When?

[Joseignacio]

http://www.priceminister.es/offer/buy/4 ... Libro.html

http://www.priceminister.es/offer/buy/4 ... Libro.html

http://www.priceminister.es/offer/buy/4 ... Libro.html

but there are some much older:

http://www.clausewitz.com/readings/Principles/

http://www.icollector.com/Military-Tact ... -_i8702773

http://en.wikipedia.org/wiki/The_Book_of_Five_Rings

http://en.wikipedia.org/wiki/The_Art_of_War

are the first tactics and strategy books that come to my head. Maybe you can get some hints...[8|]

[Sarge]

Well some of us actually have practical real experience in the field, some play video games and read Art Of War and think they have some insight.

My advice is move on straw man you'll get your AI when its done try to relax its just a game ...........[>:]

[DCWhitworth]

ORIGINAL: Sarge
Well some of us actually have practical real experience in the field, some play video games and read Art Of War and think they have some insight.

My advice is move on straw man you'll get your AI when its done try to relax its just a game ...........[>:]

Granted there is no substitute for practical experience but there are other ways to acquire knowledge on the subject.

[Sarge]

ORIGINAL: DCWhitworth
Granted there is no substitute for practical experience but there are other ways to acquire knowledge on the subject.

Don’t let straw man above confuse my original point....[;)]

ORIGINAL: Sarge

lol

For thousands of years military collages from every conceivable civilization armed with immeasurable accumulative knowledge have utterly failed in producing a simplistic AI temp plate : if this happens do this!

So …………what’s the hold up ?

[DCWhitworth]

ORIGINAL: Sarge
lol

For thousands of years military collages from every conceivable civilization armed with immeasurable accumulative knowledge have utterly failed in producing a simplistic AI temp plate : if this happens do this!

So …………what’s the hold up ?

I think the issue is that real life is an immeasurably boundless and complex thing wheras computer programs have relatively limited horizons and certainly can't *think* outside the box. A further issue is that computers do not think like humans so trying to get them to behave like them is doubly hard.

But actually you'd hit similar issues if you gave a list of written instructions/advice to a human who applied them literally and without initiative.

[Joseignacio]

First, I have been a soldier, reaching the graduation of Cabo Primero, don't know the translation but it would be like "first class corporal", the higher level in our army under sargent, however, I have not been professional, that I must admit.

Anyway, that has nothing to do with what we are writing about. I think that most of the generals who were the High Commands in the II WW had probably never been in a real front.

This is a simulation, and we are armchair generals, maybe any of us could be as good as Montgomery, Auchinleck or Rommel, but most wont. If the books were good to learn tactics and strategy for romans (Cursus Honorum included them if I am not wrong) ,

http://en.wikipedia.org/wiki/Cursus_honorum
http://en.wikipedia.org/wiki/Tribune#Ro ... y_officers

cartaginese and chinese officers among others, I think they must be valuabe no matter that, as Moltke said "No plan of operations extends with certainty beyond the first encounter with the enemy's main strength (no plan survives contact with the enemy).",

http://en.wikipedia.org/wiki/Helmuth_vo ... _the_Elder

the course of action has to be decided before, and the better the planification, the possible alternatives coverture and setting of reserves and retreat coverture measures, the better the result will be if something results wrong.

Surely you don-t believe, Alexander, Hannibal, Julius Caesar, ..., just imagined their tactics or lost enough legions till they learnt that it was not good to cross a river and leave it behind in the presence of the enemy. They were taught, with senior officers or their writings in books.

Hey guys, Steve doesn't like to mix everything up in this thread, I will copy paste all this in a new one and we can leave this for the "When?" matter, Ok?

[mavraamides]

ORIGINAL: Sarge


For thousands of years military collages from every conceivable civilization armed with immeasurable accumulative knowledge have utterly failed in producing a simplistic AI temp plate : if this happens do this!

So …………what’s the hold up ?

The problem with that sort of simple template is you would need too many rules to be practical. Even for a simple game like chess you would be talking about appx 10^40 rules for a 40 move game.

In the case of WIF, you could multiply that by at least 1,000,000. That's why some sort of hierarchical generalization is needed. Or a learning engine that can observe / play 1,000's of games and infer generalized rules.

Humans have an innate ability to recognize patterns of play or positions and generalize rules that can then be applied to them. Taking chess again for example, you may have:

1) If my opponent neglected his king's safety, launch an attack against the king side.
2) If I am ahead in pawns, trade off pieces and head to an end game.
3) If under attack on the wing, launch a counter attack in the center
...
etc.

We can take these generalized rules, recognize the patterns as they evolve from a game, and use them to narrow our move choices. This is the idea of heuristics (guided search) first introduced by the great genius Alan Turing. Perhaps if he hadn't taken his own life at the tender age of 42, AI would be years ahead of where it is now.

People who have never written an AI tend to vastly underestimate just how much thinking is going on behind the scenes by a human mind in making even the simplest tactical decision. Nothing will make you appreciate the power of your mind like trying to write a program that mimics it!

[winky51]

People who have never written an AI tend to vastly underestimate just how much thinking is going on behind the scenes by a human mind in making even the simplest tactical decision. Nothing will make you appreciate the power of your mind like trying to write a program that mimics it!

I have written AI and yes its way more complicated than people think.

Computers have linear memory in directions forward and back.

humans have dynamic memory where 1 cell can attach to numerous cells forming synapses. Way more complicated.

I made my own AI space game a long time ago (pascal) and it works. I also write AI for my scenarios for SC2. Its no easy task.

