book or using anti computer positions are no quarantee of a victory against a modern chess AI these days. If that would be the case, every top player would still be able to defeat them.

Yes, but the point is that in chess there are only a limited amount of possible moves (yes, an incredibly great amount 10 to the 40th power perhaps, but still not actually infinite) and a limited playing field.

In any wargame, the amount of potential moves is truly infinite and the gameboard, if you will, if thousands of times larger than a chess board and the amount of pieces, or units is massively greater as well.

There's actually no comparison between the ai for a good wargame and a chess ai, the wargame ai is infinitely more complex and more difficult to program. Chess ai's simply use brute force because there are so few options, relatively speaking, that brute force is the best way to go.

Wargames can't use brute force, it's not possible. They have to actually program elegant ai's capable of dealing with an infinite variety of possibilities. That's an exponentially greater challenge than chess ai's face.

lycortas2 wrote:Chess has far, far fewer available moves than any war game does. Chess AI is dealing with a very basic system with hard rules.

Doesn't every game have hard rules? And no matter how great a game war in the west is it still operates on a basic system. I don't see the difference. Chess if full of options starting from the first move. Sure there are only a handful of moves that make sense if you want to win, but the same applies to war in the west for instance. And I'm just talking about simple moves here. When we get to strategy you must make desicions like when to attack, where and how. When to defend and how. When to trade and when to sacrifice. You must formulate a plan and then execute that plan.

Brute force does nothing if the AI does not understand how to play good chess. Analyzing positions is the key. Chess AIs these days can play very strategically which was the last thing that still kept us ahead of them. They are not just tactical beasts anymore.

And like I said in my first post. The reason we have chess AIs that can beat the best of us has very little to do with how "simple" chess is and more to do with the fact that we humans have been at it for over 50 years and that's pretty amazing.

I'd rather have a chance at having a great strategic game than nothing at all. How many developers out there bother with games that require more than two fingers to play?

Yes, ai is imperfect. So join a PBEM.

Do you realize how difficult it is to build an AI.

1) Theater Awareness (moving troops to defend, attack, objectives, politics)
2) Operational Awareness (moving troops to achieve goals, either attack or defense)
3) Economic Awareness (obvious)
4) Self Awareness (coming soon)

Then building the code to make all that happen. And then with self awareness, the AI launches nukes at all the factions and comes oh so close to winning but doesn't.

vicberg wrote:4) Self Awareness (coming soon)

Then building the code to make all that happen. And then with self awareness, the AI launches nukes at all the factions and comes oh so close to winning but doesn't.

Yeh, that will be the next ageod game "Terminator Judgement Day" ... the day Athena (ageod AI) becomes self aware...

vicberg wrote:Do you realize how difficult it is to build an AI.

How what I am saying is implying that I think building an AI is easy? It took over 50 years to build an AI that can beat human in chess. That does not sound easy to me.

Regarding AI in AGEOD games, I've noticed the main issue Athena has is with building troops and organizing corps/armies. I've found the best AI games in AJE, BoA2, and Rise of Prussia, while the worst AI experince has been Revolution Under Siege (I love the game, haven't played Gold yet, but Soviet's just couldn't effectively build a military) and Espana (Nationalists AI not aggressive enough.) Thirty Years War was rough, but I haven't played it with the most recent patch so hopefully the Swedish and French now come in. All of that said I love AGEOD games and the amount of support for their games is incredible.

Personal opinion, I don't mind the "unrealistically high" casualties. Remember that players usually play more aggressively than real life. I once had a Rise of Prussia game (As Hapsburg/Russia/France/Sweden) against the AI in which the ledger reported casualties of over a million, and while not entirely realistic, it was one of the most exciting games I've ever played against the AI. Prussia fell, but France, HRE, and Sweden's military's were ruined too.

Maybe it is just because AGEOD was the first operational scale strategy game I played, but I really love the system and design. I like the game engine too, for the most part, even if it's old. I enjoy Paradox games a lot too, but those are more about building an empire rather than simulating a particular war. Commander Great War's is a lot of fun, but I prefer the AGEOD engine for simulating a war (more flavor/detail). If we're looking at large strategy games, AGEOD's engine compares well to Forge of Freedom, Storm over the Pacific, or Commander Great War.

