AI improvements? (Full Version)

All Forums >> [Current Games From Matrix.] >> [Age of Muskets] >> Birth of America 2: Wars in America



Message

JastaV -> AI improvements? (3/11/2009 12:16:50 PM)


The concept of game AI is related to a smart engine acting in a logic and flexible way to carry on a working, effective, winning strategy. Objectives achievement and consideration for enemy positions and moves are part of the strategy: the AI should be able to seek conquest of key objectives controlled by the opponent, defending at same time objects already controlled. A good planning ability is only half of the strategic duty: flexibility when executing plans is the second element.
When the AI is strictly working by a rigid scheme application, without considering the overall strategic pictures and the enemy positions we have nothing smart and we cannot even use the term AI.

PhilThib and AGEod took chances to exalt at many times their games AI, reporting it all as a highly flexible, complex, smart planning actor.
Then, the improvement of the AI in WIA was often remarked and it was the main reason to justify the publishing of a game that has so few new issues to offer as regard game features, graphic improvements, main game scenarios, when compared to old BoA.

Playing WIA after last patches publication, users had feeling AI was really improved.
IT'S FALSE!
THERE'S NO REAL AI IMPROVEMENT!

Recently, I started examining WIA files in the perspective of some modding. It was looking at scenario files I notice AGEod is now using scripted event-commands to unflexibly dictate AI decisions.
Notice, we are dealing with scenario specific setup and event files: it has nothing to see with the game engine where AI is encoded.
I catch out a mess of scripted event-commands that dictate AI posture, AI aggressiveness and enlist targets to be achived by the AI: "AI.SetAggro"; "AI.ChgLocalInterest";....
By use of such scripts AI is ordered to launch offensives in one or more regions at a scheduled date, but without any regard to the strategic overall picture.
So we have not AI, we have not Athena’s planning: AI opponent is just executing a rigidly ordered move.

Going after such scripts execution, Athena will launch that same offensive, at that same date any time the scenario is played, without flexibility and misreading player moves, opponent armies position and the strategic picture.
NO FLEXIBILITY, NO LOGIC, NO AI: just a scheme to apply in a flat, dummy way at a scheduled time!

It's remarkable these kind of commands was not used with NCP: they have been introduced in WIA as AGEod way to improve AI.
That explains why NCP received no patches and improvements out of many past months: it was not possible to "smart up" NCP AI by importing WIA.exe in the game, because there's no real improvements within WIA.exe and the related AI.
To "smart up" NCP AI AGEod programmers should have to encode tons of rigid, scenario specific scripted event-commands: they probably should have to work for months after it!

Now, but a great volume of work spent after WIA scripted event-commands we have no real AI improvements, on the contrary new troubles and AI weaknesses have been introduced.
For example, AI will keep moving against a region with great aggressiveness attacking it even with very unfavourable strength troops ratios: a suicide for the AI army.
Or AI will keep moving forces to achieved the scheduled objective even in winter season, being quickly decimated by harsh weather conditions.
Or even worst, AI main army will be launched after a scheduled objective at the wrong time, when player army are close to the AI capital that will soon fall undefended.
The examples I enlist here have all been experienced when playing WIA and mostly reported by public posts or addressed to the AGEod support.
I have no doubt AGEod support is one of the best I never meet out of many years as regard courtesy and quickeness in responce to player feedbacks: it does not mean they can fix all troubles and soon!



In conclusion,
we already have Objectives and Strategic Towns, with National Moral and Victory Point scalable values to assist AI as player in strategic choices.
The fact Athena is then to be "supported" in its decision by rigid commands points out that AI is really weak and poor.
Then scripted event-commands are game flexibility and variability killers, and their use will kill scenarios longevity too.
Of course, developers' efforts to improve AI at all cost is to be admired expecially looking to the tons of new commads to be edited per scenario. In perspective a proficient and smart use of AI dedicated scripted event-commands can make the difference expecialy when they should be used without killing flexibility and modulated by introdution of "conditions".
Unfortunatemy tera-tons of codes are to be added to go after such a solution: so, is the strategy worth? Is there no chance to actually improve and update the engine-related AI?




JastaV -> A possible solution? (3/11/2009 12:20:40 PM)


Indeed rigid, scripted event-commands can be used in a flexible way!
Flexibility and a smart use of them can be introduced by accurate use of events conditions: “conditions” are already commonly used within AGeod games command strings.

An example.
1. I edit 3 scripted event-commands scheduling offensives after 3 different objectives.
a. objective Boston
b. objective Philadelphia
c. objective Richmond

The three offensives will be scheduled to be active at the same date, any with a 33% activation chance and they mutually exclude.
With that, just 1 out of the 3 available will be randomly launched: this way I'll save variability any time the scenario is played.
Then, using a time span, (max and min. dates) to activate the script I'll introduce variability as regard time.

