User avatar
Pocus
Posts: 25673
Joined: Wed Oct 19, 2005 7:37 am
Location: Lyon (France)

What are acceptables advantages for the AI player?

Tue Oct 13, 2009 4:54 pm

Hi,

We would like to know what, for you, is acceptable, as advantages given to the AI, for these levels of difficulty:

Easy, Normal, Difficult, Very Difficult, Insane

For examples:
Production bonus, cohesion recovery bonus, combat bonus, activation bonus, leadership bonus, diplomacy bonus.

But also, 'AI helpers', a brand new category we are considering... AI helpers are new algorithms for the AI which can "twist" the rules, let me give you an example:

An AI helper algorithm would be the possibility for the AI to build a supply depot in a region, without a supply wagon there, but only if the region and the adjacent ones are void of any enemy, and if there is a structure with a friendly combat unit in the target region. If ok, the AI would expend resources and a depot building would be initiated.

The same kind of test would be used to repair a railroad, even if there is no friendly unit to rebuild 'physically' the railroad. The abstraction here is that without an enemy, it is acceptable to allow this 'twist' of rule for the AI, which, let's be frank, can have at most the brain power of a children (and you are not mean enough to beat a children, are you?) :neener: .

So, are AI helpers pure cheat that should not even be considered, or can they be considered, as an option, or integrated implicitly starting a certain difficulty level (which one?). Last question, what 'Helpers' you would accept (automated troops ferrying over the ocean in time of peace, etc.?)

Feel free to give your opinion(s)!
Image


Hofstadter's Law: "It always takes longer than you expect, even when you take into account Hofstadter's law."

User avatar
Jim-NC
Posts: 2981
Joined: Wed Feb 25, 2009 4:21 pm
Location: Near Region 209, North Carolina

Tue Oct 13, 2009 5:17 pm

IMHO, the railroad repair is necessary. I would think that giving bonuses at difficult would be appropriate (with bigger bonuses at very difficult and insane).

Other "helpers" could be a mixed bag (some good, some bad). I would have them as options (you get to decide how difficult you want to make your game). That way you can customize the game to your liking.
Remember - The beatings will continue until morale improves.
[SIGPIC][/SIGPIC]

User avatar
lodilefty
Posts: 7616
Joined: Sat Aug 11, 2007 3:27 pm
Location: Finger Lakes, NY GMT -5 US Eastern

Tue Oct 13, 2009 6:06 pm

It's pretty easy to set a 'helper' to an AI level. So we can already set the amount of help that way....

What's harder is to really test the effect! Takes many game hours of 'play-tweak-play-tweak' to get it 'right'..... :bonk:

...and 'right' for me may not be 'right' for you. :blink:

I [obviously, for WIA users :D ] strongly support 'helpers', as types suggested by Pocus and more [Local Interest by region, Agression, force leaders active, unlocking a stack early, even historically-based 'major move events' like the British evacuation of Boston ;) ]
Always ask yourself: "Am I part of the Solution?" If you aren't, then you are part of the Problem!
[CENTER][/CENTER]
[CENTER]Visit AGEWiki - your increasingly comprehensive source for information about AGE games[/CENTER]

[CENTER]Rules for new members[/CENTER]
[CENTER]Forum Rules[/CENTER]

[CENTER]Help desk: support@slitherine.co.uk[/CENTER]

User avatar
Franciscus
Posts: 4571
Joined: Fri Apr 20, 2007 8:31 pm
Location: Portugal

Tue Oct 13, 2009 6:27 pm

Very interesting question. First, let me tell you that I do not know nothing about AI. :)
With that in mind let me tell you what I as a player do not like - to see the AI using different rules from my own and also seeing a super- human AI. So I tend not to like options that give AI forces increased strength, resistence, cohesion, etc. If we go that way, one might simply to program the AI to be armed with laser guns vs our rifles :D . "very difficult" levels tend to bend reality in those ways and I do not choose them to play. As to the AI helpers exemplified by Pocus, at least the examples cited are simply a simplification of the rules, that I would not mind seeing implemented also for the human player. For instance, the difficult thing about building a depot for the AI IMHO is the decision to build it and where. Afterwards, if the process is stramlined, it should be also for the human player - kind of what happened with the rules about division formation in AACW.
What I would not mind at all to see for instance in VGN as an AI helper would be...scripts ! :bonk: . Let me explain. As I tend to like to play historicallyand as Ageod games are usually full of events, I would like to see The Ai being controlled by scripts after a significant event affects her. For instance (and I know nothing about VGN) suppose that Prussia is played by AI. If a germanic unification events would trigger, scripts that would emulate what an average human player do would trigger (diplo actions, raising armies, etc) and if the AI lacked at the moment the necessary resources they would be given automatically to the AI). I recognize though that to other players this historicity would not be to their liking.
Regards