My personal hope for AGEOD is that they try a smaller scale game. I think American Revolution (BoA3) or the English Civil war would be perfect. Maybe go to the Middle Ages for the War of the Roses/Hundreds Year War. I think that would let Athena remind players what she can do!

Personally, the reason I keep buying AGEOD games are:

1) I find the engine/system fun enough to play around with
2) They support their games for a very long time

There are, of course, secondary reason, like the art is very good and I only buy games that touch on historic periods I enjoy

I`m perfectly aware that this game (or any other AGEOD game) may not be fully playable at first, specially when it comes to balance and AI, I don´t like it, but it`s the state of the "industry" when it comes to strategic games, but I do believe they make an honest effort polishing their games for a long time and engaging with their community

The other option is quitting this hobby for good, there is no market for this games to allow for big budgets, even Paradox, which is probably the most successful company in this genre depends on the community for polish and balance, the same goes for other companies like, for example, Battlefront with their very particular system

If I could pick one thing, I would like for the AI to build and manage it`s armies and corps rationally, maybe just for flavor, but many times it`s formations make no sense to a human, maybe they make sense mathematically, I don´t know, but it brakes immersion

All said, it`s only a "scam" if you are totally new to the issue, and even then a "scam" is too a strong a word, the game installs and runs

The tragedy in computers games is that for the most part each studio reinvent the wheel when it comes to AI. It's a pity that companies like Microsoft or the like have not done the equivalent of DirectX but for AI. But graphics sell, AI, not so much. So we do what we can with our feeble minds.

Hubu wrote:Again, another complitely broken game AI wise.
Please advertise your games as PBEM only, because solo, it's hardly playable. 45k vs 55k battle results in a 75k+ casualties?
Same bugs, same retarded AI, same flaws that broke the WWI game.

You lost me as a customer. This level of quality, endearing at first, doesn't suit a 30+€ priced game.

I will just tell you one thing Ageod: get your shit together, and be straightforward toward your customers; be honest about the fact that they are, at first, your free-of-charge late beta testers.

This is a fruitful business practice, I know, but it doesn't pay off in the long term, and I hope you realize it.

Harsh. Very harsh. That a game like this is very roug hat the startis nothing new. Games like this, Slitherine or Paradox tend to be very rough at the start. They're complex to program and developped by small studio's for a very select audience.
I'm probably going to give the grand campaigns a try after a patch or 2. Not yet for now and I expected this in fact.
Meanwhile the grognards here are already laying a good foundation of the future. As is typical.

Aurelin wrote:Chess Ais are not *that* strong. Play outside the book (Like the Mieses Opening.) It involves playing conservatively for a long-term advantage that the computer is unable to find in its game tree search. This will frequently involve selecting moves that are believed to be sub-optimal in order to exploit known weaknesses in the way computer players evaluate positions.

Now Go Ai, on the other hand, is pathetic if played on a full size board.

I will disagree with your comment about Chess AI's "are not that strong..." As a semi-professional and Expert chess player (USCF rating 2079), I am quite aware of the various commercial chess programs available (i.e Fritz, Gull, Houdini etc..) and myself have been crushed by them so routinely, that I know longer play against them even to train. Komodo 9 for example could EASILY crush the human world champion today. Kasparov thought Deep Blue was "cheating" (getting human aid, and it was not). He would be shocked and humiliated today by the strength of Stockfish or Komodo (programs that make Deep Blue look like an amateur). Do some research on these programs.

Scottus wrote:I will disagree with your comment about Chess AI's "are not that strong..." As a semi-professional and Expert chess player (USCF rating 2079), I am quite aware of the various commercial chess programs available (i.e Fritz, Gull, Houdini etc..) and myself have been crushed by them so routinely, that I know longer play against them even to train. Komodo 9 for example could EASILY crush the human world champion today. Kasparov thought Deep Blue was "cheating" (getting human aid, and it was not). He would be shocked and humiliated today by the strength of Stockfish or Komodo (programs that make Deep Blue look like an amateur). Do some research on these programs.

I play Fritz 14, regularly, not lost yet. Going to get the 4-core version to see what it can do.

Aurelin wrote:I play Fritz 14, regularly, not lost yet. Going to get the 4-core version to see what it can do.