2. Then I can work to make the activation choice smart, by introducing specific activation conditions.

a. A 3/2 troops superiority in the area could be a condition for Athena launching its scheduled offensive.
b. Athena will not move on Boston and Philadelphia without being in control of New York and Albany: that way AI will strike on offensive with flanks and rear secured.
c. Athena will not take offensive posture when enemy armies are close to the AI side capital-town: defending the capital, or any own side controlled major city is a priority.
and so on......

Now, looking at that we could do by use of smart conditions we can built up alternative, “perfect” operative plans, (history teaches there’s no perfect plan!) to be picked up by Athena at the right time.
So, Athena will launch its offensive against one of many objectives by a random choice, saving game variability and scenario playing-longevity. Then the choice will be restricted by preliminary achievement of local superiority, (a condition commonly used in military offensive operations).
Of course tons of scripts are to be edited... a hard duty for game designers.
Producing "easy" scripted event-commands, without conditions and multiple alternatives is most simple and direct: but it produces all the side-effect we know.

In conclusion, activation conditions and elaborate plans need a mess of dedicated command-strings without really wiping out the risk of failure. In fact, they produce anyway rigid operative schemes that out of many conditions used will not grant a smart execution. There always will be some condition or event we cannot foresee when editing AI plans by scripted event-commands…..
Anyway that is a minor trouble: even historical and modern military plans cannot foresee all possible incidents; then, “incidents” altering balance of power and making things uncertain are the actor behind interest for military history!



GShock -> RE: A possible solution? (3/11/2009 4:29:35 PM)

Might be the reason why i quit playing aacw an year ago and i could never even complete a wia campaign...

good to know why it is so bad but i frown at the fact theres no easy way to change any of this.

...now if they fixed things a bit at least we could play a pbem...



Blueprint -> RE: A possible solution? (3/11/2009 11:09:39 PM)

These new commands are implemented in AACW since July 2008 and in use in my AACW mod since August 2008. NCP being posterior to AACW these commands could be implemented in NCP.

These commands aren't what JastaV names scripted events. Hes talking before knowing. For my own, I'm with Lodilefty, the only one to have used since some months these new commands to model AI. And by using conditions too since August 2008 to fire these commands.I'm happy to se JastaV discovering at last what is from a long time discussed in the Ageeod forums. There's even a third AI command implemented from the start.

AI can be "guided" too by playing wih objectives, fixing of forces. Things I used in my AACW mod since 2007.

I did that by the trial and error process ( yes a little different of the "Ignre and conclude" seen in the precedent posts). What I noticed is AI is reacting differently when you replay the same turns. there are patterns of course but variations too. And variations are sometimes very large. I've played hundereds of games. AI is sometimes doing things very different. Why? In part because the AI is built to choose different moves on a given situation, in part because the AI is reacting to enemy moves and when i tested different styles of play I've got different AI reactions. so long for the scripting theory endorsed by JastaV.

These commands are just helping AI to concentrate on some objectives and play more or less offensively in a context based for its assesments on several factors whose the more important are the enemy threaths. Adding these commands is just adding to the tactical AI a better assesment of the strategic situationS. This could only be done until now by tweaking objectives values of regions, a trick used by JastaV in his now unpublished NCP mod.

Now is the AGE AI perfect? No. Is the AGE AI predictable?Partly yes. Is it another wargame having an AI as smart as described by JastaV? I don't think? Should JastaV and Gshock learn before to write? definitly yes.

Time spent by JastaV to expose his uasserted theory could have been used in a more convenient way, like playing "Risik" (Risk) : http://www.matrixgames.com/forums/tm.asp?m=1997686






Blueprint -> RE: AI improvements? (3/11/2009 11:43:48 PM)

quote:

ORIGINAL: JastaV




Playing WIA after last patches publication, users had feeling AI was really improved.
IT'S FALSE!
THERE'S NO REAL AI IMPROVEMENT!



Now, but a great volume of work spent after WIA scripted event-commands we have no real AI improvements, on the contrary new troubles and AI weaknesses have been introduced.
For example, AI will keep moving against a region with great aggressiveness attacking it even with very unfavourable strength troops ratios: a suicide for the AI army.
Or AI will keep moving forces to achieved the scheduled objective even in winter season, being quickly decimated by harsh weather conditions.
Or even worst, AI main army will be launched after a scheduled objective at the wrong time, when player army are close to the AI capital that will soon fall undefended.
The examples I enlist here have all been experienced when playing WIA and mostly reported by public posts or addressed to the AGEod support.






There is somtehing I can't understand: either common players are just so dumb they can be deceived about WIA AI smartness, or the AI is really better than average so the weaknesses related aren't so terrible as pointed out.

Or jastaV post is just another mess...



GShock -> RE: AI improvements? (3/12/2009 4:50:29 AM)

I doubt it is humanly possible to build a scenario with an AI perfectly instructed how to play. An AI that might still lose but give you a challenge under these conditions. It doesn't for Athena even at the hardest handicap on the player is not a troublesome opponent.

The point on WiA still remains. I am yet to conclude a single campaign without winning in the first few turns to a very lousy AI.