User avatar
W.Barksdale
AGEod Grognard
Posts: 916
Joined: Tue Feb 05, 2008 8:17 pm
Location: UK

Tue Oct 13, 2009 7:38 pm

I like the idea of 'AI helpers' for railroad and depots. With a little tweaking this may help her during winter campaigns that she often seems so fond of.

The bonus... I think the best would be, if possible, just a production bonus which increases with the level of difficulty. The others make her forces seem 'super-human'.

I always found it strange when playing on the higher difficulty levels (in AACW) how Athena could move so quickly and strike so hard, yet still leave her capital undefended. As she often will not make the 'optimal' decision regarding resource collection, a production\manpower bonus may really help her behevior in this regard.
"Tell General Lee that if he wants a bridge of dead Yankees I can furnish him with one."
-General William Barksdale at Fredericksburg

User avatar
squarian
Brigadier General
Posts: 485
Joined: Thu Aug 07, 2008 7:41 pm
Location: Pittsburgh, Pennsylvania

Tue Oct 13, 2009 8:32 pm

Pocus - First of all, thanks for thinking about this - it's the sort of thing which does AGEOD so much credit as a game company. In replying, I should point out that I'm thinking as much of WIA as AACW (the two I own, so far).

Some people feel differently, but I accept that the AI is going to need some help to provide a challenging game.

That being said, in general and at all levels I would prefer the help to be invisible to the player as much as possible. An extreme and absurd example to the contrary would be a new stack of troops which suddenly materializes near my own troops - in other words, I don't want to play against an AI which can blatantly wave a magic wand to help itself.

Generally, I'd prefer "normal" to equate to the AI playing by the rules as written, without much or any help.

At higher levels, I think that any "cheats" ought to be fairly discreet and behind the scenes. Logistical cheats, like repairing rails or building depots - sure. Perhaps a slight cohesion-recovery bonus, say 5-10% - especially if tied to particular AI offensive war-plan agendas, to simulate the build-up for a big push or a period of recovery after defeat by particular forces rather than across the board. Activation and FoW bonuses, which are already options, are reasonable. I'd be happy for the the AI to have additional advantages which do not pertain directly to combat and movement. The AI should not "cheat" in the basic mechanics, but making it easier for the AI to repair and deploy its assets - in other words, logistical help - seems reasonable to me.

What I would be most interested in are "AI helps" with the more complicated sorts of planning which humans do naturally. An example would be the Boston evacuation events in WIA - the AI seems to have a hard time coordinating amphibious operations, so by all means give it some help there. I don't regard this as a "cheat" like giving it a combat bonus - this is simply a refinement to the AI's thinking, in my view. Likewise, where humans can fairly quickly grasp the need to concentrate troops for an offensive push, the AI seems not always able to do so - and an additional boost to it's abilities in that situation seems normal and desirable.

It may be that this amounts to nothing but asking you to invent a genius - and if so I apologize for being unhelpful. But examples like automatically delivering troops over seas, when the embarkation/transport/debarkation routine is simply too difficult for the AI to carry out, seem perfectly justifiable to my mind.

Finally, I'd just point out that Clovis seems to have achieved real success using the AI-related events you have already provided. A dynamic set of objectives, which can be tailored to direct the AI in response to changes in the state of play, might be the best possible improvement. Whether that's done by you as part of the .exe package or retrofitted by modders just depends on the availability of time and resources.

Thanks again for gnawing away at this problem.

User avatar
arsan
Posts: 6244
Joined: Tue Nov 28, 2006 6:35 pm
Location: Madrid, Spain

Tue Oct 13, 2009 8:43 pm

Hi!

I agree with the previous posters in that i don't like direct and blatant bonuses, specially combat bonuses that brake the historic simulation by destroying the suspension of disbelief.
AI units that fight as twice their number or move twice as faster is s typical example. I never use hard or very hard settings because of that. :(
Instead i don't have much problems with bonuses and AI helpers that are not as obvious. For example, production bonus (within certain limits) are acceptable for me. Or some cohesion/attrition recovery bonus (humans routinely know how to rest his troops and the AI may not do it as good as them) or slight activation bonus like the current one.