Under what time conditions? Wow, you should be sending fear throughout the chess world considering even a 10 year old prototype version of Fritz drew Kasparov and even beat Kramnik. Of course if you are playing on a low end machine you might stand a chance and if so, play some speed chess with it for some humility lessons.

Even low end machines today are probably stronger than the machines Kasparov was playing against 12 years ago.

Fritz, despite being on the weaker end of Chess software today (although 15 looks strong) still will give any non master a beating.

If you are pounding Fritz like you say you are (I doubt it), pick up a copy of Komodo 9.

Look this up: "Komodo vs Humanity: Nakamura Will Give It A Go" and read about the thrashing Komodo gave human GM's while spotting them material. Hikaru Nakamura is the #1 US player and #4 in the world and will be playing the Komodo program with the human getting material odds. Recently, he lost two and drew two against Stockfish, a weaker program which Komodo destroyed.

My entire point is that you claim "Chess A.I's are not strong"... as a Candidate Master, considering my own personal experience with Chess A.I. , and the spanking they have been giving Masters and even GM's (not to mention the worlds best players as well as the champion), for over 12 years, I would say you are quite wrong. GM's don't stand a chance today without being spotted one or two pawns. I call that beyond strong Chess A.I. In fact, Humans are really no longer the world champion in Chess... machines are.

Yeah. Chess engines are at 3000+ ELO now. Carlson is 2860 or so. You are not regularly beating Fritz 14 unless your in sparring mode and its handicapped in some way. The above poster says I seriously doubt. I'd go further and say impossible.

Wargames are like chess with a thousand pieces. I always liked that definition.

I am happy with WON now the combat % can be changed.

Anybody thought of introducing odds into chess e.g. queen vs pawn 90%, pawn versus queen 10%?

Taillebois wrote:Wargames are like chess with a thousand pieces. I always liked that definition.

I am happy with WON now the combat % can be changed.

Anybody thought of introducing odds into chess e.g. queen vs pawn 90%, pawn versus queen 10%?

..with fog of war, so you only see within 2 squares of your pieces...
...with random combat results, based on characteristics of the pieces....
....with terrain in each square affecting movement and combat....
....with logistics, where pieces out of supply starve.....

Crimguy wrote:Yeah. Chess engines are at 3000+ ELO now. Carlson is 2860 or so. You are not regularly beating Fritz 14 unless your in sparring mode and its handicapped in some way. The above poster says I seriously doubt. I'd go further and say impossible.

Agreed... Been playing for over 40 years. There is no way anyone is beating any version of Fritz and claim they never lost a game unless the are Bobby Fischer incarnate... and even then not on any serious game mode or on some POS rig.

The point is the guy said that "Chess A.I's are not that strong. Judging by the CRUSHING string of victories against the worlds best GM's over the past 12 years, I would say he's wrong. I am sure the GM's would agree. Today, programs like Stockfish and Komodo are regularly beating Masters and Grand Masters at two pawns or more odds! The GM's stand NO chance with even odds.

My guess is it's all mathematics of a sort not my best subject. Chess Maths is probably so much simpler than the most basic wargame. Also Chess has been around for so long that the play testing has all been done how many times in a Chess game does anyone now do something completely different. No expert however just think that by the time that get a A.I done for any of the larger Wargames that can match the unpredictable human nature is the time we will all be holidaying on Mars

typhoon wrote:My guess is it's all mathematics of a sort not my best subject. Chess Maths is probably so much simpler than the most basic wargame. Also Chess has been around for so long that the play testing has all been done how many times in a Chess game does anyone now do something completely different. No expert however just think that by the time that get a A.I done for any of the larger Wargames that can match the unpredictable human nature is the time we will all be holidaying on Mars

Wargame AI is beyond insane to program one even of a mediocre level (vis a vis Chess A.I.). Way more complicated. However, there are new ideas in A.I. design that we will be seeing in the next few years that will be taking advantage of some of the new CPU design ideas etc.

No one is claiming Chess A.I. and wargame A.I. are similar in complexity. Someone above made the comment that Chess A.I. is "not strong". He is wrong. Also, note that even though chess as compared to typical wargames is seemingly simplistic, there are more possible chess moves then there are atoms in the known universe!