I carry on to spot problems which i report hoping that something will change...and it always has changed for the better. This is so far my entire experience with this game. I should be paid to playtest it instead of having paid to playtest it.
We are past the 6 months and reaching the Bday and the game is still heavily flawed. Ok, let's take away the AI problem and focus on multiplayer... you siege opponents and they gain strenght while you lose it to weather attrition...you must admit it's not a good display. I wouldnt play it...and in fact, i don't. I wish i could but i just can't and you can expose JastaV's second reasons or mine (for all i care) a hundred times.

The bugs are there.

I didn't pay for this game to be a volounteer. I paid to play it and i never could.
There's companies here at matrix whose successful projects need 1 patch per year and who never have a ctd at v1.0. Their games cost much less, are quite enjoyable, their support is decent, their volounteers don't harass people and the AI is enjoyable...there's no excuse.

quote:

There is somtehing I can't understand: either common players are just so dumb they can be deceived about WIA AI smartness,
or the AI is really better than average so the weaknesses related aren't so terrible as pointed out,
Or jastaV post is just another mess...


There are no players, "Blueprint".
Minimal traffic @ ageod, no traffic at all here @ matrix; It is as evident as the sun in a clear sky...you just can't miss it.



Blueprint -> RE: AI improvements? (3/12/2009 8:47:01 AM)

Gshock, you have been removed from the Ageod beta group. Since, you're posting here about WIA, and the current conclusion you can't play it because of a bug . Do you really believe someone knowing this will not conclude you have for personal reasons an agenda against AGEOD? If you're right about the WIA unplayability, the real thing many would make would be shelving the game, forget AGEOD and go after other things.

You were volunteer from a long time for AGEOD. And if you were removed, it was against your will and for something totally different than troubles in AGEOD games.I will not state what was the reason, but I should point out until this removal you were rather enthusiast about Ageod support and games...

For the rest of your statement about WIA sells and state, anyone will forge its own opinion about the value of your argument in light of your partiality. This thread, where you jumped in the JastaV wagon to assess on wrongly base the Ageod AI behaviour is just another example of your need to transform a mouse into a bear, just for bashing AGEOD.




GShock -> RE: AI improvements? Siege improvements? (3/12/2009 9:24:41 AM)

Are you guys going to fix the problem of besieged troops receiving replacements and food inside the settlement while the forces laying siege starve and die to attrition?

Seems quite about time to fix it...the game was released 8 months ago. [:)]





Arsan -> RE: AI improvements? Siege improvements? (3/12/2009 11:28:02 AM)

quote:



The point on WiA still remains. I am yet to conclude a single campaign without winning in the first few turns to a very lousy AI.




Bullshit [:)]

Would you please care to post the campaign saves where you win on the first turns as you state?? It would be interesting to see [:D]
I mean a campaign, as you said. Not a 5 turn scenario where you win on turn 4 [;)]

You must be a genius or i may be stupid, as i'm currently playing and enjoying a 1755 campaign as the Brits against the AI and after 4 and a half years of war (more than 50 turns) i still need to take Louisburg and Quebec and most of Canada.
I know i will win on the end, but the Ai is giving me a lot of work (and fun) in the process.

I've playing AGEOD games against the Ai since BoA (2006) and still find the AI fun. Better than 99% of the other AI's.
You can't ask an AI to be a military genius. Just a worthy opponent. An AGEOD AI definitely is. One of the best around.
And its still improving with each patch. [&o]

You know what i think about you two (Gshock and JastaV) guys: that you are pissed off personally with AGEOD but so hooked on their excellent games than you cannot stop playing and thinking and posting about them even if just to ramble and lie.
Yours is a sad, sad story [8|]



dunnsa -> RE: AI improvements? Siege improvements? (3/12/2009 12:07:44 PM)


quote:

ORIGINAL: GShock

Are you guys going to fix the problem of besieged troops receiving replacements and food inside the settlement while the forces laying siege starve and die to attrition?

Seems quite about time to fix it...the game was released 8 months ago. [:)]





Yes. But only if ports are also blocked.

I wasn't aware that you require a detailed update of beta activities for WIA.

The game I assume you got for free while still a beta will have another free update.



Stwa -> RE: AI improvements? Siege improvements? (3/12/2009 12:12:46 PM)

I always liked the AI in BoA a lot better than the one that came with WIA. WIA's AI was overly aggressive on the low settings IMHO. Sometimes I like the AI to just sit on the main objectives and force me to attack. It took a lot of significant mods of other things (like the removal of tracks), to help the AI out a bit.

As far as I am concerned BoA was the best game in the series. (after modification at least). This partly reflects my intense liking for the period and the style of warfare.

The more you modded this game closer to Phils original design, the better it got.

-------------------------------------------------------------------------------------------------------------------------------------
BTW for BOA - the 1775/1776 campaigns had a lot of scripts there were "dynamic" sort of speak, that would insure that arrival times and destinations for the main reinforcements were randomized in time and location, most notable would be the French and Spanish arrival as allies.

The original 1755 campaing did not have this kind of randomization built into its scripts for arriving reinforcements, but over time I changed these.