Regarding AI helpers like the ones Pocus proposes, i'm completely in favor. IMHO everything that a player routinely does but the AI may have trouble with can get an AI helper.
I mean basic "houskeeping" things like ferrying troops across sea, keeping basic supply system creating some depots and fixing rails, setting up basic garrisons on important towns, heading back to winter quarters in winter, using the right leaders for the right jobs (like raiding leaders for irregulars or cavalry leader for cavalry etc...)
IMHO, the general idea would be: as long as the "cheat" make the game more enjoyable by improving the AI and is not so noticeable that breaks the suspension of disbelief its OK with me.
Another interesting AI helper options are the bigger "event like" helpers. Something like what Franciscus explains (based on historical events) look good to me.
Another kind would be things that helped the AI to overcome some of his limitations.
In WIA, the event that evacuates the forces trapped at game start on the untenable Boston siege is a good example. The AI may evacuate them by herself, but most probably not faster or thoroughly enough to not cripple them by starvation.
To get out of Boston is a no brainer decision so, why don't help her do it right so the game becomes much more interesting for the human USA player??

For example on the AACW something similar could be done by setting up some USA naval invasion events (with variable destinations and chances) so the human CSA would not be assured that their coastal towns were out of the harms way.

Ideally, all this should be done optionally, so people that loath any kind of AI "cheats" would not be obliged to cope with them. :)
Maybe...
a Very Easy level with bonus for the human player,
an Easy level with no helpers to the AI,
a Normal level with AI helpers but not bonuses,
a hard level with AI helper events and some bonus (maybe indirect ones like the productions bonus)
and a very hard with all of this and some combat, movement etc bonus

Just my 2 (long) cents! :)

User avatar
squarian
Brigadier General
Posts: 485
Joined: Thu Aug 07, 2008 7:41 pm
Location: Pittsburgh, Pennsylvania

Tue Oct 13, 2009 9:27 pm

arsan wrote:Maybe...
a Very Easy level with bonus for the human player,
an Easy level with no helpers to the AI,
a Normal level with AI helpers but not bonuses,
a hard level with AI helper events and some bonus (maybe indirect ones like the productions bonus)
and a very hard with all of this and some combat, movement etc bonus



Seems like a reasonable scheme to me.

User avatar
Pocus
Posts: 25673
Joined: Wed Oct 19, 2005 7:37 am
Location: Lyon (France)

Wed Oct 14, 2009 8:02 am

It seems everybody dislike combat bonuses, so removing them completely is something to be considered seriously.

As for AI helpers, please differentiate between:

- generic helper, done by code (logistical helpers like placing a depot)

- scripted events that modify the AI behavior (like in Clovis mod). Here it is an improvement of AI behavior, done by script, and is not a twist of the rules, so they are 100% legit and are not even considered 'helper' for me, as I see them.

- scripted events that allow the AI to do things it can't do normally, like evacuating Boston by script, with a kind of 'magic wand' approach. Definitively an helper, but a scripted one, so I leave that to modders.

I would like your opinion mostly on the first category in fact, additional AI code which ease its logistical ineptness (no magic teleportation of combat units, etc.) by having her play with slightly different rules, on very specific problems.

@ Franciscus, specifically. I'm unsure that, if the AI get these kinds of 'streamlining', it should be legit to give them automatically to the players too. Because the basic premise, that AI and player can play on the same ground, is wrong by essence. You are (for most of you ;) ) being of flesh with a very powerful brain and 20+ years of complex cognitive development. The AI, is not an AI as in sci-fi movies, but just some piece of code assembled by a developer which has at most some hundreds hours of work to do his best, knowing that what he can do will approach at most a few % of the proficiency of a talented wargamer.

So things being by essence not equal, why in this case give to the player all the 'helpers' of an AI? Because this is not fair for the player? ;)
Image


Hofstadter's Law: "It always takes longer than you expect, even when you take into account Hofstadter's law."

User avatar
Carnium
Posts: 2115
Joined: Thu Jan 17, 2008 2:08 pm
Location: Slovenia

Wed Oct 14, 2009 9:50 am

Are there any plans to link those AI helpers and bonuses with the level of game difficulty ?
Some preset rules would be good to have, especially for beginners and for testing purposes.

My wishes for difficulty levels are as follows:

Very Easy --- combat and activation bonuses for players. Suitable only for novice players.