[Sarge]

ORIGINAL: Joseignacio
Hey guys, Steve doesn't like to mix everything up in this thread, I will copy paste all this in a new one and we can leave this for the "When?" matter, Ok?

Agreed ,
I’m certain if Matrix is stating the release will have AI then I have no reason to be suspect .


I don’t venture down to the WIF forum often so I my be guilty of missing Matrix statements on the contrary ?.

[Kham]

No statements to the contrary.

It's just gonna take time and its hard on the patience - especially if you apply a realistic expectation of how good it can become

[Triboga]

But, finally, with or without high AI, have any new release date?

[wworld7]

ORIGINAL: Triboga

But, finally, with or without high AI, have any new release date?

Not yet.

Perhaps with the Dec 1, 2009 update we will hear something.

[micheljq]

From my personal point of view the dev should never give a release date unless the product is really finished and ready to go to the market. In the industry they give vague answers like Q2. 2010 for example, this way they are not pressured to release an unfinished product.

[Joseignacio]

ORIGINAL: micheljq

From my personal point of view the dev should never give a release date unless the product is really finished and ready to go to the market. In the industry they give vague answers like Q2. 2010 for example, this way they are not pressured to release an unfinished product.

I think he made the dates public to have an extra push to meet them, since all the fans would know them.

And I think now he doesn't (by now) because the time to dedicate to the bugs is unpredictable.

[micheljq]

ORIGINAL: Joseignacio

I think he made the dates public to have an extra push to meet them, since all the fans would know them.

And I think now he doesn't (by now) because the time to dedicate to the bugs is unpredictable.

I agree with that, but I mean there is unpredictable bugs happening in the development of all PC games anyway. So they never know when the product will be finished until it is. It is like that in all the industry, MWiF is not an exception. [:)]

[SamuraiProgrmmr]

ORIGINAL: micheljq

From my personal point of view the dev should never give a release date unless the product is really finished and ready to go to the market. In the industry they give vague answers like Q2. 2010 for example, this way they are not pressured to release an unfinished product.

I would just like to say that we (as a group) have been granted more transparency into this project than any other computer game project in history (correct me if I am wrong).

The benefits of that transparency are of immense value. Experienced players of the game have had the chance to influence decisions. Inexperienced players of the game have had a chance to point out the pieces that needed more explaination or information on screens. This has, IMO, made this a much better product.

So what if we were given a PROJECTED date (or dates) that were not met.

What is more disappointing... a missed target date or an inferior product?

To Reiterate : Matrix Games NEVER said "This is going gold on xx/xx/xxxx".

We were given the privilege of seeing behind the scenes as this was built.

Please don't make them regret that. I wish more companies would adopt this transparency.

[micheljq]

Transparency does not mean they are obliged to give an early exact release date, anyway they will do what they feel is better to do.

That's being too much of a nice guy, I personally think.

[Caquineur]

ORIGINAL: SamuraiProgrammer

ORIGINAL: micheljq
From my personal point of view the dev should never give a release date unless the product is really finished and ready to go to the market. In the industry they give vague answers like Q2. 2010 for example, this way they are not pressured to release an unfinished product.

I would just like to say that we (as a group) have been granted more transparency into this project than any other computer game project in history (correct me if I am wrong).

The benefits of that transparency are of immense value. Experienced players of the game have had the chance to influence decisions. Inexperienced players of the game have had a chance to point out the pieces that needed more explaination or information on screens. This has, IMO, made this a much better product.

So what if we were given a PROJECTED date (or dates) that were not met.

What is more disappointing... a missed target date or an inferior product?

To Reiterate : Matrix Games NEVER said "This is going gold on xx/xx/xxxx".

We were given the privilege of seeing behind the scenes as this was built.

Please don't make them regret that. I wish more companies would adopt this transparency.

SamuraiProgrammer, I agree 99,99% with what you just said.
The 0,01% missing is about the part I put in bold. In fact, in at least one monthly report, Steve said :

ORIGINAL: Shannon V. OKeets
April 1, 2009 Status Report for Matrix Games’ MWIF Forum

Accomplishments of March

Project Management
Firm release date for MWIF product 1 is July 27th, 2009.

But it's, IMO, a minor point, as I totally agree with the fact that a superior product is better than a "not-missed" target date, at least in the case of MWiF.

Alain

[Triboga]

To be honest I prefer have a released date, but doesn’t care if finally didn’t arrive to it. Is just another milestone to know where the game’s development is. Personally, I appreciate this effort, and the fact of give a date and try to reach it.

Playing board game is easy to notice how many variables you have to be aware, and (as far as I know I’m not developer[:'(]) can’t imagine how many of them must be in a “decision tree” of this kind of game. But in my opinion (well is a fouled opinion of a person who wants to boost the game release [:D]), even with the best AI, this kind of game is not the kind of you enjoy playing against a evil machine, it’s a huge approach of WWII and –always for me- is an approach which need the human factor, we’re not talking about a battle or even a WWII game designed for PC, thought it was not a game developed for PC but for board game

[bo]

ORIGINAL: micheljq

ORIGINAL: Joseignacio

I think he made the dates public to have an extra push to meet them, since all the fans would know them.

And I think now he doesn't (by now) because the time to dedicate to the bugs is unpredictable.

Mich I have always respected yours and Joses opinions, always, but this bug thing annoys me to no end, if the code is written right in the first place why is there so many so called bugs, I trully dont get it. I have not posted because I will end up saying the wrong thing but I do read the posts every 2 weeks or so.

BO
I agree with that, but I mean there is unpredictable bugs happening in the development of all PC games anyway. So they never know when the product will be finished until it is. It is like that in all the industry, MWiF is not an exception. [:)]