Most all of the smaller scenarios were fairly rigid, since they were vignettes, intended to educate the player and provide some historical background.



JastaV -> RE: AI improvements? Siege improvements? (3/12/2009 3:42:29 PM)

quote:

Blueprint:
NCP being posterior to AACW these commands could be implemented in NCP.

Indeed.
In past weeks I experienced myself with their use in NCP mods. It was so I guessed their effectiveness could be improved by extensive use of triggering conditions. Then, pools of alternative commands dictating priority to different scenario objectives to be randomly choose from could keep variability high.


quote:

Blueprint:
These commands aren't what JastaV names scripted events. Hes talking before knowing.
These commands are just helping AI to concentrate on some objectives and play more or less offensively in a context based for its assesments on several factors whose the more important are the enemy threaths. Adding these commands is just adding to the tactical AI a better assesment of the strategic situationS.



A “live” example from WIA Pequot War scenario could help illustrating things……


GBR; evt_nam_Boost_AI;StartEvent;1
NULL;NULL;Conditions;NULL
NULL;NULL;MinDate;1636/12/01
NULL;NULL;CheckAILevel;1
NULL;NULL;Actions;NULL
NULL;NULL;AI.SetAggro ;$New_England|200
NULL;NULL;AI.ChgLocalInterest;$Middleton|150|$NewLondon|200|$Boston|20|$Providence|20|$Norwich|20
NULL;NULL;EndEvent;NULL


The condition for events activation is a fixed date: 1636/12/01. Then the fact GBR side is to be AI controlled.
It means the support to AI will be triggered any time Pequot scenario is played for the player = Indian side at the scheduled date.
No conditions for preliminary controls of objects will be considered. No preliminary evaluation of the strategic overall picture.
The activation , (actions) will introduce “changes” influencing the AI behavior, with an increased AI aggressiveness in the New England macro-area and major interest for some specific regions as New London, (200), Middleton (150).

BTW, Pequot War scenario is played over an extremely small area of the map: 16 regions at all and exclusively located in the New England theater.
Then there are only 3 Strategic towns: Mohigan Village, Misistuk, Hartford.
Middleton region Pequot Village is the only Objective in the scenario.
It’s dramatic to observe the AI is not able to rationalize and take its smart decisions over a so mall gaming area, with very few Strategic towns, just one objective and very few units on play.


quote:

Blueprint:
Now is the AGE AI perfect? No. Is the AGE AI predictable?Partly yes. Is it another wargame having an AI as smart as described by JastaV? I don't think? Should JastaV and Gshock learn before to write? definitly yes.


AI smartness!
This is a very interesting point.
I played WIA, NCP, AACW, BOA extensively: curiously, but the side played out on many scenarios against AI, It always resulted in my victory. Most of time a decisive victory. As a player I could draw strongly unbalanced scenarios as Waterloo playing French side!
I have a rich collection of games. I played all Europe Universalis ser. Games; I played all Total War Ser. Games, (recently I’m going after ETW!). Let’s add to the list Sins of a Solar Empire, ( I bought it because of it’s stellar reputation on the net); then Warhammer Mark of Chaos Series. That to quote game still installed in my HDs.
Is AI from any of this games superior to Ageod products one?
No!
No!
No!
I never said it!


Warhammer Mark of Chaos battles are fun: great to be played and great lo look to because of an amazing graphic!
Sins of a Solar Empire has a great graphic standard too: but AI is stupid to the point you cat nail it at will.
Europe Universalis ser. offers a walk through history and a wide choice of powers to select from. You can take the challenge to lead a great power to world wide domination, or you can rule a small power looking to achieve cultural or mercantile superiority. Anyway, here too player wins!
Total War Series games? Great fun again. The battles resolution in ETW is impressive as regard graphic and motion capture detail. It seems to be on the set of a great historical movie. It’s all: AI is so dumb to wipe out full units because of friendly fire… then, the game has just been released and fixes are needed and will come for sure!

I never said other strategy games, (turn based or real time) are superior to AGEod titles.
I never spent so many time after a strategy game as I did after AGEod titles!
That modding; that playing: mostly modding and play-testing indeed.
Did that because looking to denigrate AGEod and its products?
No!
No, No!
I would not spent months modding NCP and helping the program developers in such a perspective!
My attention and care to AGEod games is dictated by the great potential the games have.
OOB are extremely accurate and detailed: yes there are mistakes in some scenarios! We can fix them of live with such problems.
But think of a thing: I worked for months going after sources and then encoding data editing the “Prelude to Leipzig” scenario, (it’s part of my NCP mods). I assume something like that happened with default scenarios with peoples going mad after any of them.
Why?
Why we should dedicate so long care to OOB editing, to researches, to events adding flavor, to the creation of a detailed historical map, when all is then wasted by a dumb AI, ( not better not worst than other games AI) or by game issues?

Is so hard to catch out a good AI?