Easy --- only non combat bonuses and helpers that would help the AI in organizing and recruiting its armies. Suitable for players that play once in a while.

Normal --- all non combat bonuses that would help the AI in organizing and recruiting its armies plus small activation bonus, small cohesion recovery bonus, small leader and unit spotting bonus and small diplomacy bonus
Suitable for semi-regular players.

Difficult --- all non combat bonus plus cohesion recovery bonus, small combat bonus, activation bonus, unit spotting bonus, medium diplomacy bonus
Suitable for veteran and regular players

Very Difficult --- every AI bonus and helper you can think of, but not at insane level
Suitable only for experts.

Basically bigger bonuses an more helpers for higher level of difficulty.

Some special AI helpers that I would like to see:
- auto garrison feature (like home guard or home defense) for vital AI cities plus fixed garrisons for all captured vital/strategic areas.
- auto repairs for the AI structures, infrastructure and maybe even units
- variable sea invasion script that would be more difficult for the human player to guess and guard the invasion threatened cities/areas (something that I would love to see in WiA)
- anything that can help the AI in re-suppling the units and keeping it active, like your supply depot auto-build or even a special bonus for supply production for the AI ?

richfed
Posts: 902
Joined: Tue Nov 21, 2006 9:50 pm
Location: Marion, North Carolina, USA
Contact: Website

Wed Oct 14, 2009 12:01 pm

I don't use combat bonus options, but having them AS AN OPTION, does no harm. I would leave them in as is.

As far as AI tweaks - rail repair: Good - supply issues: EXCELLENT. You know me, ANYTHING that will help AI troops stay in supply is on my A-list, ESPECIALLY, if it results in solid amphibious assaults and then sustained interior invasion.

In a PBEM game, I have noticed the effect of good blockading technique [by my opponent, not me]. Improving this in the AI would also be a plus.
[color="DarkRed"][SIZE="2"][font="Book Antiqua"]"We've caught them napping!"[/font][/size][/color]

User avatar
lodilefty
Posts: 7616
Joined: Sat Aug 11, 2007 3:27 pm
Location: Finger Lakes, NY GMT -5 US Eastern

Wed Oct 14, 2009 12:10 pm

Sieges.
[INDENT]The AI often abandons a siege after 2-3 turns [going out of supply perhaps] and I can't ever recall seeing the AI position ships to blockade a port during a land siege. This seems to be a bit of logistics [to feed the besieging army] and Naval AI.[/INDENT]

Breaking Sieges.
[INDENT]The corollary of sieges. AI rarely sends relief expedition. If the Naval AI starts supporting sieges, then the Naval AI has to give high priority to breaking a blockaded port also..[/INDENT]
Always ask yourself: "Am I part of the Solution?" If you aren't, then you are part of the Problem!
[CENTER][/CENTER]

[CENTER]Visit AGEWiki - your increasingly comprehensive source for information about AGE games[/CENTER]



[CENTER]Rules for new members[/CENTER]

[CENTER]Forum Rules[/CENTER]



[CENTER]Help desk: support@slitherine.co.uk[/CENTER]

User avatar
McNaughton
Posts: 2766
Joined: Wed Mar 21, 2007 8:47 pm
Location: Toronto, Canada

Wed Oct 14, 2009 1:16 pm

One thing I did for the AI for my AACW mod (that never quite made it) was to give slight supply bonus' for all elements at a 25% chance. Since the AI likes to get their forces out of supply, then retreats (causing manpower losses, cohesion weakness, and basically making a viable combat force a usless hulk), I found that this mitigated this AI behaviour somewhat. The AI did not become more aggressive, and tended to follow the same patterns (advance, retreat), but was in significantly better shape after doing so.

Basically, an event was in place looking for elements that were at a certain limit of supply, and provided extra supply to that element (plus the element had to be on the AI side, plus at a certain difficulty level).

One thing not really mentioned yet in AI bonus', is also things in place to limit the player to follow the rules of the game. For example, many players have 'house rules', such as if McClellan is a 3* general, he must be CinC of the main Union Army instead of what other players do, and put him out west (and put Grant as CinC of your main army).

All too often players complain about AI cheating, but ignore their own (above situation included). I hear of players knowing AI patterns, and purposely working against them (if the AI scatters their forces, the Human concentrates, if the AI concentrates, the Human scatters and envelops, etc.).

The AI will never react as quickly as the human can, and especially a human who puroposely plays to break the rules. One thing I think that needs greater focus in all games is to limit the freedom of players to do anything that they want, given that these are historic wargames.

Anyone playing Germany during WW2 should have to deal with the annoyance of Hitler interfering with your plans. One idea I would have liked to see for HoI2 when modding was events in place for Germany where you would get events of Hitler meddling. However, many people did not want this, as the desire for absolute freedom was so great. However, absolute freedom for the player is incorrect, when we are forcing the AI to follow certain rules...

User avatar
Anguille
Posts: 186
Joined: Thu Sep 28, 2006 2:38 pm
Location: PETIBONVM

Wed Oct 14, 2009 1:30 pm

This is a very good question. However, before saying what kind of bonus or help we should give to the AI, the question is: what are the weaknesses of the AI in comparison to the Human Player? This is something that can be completely different from one game to the other...and blatant weaknesses should be worked on before a game is released. Imho, it's not good to give a bonus if the AI is not weak in an area.

For example:

- if the AI has problems to resupply it's troops, a ai-help should be considered
- if the AI has problems with production, a production bonus would be adequate.

Beta-testing should see where the weaknesses are.

User avatar
lodilefty
Posts: 7616
Joined: Sat Aug 11, 2007 3:27 pm
Location: Finger Lakes, NY GMT -5 US Eastern

Wed Oct 14, 2009 2:37 pm

McNaughton wrote:<snip>
Basically, an event was in place looking for elements that were at a certain limit of supply, and provided extra supply to that element (plus the element had to be on the AI side, plus at a certain difficulty level).

<snip>


Slightly OT, but what command syntax did you use to evaluate Supply? :confused:

VERY applicable in WIA perhaps ;)
Always ask yourself: "Am I part of the Solution?" If you aren't, then you are part of the Problem!
[CENTER][/CENTER]

[CENTER]Visit AGEWiki - your increasingly comprehensive source for information about AGE games[/CENTER]



[CENTER]Rules for new members[/CENTER]

[CENTER]Forum Rules[/CENTER]



[CENTER]Help desk: support@slitherine.co.uk[/CENTER]

User avatar
Franciscus
Posts: 4571
Joined: Fri Apr 20, 2007 8:31 pm
Location: Portugal

Wed Oct 14, 2009 3:27 pm

Pocus wrote:It seems everybody dislike combat bonuses, so removing them completely is something to be considered seriously.

As for AI helpers, please differentiate between:

- generic helper, done by code (logistical helpers like placing a depot)

- scripted events that modify the AI behavior (like in Clovis mod). Here it is an improvement of AI behavior, done by script, and is not a twist of the rules, so they are 100% legit and are not even considered 'helper' for me, as I see them.

- scripted events that allow the AI to do things it can't do normally, like evacuating Boston by script, with a kind of 'magic wand' approach. Definitively an helper, but a scripted one, so I leave that to modders.

I would like your opinion mostly on the first category in fact, additional AI code which ease its logistical ineptness (no magic teleportation of combat units, etc.) by having her play with slightly different rules, on very specific problems.

@ Franciscus, specifically. I'm unsure that, if the AI get these kinds of 'streamlining', it should be legit to give them automatically to the players too. Because the basic premise, that AI and player can play on the same ground, is wrong by essence. You are (for most of you ;) ) being of flesh with a very powerful brain and 20+ years of complex cognitive development. The AI, is not an AI as in sci-fi movies, but just some piece of code assembled by a developer which has at most some hundreds hours of work to do his best, knowing that what he can do will approach at most a few % of the proficiency of a talented wargamer.

So things being by essence not equal, why in this case give to the player all the 'helpers' of an AI? Because this is not fair for the player? ;)



My dear Pocus, I have absolutely no pretension to correct you in anything pertaining to AI programming :) . You have designed one of the best AI's out there, and I am certain that anything you do in the future will be of the same quality. :coeurs:
Maybe I misunderstood you, or maybe I am confusing AI with design/rules. The way I read your example, pertaining specifically to depots, the main difference for the AI would be the lack of need for the AI to have supply wagons present in order to build a depot. The rest of the conditions you mentioned are just basic rules - regions without enemies, with structures and friendly units (conditions that would also apply to the human, I presume). So I read it as just a streamlining of the process of building depots. Let me say also that at least to my kind of playing that level of micromanagement is tedious - I do not like to have to build x wagons and carry them to y spot to build a depot. I want to be able to DECIDE where and when to build a depot, and then , if I have sufficient resources, the process would be automated. The same goes for forts, for instance. Again, the way I read it, I fail to see how the AI helper you mentioned would help the AI DECIDE when and where to build depots - unless she would build depots in EVERY region without enemies, with structures and friendly units - which would be kind of strange :bonk: . The problem as I see it is that that kind of helper would not help AI take good strategic decisions - and that's the AI's great shortcoming, IMHO