I’m a classified chess player.
For the same money you buy any of the games I quoted you can buy a PC chess program. Then depending on the difficulties setting you will chose you will soon reach a point when you could not match the AI!
Is so hard to develop an efficient AI for strategy games?

The spirit of my thread was not looking to denigrate the AI but to underline two questions.
1. Is there any chance to get a flexible and challenging AI without need to encode scenario specific commands supporting Athena moves?
2. Supposing scenario specific commands supporting the AI are a need we should get them as flexible as we can… even at the cost of a hard, long editing job.



quote:

To Clovis:
This could only be done until now by tweaking objectives values of regions, a trick used by JastaV in his now unpublished NCP mod.



Clovis, you are wasting no chance to remark the affair of my NCP mods.
They came out of my efforts: most of time boycotted by other AGEod team-mates.
You denigrated and spammed them!
Then you are here to claim right of free access to them again.
As I had chances to say in the past my mods public availability is not a problem: not for me.
Is AGEod somehow interested after them?
Well, they have just to contact me: I trust we’ll agree over their public availability terms.
But, you Clovis, you are not AGEod!



quote:

To GShock:
...now if they fixed things a bit at least we could play a pbem...


You are a bit off-topic here.
We do not need AI improvements to play PBEM!



GShock -> RE: AI improvements? Siege improvements? (3/12/2009 7:18:03 PM)


quote:

ORIGINAL: lodilefty


Yes. But only if ports are also blocked.

I wasn't aware that you require a detailed update of beta activities for WIA.

The game I assume you got for free while still a beta will have another free update.


U assume wrong Lefty. Check with your accounting department.

I don't require a detailed update of beta activities, I am just telling you that besieged troops can't be allowed to receive replacements regardless of the presence of a port.

Then you may call it update but until the sieges are fixed, i call them post release beta processing as per standard AgeOD procedure.



Arsan -> RE: AI improvements? Siege improvements? (3/12/2009 8:10:59 PM)

quote:

GShock
I am just telling you that besieged troops can't be allowed to receive replacements regardless of the presence of a port.


Why not?? Just because you said so?
any historical or rational backup to explain why not blockaded ports cannot be used to receive replacement in addition to supplies??
Sorry, but it's not your game you cannot impose your arbitrary ideas [;)]



Blueprint -> RE: AI improvements? Siege improvements? (3/12/2009 9:22:01 PM)


quote:

ORIGINAL: GShock



Then you may call it update but until the sieges are fixed, i call them post release beta processing as per standard AgeOD procedure.


per standard AgeOD procedure...partiality is surfacing fast...



GShock -> RE: AI improvements? Siege improvements? (3/12/2009 9:23:35 PM)

Simply because by definition, a siege is the encirclement of a settlement aimed to starve the defenders. This means supplies or reinforcements can't pass.

There might be settlements where the port is annexed to the settlement and in this case that's the FIRST part of the settlement to be overrun when the siege begins and since this part is OUTSIDE the walls (and it has to be otherwise attackers would enter the settlement freely and overrun it) without a force stationed outside the port would automatically be blockaded when the siege begins.

Leningrad is perhaps the only siege that didn't follow this guideline. In fact, it didn't fall. [:)]
It was more of a constant bombardment trying to force submission but not a real siege, in fact supplies were being shipped from the non-encircled side of the town including across the ice and through the port. That's exactly because it was NOT a siege. Still this example is not applicable. In WW2 Leningrad was about 50 times bigger than any of the settlements present in WIA and the germans simply didn't have the numbers to encircle it. This detail is omitted in the abstraction of WiA, you can siege any structure size with any number of units which means...you either siege and autoblockade or you don't siege at all (we need a minimum manpower requirement which is missing in this case so the best solution should be to siege and autoblockade). Remember the player has no control of where the replacements will go and surely wouldn't send them to a besieged settlement if he could. He would rather send them to the forces going to relieve the siege; the so called rescue party.

Furthenmore how does the situation stand in settlements that do not have a port but which do have a depot? U can't blockade those so what...free supplies and free replacements for the besieged while the besiegers starve and die to attrition?

You do understand that however you want to put it, there's a problem here? Let's not hassle and see what comes in this update. If siege timing has been altered and new prototype works well the impact of supplies and replacements in favor of a besieged force might not be so huge on gameplay. Can't really say right now.



JastaV -> RE: AI improvements? Siege improvements? (3/12/2009 9:29:37 PM)

There're not arbitrary ideas, but just ideas!
Best way to deal with them is to download last updated version of DB files and verify how own ideas work in a modded scenario.
Most of time modders will return over own steps after testing!

Unfortunately, GShock's requests over different management of replacements and reinforcements during a siege, (especially as regard besieged forces), cannot be managed by a simple scenario modding attempt. Guess an engine, hard coded script re-working is needed.