Returning to scripts, I am afraid I do not agree that AI helper scripts like the Boston evacuation one are just to be left to the modders. Again, I do not understand a thing about AI's, but I am pretty sure that even comparing them with a child's brain is way too flattering (for the AI, of course :D ). Some people do not like scripts. On the contrary, I think that in the current state of PC gaming they can be a good help. For instance, I would love to see scripts ordering AACW USA AI to mount solid amphibious invasions. It would not be completely predictable - the script would be triggered between certain set dates, with some randomness, the target regions would be picked from a list, depending on importance of the target, defenses, an element of randomness, etc. But the forces would be scripted to be well organized, well led and of sufficient force to be a threat. Why not ? I for one would very much prefer the existence of such "predictable" scripts in a HISTORICAL strategy game as AACW than the current relative inability of the AI to effectively menace my southern coast.

Finally :) , about supply helpers - a word of caution - I would absolutely do not like if the AI would send forces out to the Appalachians in mid winter and they would not suffer due attrition - completely illogical, against the "immersiveness" that I so much like. On the contrary, ideally the AI should be helped to not do such blunders - maybe a script triggering in December ordering the AI units to return to base if possible ? ;)

Best regards.

User avatar
Generalisimo
Posts: 4176
Joined: Wed Jun 07, 2006 10:03 pm
Location: Buenos Aires, Argentina
Contact: ICQ WLM

Wed Oct 14, 2009 3:33 pm

Franciscus wrote:Returning to scripts, I am afraid I do not agree that AI helper scripts like the Boston evacuation one are just to be left to the modders. Again, I do not understand a thing about AI's, but I am pretty sure that even comparing them with a child's brain is way too flattering (for the AI, of course :D ). Some people do not like scripts. On the contrary, I think that in the current state of PC gaming they can be a good help. For instance, I would love to see scripts ordering AACW USA AI to mount solid amphibious invasions. It would not be completely predictable - the script would be triggered between certain set dates, with some randomness, the target regions would be picked from a list, depending on importance of the target, defenses, an element of randomness, etc. But the forces would be scripted to be well organized, well led and of sufficient force to be a threat. Why not ? I for one would very much prefer the existence of such "predictable" scripts in a HISTORICAL strategy game as AACW than the current relative inability of the AI to effectively menace my southern coast.

I think Pocus means by "letting this to modders" is that those AI helpers events can be done by lodi, Gray or me... so, he is not the one behind them. There is no need for him to "waste" his precious time on them when "a modder" can use the commands available to script them. ;)
On the other hand, the ONLY one that can twist the rules is him... ;)
So, there it is the difference... :thumbsup:
"History is the version of past events that people have decided to agree upon."
Napoleon Bonaparte


BOA-AAR: ¡Abajo el imperialismo Británico! (en español)

AGEOD Facebook Fanpage - news & screenshots about the upcoming games!

User avatar
Generalisimo
Posts: 4176
Joined: Wed Jun 07, 2006 10:03 pm
Location: Buenos Aires, Argentina
Contact: ICQ WLM

Wed Oct 14, 2009 3:38 pm

Pocus wrote:- generic helper, done by code (logistical helpers like placing a depot)

I really like those, as long as they are simulated seamlessly inside the game... like you said, special rules that are outside the sight of the player.

I really hate when the AI starts using his magic wand and new units start raising near the frontline (very usual in every combat game to make the game harder :( ).

I also hate the combat bonuses, specially when they are too big... because they just kill the "feel" of the game in my opinion.
But an option is always good... there must be someone out there that uses them... ;)
"History is the version of past events that people have decided to agree upon."
Napoleon Bonaparte




BOA-AAR: ¡Abajo el imperialismo Británico! (en español)



AGEOD Facebook Fanpage - news & screenshots about the upcoming games!