As regard the specific idea I can trust it!
Anybody reasonable will do: during a siege, besieged forces will not get external replacements for sure!... although looking at history: see Mantua siege during first Napoleon's campaign in Italy!
Most wargames, (boardgames expecially) have detailded rules as regard prohibition for besieged forces to take replacements!
Guess, improving AGEod engine according to GShock's suggestion could be a great hit and a positive achivement for WIA as other Ageod games, and for players looking to historical accuracy too.
It's then to be said that the issue is known and common to all sides: player and AI managed. As so, it is not an issue unbalancing the game and players could live with it till the time AGeod Dev. & Suppport team will fix it.
I trust they'll work to fix it!

BTW, I understend GShock disappointment, having till to wait for fixes and improvements so many monts after WIA publication.





Arsan -> RE: AI improvements? Siege improvements? (3/12/2009 9:38:33 PM)

quote:

ORIGINAL: GShock

Simply because by definition, a siege is the encirclement of a settlement aimed to starve the defenders. This means supplies or reinforcements can't pass.

There might be settlements where the port is annexed to the settlement and in this case that's the FIRST part of the settlement to be overrun when the siege begins and since this part is OUTSIDE the walls (and it has to be otherwise attackers would enter the settlement freely and overrun it) without a force stationed outside the port would automatically be blockaded when the siege begins.



You might have heard about Boston 1775 siege for example [;)]
http://en.wikipedia.org/wiki/Siege_of_Boston
No need to go to WW2 to look for examples of need to blockade a port to effectively siege...

Another WIA era example: Louisbourg
http://en.wikipedia.org/wiki/Siege_of_Louisbourg_(1758)




Blueprint -> RE: AI improvements? Siege improvements? (3/12/2009 10:03:51 PM)



quote:



Indeed.
In past weeks I experienced myself with their use in NCP mods. It was so I guessed their effectiveness could be improved by extensive use of triggering conditions. Then, pools of alternative commands dictating priority to different scenario objectives to be randomly choose from could keep variability high.



Great.Now the problem is your NCP work has been removed from you the day you decided to leave AGEOD forum. We 're now 3 months after, and your mod isn't yet dowloadable. You may cliam because of intellectual property concern, lack of other site than Ageod or anything. i just persist to think it's pretty interesting to look at someone who reutinely ask for bug fixing or feature development in blunting terms, to say the last, but is for its own unable to just to deliver his own work in a more convenient way than the recent offer to send the mod by e-mail...


About the Pequot war scenario:

1) it's a very small scenario indeed. I just would know with so few objectives and time span why the AI should be prepared to face a large variety of situation; Any small scenario is drawing generally the same picture with slight variation.

2) The AGE AI is able to deal with these variations. The command event was introduced in this scenario, as it was openly discussend in Ageod forum, because the pequot AI wasn't sufficiently aggressive. the evnt is just inciting pequot Ai to take the offensive. Now, if the Pequot hadn't taken any offensive in reality, the war would never had occurred. And there would have been no need to simulate this.

3) You're maybe a chess player, a good modder but you're once again doing the worst in argumentation: concluding by wrong generalization based on non-tested and very small experiments or halfly learned facts.Obviously, you never had played the Pequot war scenario or at last not sufficently to feel the limits of your post.


For my own,since one year, I'm working on AACW AI. I'm reading each day the AI log file produced by the game, each log being thousand lines long.i've scripted events to tailor AI. i've ssen AACW updates improving AI behaviour on what can't be modded. You wonder why working on OOB and modding stuff if the AI is dumb?

Simply because by experience and work, I've got a better AI and if my modding efforts have had such a results, it's only because the hardcoded AI is a good one, contrary to your obtuse and wrong assesments you're displaying routinely since some weeks here.

What I say is just backed by work, not by hate...you've just predicted two days ago on this forum the next AGEOD game will be full of bugs and unplayable. Do you really believe beyond your circonvoluted prose, based on the contradictory fact you're discussing modding about a game you consider unplayable, anyone will not spot your real motives?












quote:

Why we should dedicate so long care to OOB editing, to researches, to events adding flavor, to the creation of a detailed historical map, when all is then wasted by a dumb AI, ( not better not worst than other games AI) or by game issues?

Is so hard to catch out a good AI?


Manifestly yes considering your "dumb AI" games list... Now what's lacking is your 71 posts in any forum about the unplayability of this other games.



Blueprint -> RE: AI improvements? Siege improvements? (3/12/2009 10:21:53 PM)

quote:

ORIGINAL: JastaV

There're not arbitrary ideas, but just ideas!
Best way to deal with them is to download last updated version of DB files and verify how own ideas work in a modded scenario.
Most of time modders will return over own steps after testing!

Unfortunately, GShock's requests over different management of replacements and reinforcements during a siege, (especially as regard besieged forces), cannot be managed by a simple scenario modding attempt. Guess an engine, hard coded script re-working is needed.

As regard the specific idea I can trust it!
Anybody reasonable will do: during a siege, besieged forces will not get external replacements for sure!... although looking at history: see Mantua siege during first Napoleon's campaign in Italy!







except by sea...You're displaying here the same care to back up your historical assesment than in evaluating Ageod AI value. Look a little to the Arsan links before... And I didn't known Mantua is a sea harbour....There's really a trouble with your examples because they are always out of context.

If we had to call bugs errors in your posts, I fear we would conclude the situation to be beyond any fixing hope...



GShock -> RE: AI improvements? Siege improvements? (3/12/2009 11:48:45 PM)

quote:

ORIGINAL: Arsan

You might have heard about Boston 1775 siege for example [;)]
http://en.wikipedia.org/wiki/Siege_of_Boston
No need to go to WW2 to look for examples of need to blockade a port to effectively siege...

Another WIA era example: Louisbourg
http://en.wikipedia.org/wiki/Siege_of_Louisbourg_(1758)



I'm not saying it's not possible for supplies and replacements to pass, i am saying since there's no safety control on the minimum size of a force required to siege a settlement this development choice leaves no solution but to exclude supply and reinforcements from all besieged structures regardless of the presence of a port.

It looks to me the best solution is to add such check adding a minimum requirement of units (or men) to siege a structure according to its level. When in presence of a port, more units would be required to also achieve the blockade by land (unless you blockade from the sea of course so it's 1 check for ordinary settlements and 2 checks for settlements with ports).

Another important achievement of such control would be that the party inside the unbesieged structure, (control 1 or 1+1 not beaten by manpower), would be allowed to get out of the structure and into the region, handle the posture buttons changing the MC in the zone and actually react to a "siege attempt" without need of assistance from rescue parties. This thing would definitely help the AI fight better especially since it's still incapable of real sea operations and solve once and for all the problem that 100 men can siege effectively against 100.000 at present time.

There are hundreds of settlements scattered on the map and the problem is not just port-related. This choice was illogical but it's never too late to change it. Actually i'd say it's quite time to do something about the sieges.

Anyway...as i said, let's see this new prototype and hope at least the problem can be reduced but i think without this double check it can never be eliminated unless someone else has a better idea in merit, in which case i'd love to hear it.



Blueprint -> RE: AI improvements? Siege improvements? (3/12/2009 11:57:07 PM)

if 100 men sieges against 100.000 at present time, the besieged will wipe out the besieging force at will, ending the siege.[8|]





Stwa -> RE: AI improvements? Siege improvements? (3/13/2009 8:16:28 AM)

BTW,

The Perquot scenario is really out of place in the WIA game. Its pretty shallow scenario anyway. If someone doesn't like it per se. Then just delete it. Out of sight -- out of mind. [:)]

Even, with BoA, once in a great while I while go back a play on of the small scenarios. But for the main part, I really only play 1755, 1775, 1813, and my version of 1778.



Arsan -> RE: AI improvements? Siege improvements? (3/13/2009 9:51:03 AM)

quote:

ORIGINAL: GShock

Another important achievement of such control would be that the party inside the unbesieged structure, (control 1 or 1+1 not beaten by manpower), would be allowed to get out of the structure and into the region, handle the posture buttons changing the MC in the zone and actually react to a "siege attempt" without need of assistance from rescue parties. This thing would definitely help the AI fight better especially since it's still incapable of real sea operations and solve once and for all the problem that 100 men can siege effectively against 100.000 at present time.



Don't you know that the force under siege can be just dragged outside the structure and into the same region anytime you like and with an offensive posture will attack and try to expel the besiegers?? [&:]
It's been like this since BoA. 1.0 three years ago for what i know.[;)]
No need to wait for a relief force. The sally button is just to coordinate the attack from the units inside the structure with the arrival to the region of an relieve force, so both attack the besieger together.





dunnsa -> RE: AI improvements? Siege improvements? (3/13/2009 11:54:45 AM)


quote:

ORIGINAL: Stwa

BTW,

The Perquot scenario is really out of place in the WIA game. Its pretty shallow scenario anyway. If someone doesn't like it per se. Then just delete it. Out of sight -- out of mind. [:)]

Even, with BoA, once in a great while I while go back a play on of the small scenarios. But for the main part, I really only play 1755, 1775, 1813, and my version of 1778.


Eh? Wars in America? [&:]

Pequot War.
Fought in America

Of course, you are welcome to your opinion.[:)]



JastaV -> RE: AI improvements? Siege improvements? (3/13/2009 1:35:38 PM)


quote:

ORIGINAL: Blueprint
And I didn't known Mantua is a sea harbour....There's really a trouble with your examples because they are always out of context.

If we had to call bugs errors in your posts, I fear we would conclude the situation to be beyond any fixing hope...


Mantua is and was a river port on Mincio.
Austrian used flatboats to resupply their besieged garrison night-time.



Stwa -> RE: AI improvements? Siege improvements? (3/13/2009 5:08:02 PM)


quote:

ORIGINAL: lodilefty


quote:

ORIGINAL: Stwa

BTW,

The Perquot scenario is really out of place in the WIA game. Its pretty shallow scenario anyway. If someone doesn't like it per se. Then just delete it. Out of sight -- out of mind. [:)]

Even, with BoA, once in a great while I while go back a play on of the small scenarios. But for the main part, I really only play 1755, 1775, 1813, and my version of 1778.


Eh? Wars in America? [&:]

Pequot War.
Fought in America

Of course, you are welcome to your opinion.[:)]


Lodi, I mean 17th Century stuff just ripped into a game that was really ment for 18th Century. For that matter, you didn't include any Conquistador stuff, right? In a way that was Wars in America as well. Plus, I don't think there are any real graphics for this very early era, maybe I missed em.





berto -> RE: AI improvements? Siege improvements? (3/13/2009 5:47:33 PM)

quote:

ORIGINAL: JastaV

Is so hard to catch out a good AI?

...

Is so hard to develop an efficient AI for strategy games?



Apparently, for some games, for some scenarios, the answers are: yes, and yes.

When I am disappointed with the AI--not all the time, just some of the time; not just for AGEOD games, but for games from any publisher--my solution is:

Hot-seat Solitaire!

I've been in the war gaming hobby for 45+ years. Years ago, when the hobby was even more obscure, when it was a major hassle to coordinate extended play sessions with other like-minded players (few in number; widely dispersed, most likely interested in a different game and/or historical period), I developed the habit of playing 99% of my games solitaire. It's a habit that has sustained me through most of my game playing life.

The main thing you give up is Fog of War, but you can fake this. "House" (mind) rules.

When I am disappointed with the AI, my response is not to dump the game, much less to denigrate the author/publisher. It is simply to "play" the "game" differently. Against myself. (PBEM? Not for me. I don't need to explain why--again. I've got my reasons. Don't ask.)

Hot-seat solitaire. Works for me!



JastaV -> RE: AI improvements? Siege improvements? (3/13/2009 9:42:12 PM)


quote:

ORIGINAL: berto

quote:

ORIGINAL: JastaV

Is so hard to catch out a good AI?

...

Is so hard to develop an efficient AI for strategy games?



Apparently, for some games, for some scenarios, the answers are: yes, and yes.

When I am disappointed with the AI--not all the time, just some of the time; not just for AGEOD games, but for games from any publisher--my solution is:

Hot-seat Solitaire!

I've been in the war gaming hobby for 45+ years. Years ago, when the hobby was even more obscure, when it was a major hassle to coordinate extended play sessions with other like-minded players (few in number; widely dispersed, most likely interested in a different game and/or historical period), I developed the habit of playing 99% of my games solitaire. It's a habit that has sustained me through most of my game playing life.

The main thing you give up is Fog of War, but you can fake this. "House" (mind) rules.

When I am disappointed with the AI, my response is not to dump the game, much less to denigrate the author/publisher. It is simply to "play" the "game" differently. Against myself. (PBEM? Not for me. I don't need to explain why--again. I've got my reasons. Don't ask.)

Hot-seat solitaire. Works for me!


From an old boargamer, (but not as old as you) to an old boardgamer: I know that practice and I spent so many times after it !
But I moved from the board to PC just for seeking an opponent.
Then, going after your suggestion, many gamers will stop buying PC games...... guess it will not greately help authors/publishers.



GShock -> RE: AI improvements? Siege improvements? (3/14/2009 12:52:38 AM)

Well it appears that according to a thread message in last January 2008, the besieged force is actually able to lift the siege by sallying out without the help of a rescue party (+ sortie button). It was fun to read that at that point in time, and so much past the release, several gurus of the game had no idea of how actually the procedure worked and if it was at all possible. Me myself, as you may have noticed, knew nothing about this.

So actually Clov...ehm "Blueprint" is is right. Perhaps it might be a good idea to have the DEVs prepare a server for RSS news and mailing newsletter so that when rules change all of those who can't read the forums everyday (and take notes of the changes applied by 2957239 patches per year) can still play knowing what the rules and available choices are.

Source: http://www.ageod.com/forums/showthread.php?p=65621

Meanwhile, and reconnecting to the "logical choice of allowing replacements and reinforcements even during a siege"

I have explained that since the game doesn't check for manpower even 100 men can siege a structure with 100.000 inside and luckily "blueprint" is right, that force would be blown away by the defenders sallying. Still, in the meanwhile defenders would be under siege till next turn's resolution phase while the structure produces reinforcements and gives replacements to the besieged party.

Some are really convinced that the blockade would stop this from happening...so this choice is WAD (it means that in absence of a port, any depot city will not only produce supplies but also give replacements to the besieged party while the besiegers starve and die of cold)...so I have taken 5 minutes to search a bit ... and i found...


I have several line replacements, just sitting there, but my depleted units are not filling up. Why?
A unit needs to be in either a depot, harbor, or a level 2 city (not besieged) in order to draw replacements. Note also that partisans and irregulars can regain hits for free if a region is very loyal. 
Source: http://ageod.nsen.ch/aacwwiki/FAQ#I_have_several_line_replacements.2C_just_sitting_there.2C_but_my_depleted_units_are_not_filling_up._Why.3F

It appears to me the text specifies in order to receive replacements, the city must not be besieged and there's no mention about blockading ports.



Page: [1] 2 3   next >   >>

Valid CSS!




Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI
0.9375