User avatar
Franciscus
Posts: 4571
Joined: Fri Apr 20, 2007 8:31 pm
Location: Portugal

Thu Oct 15, 2009 12:58 pm

Generalisimo wrote:I think Pocus means by "letting this to modders" is that those AI helpers events can be done by lodi, Gray or me... so, he is not the one behind them. There is no need for him to "waste" his precious time on them when "a modder" can use the commands available to script them. ;)
On the other hand, the ONLY one that can twist the rules is him... ;)
So, there it is the difference... :thumbsup:



OK, I see your point. Nevertheless, let me remind you what I think is usually the point of view of an average buyer of PC games - he does not mind who makes the features of the game he bought - the designer, programmer, coordinator, beta tester, or even the janitor of the company :D - he usually does not even care about the features, but only about the end result. And this is what he gets in the game he buys and in the patches that sometimes (not allways) he searches and downloads from a web page.
Whatever is on the official game at release time or in a patch, is not a mod, and ultimately is the responsability of the game devs, regardless of who makes what feature/script/routine/graphic/pic, whatever. Ideally also, most of the fundamental features should be "in" at release time, and the usual "work in progress" status of our games should be kept at a minimum.
Mods, however, are very different - they are made voluntarily by one or more individuals, are not "official", and are used by a very small fraction of the actual players.

So, the existence of hipothetical scripts like the Boston evacuation one in future Ageod games is of course a decision of Ageod. If they are contemplated in the "official" game, I do not really care who makes them. Of course, I will also be happy to see them as a mod, but it is not the same thing, and most of the game buyers will not benefit from them.

Regards.

User avatar
Pocus
Posts: 25673
Joined: Wed Oct 19, 2005 7:37 am
Location: Lyon (France)

Thu Oct 15, 2009 1:42 pm

We are on the same wavelength Franciscus, no worry. What is in the game or is not is anyway in the end the responsibility of the company doing the game. That's why we always said that what our coordinators make as scripts, when included in patches has our full support.

Also, when I said that I left to modders "scripting the AI", implicitly for me it means doing add-on, but the core "thinking processes" must all be there, i.e we are to deliver a complete game, with an AI able to do all that is needed to operate.

"Scripting the AI" is a possibility left to modders because there is always room for improvements, but as a company, we must draw a line, or no games would ever recoup the expenses. What I'm proposing in fact is to think of a new way of helping the AI, coming from a pragmatic understanding: a given dev team never has enough time (except perhaps Blizzard) to make everything as they want, especially for an AI. So why not decide to be the most efficient possible by creating some "twist in the rules", only for the AI, in order to solve big problems for her, in a mere few hours of coding? And with this time saved, we can concentrate on others things (the intent is not to be able to take more vacation ;) ). Pragmatism so.

But fear not, there are things we don't want to do. For example Athena respect strictly the fog of war. I know some games don't, and this surely spare several hundreds of hours, and increase enormously the efficiency of an AI, to be omniscient. But here this would be plain cheating for us.
Image


Hofstadter's Law: "It always takes longer than you expect, even when you take into account Hofstadter's law."

User avatar
Generalisimo
Posts: 4176
Joined: Wed Jun 07, 2006 10:03 pm
Location: Buenos Aires, Argentina
Contact: ICQ WLM

Thu Oct 15, 2009 3:01 pm

Franciscus wrote:OK, I see your point. Nevertheless, let me remind you what I think is usually the point of view of an average buyer of PC games - he does not mind who makes the features of the game he bought - the designer, programmer, coordinator, beta tester, or even the janitor of the company :D - he usually does not even care about the features, but only about the end result. And this is what he gets in the game he buys and in the patches that sometimes (not allways) he searches and downloads from a web page.
Whatever is on the official game at release time or in a patch, is not a mod, and ultimately is the responsability of the game devs, regardless of who makes what feature/script/routine/graphic/pic, whatever. Ideally also, most of the fundamental features should be "in" at release time, and the usual "work in progress" status of our games should be kept at a minimum.
Mods, however, are very different - they are made voluntarily by one or more individuals, are not "official", and are used by a very small fraction of the actual players.

So, the existence of hipothetical scripts like the Boston evacuation one in future Ageod games is of course a decision of Ageod. If they are contemplated in the "official" game, I do not really care who makes them. Of course, I will also be happy to see them as a mod, but it is not the same thing, and most of the game buyers will not benefit from them.

Regards.

I will not say we disagree...
I was just explaining to you what Pocus meant with "I will leave that to modders"... nothing more. ;) :D
"History is the version of past events that people have decided to agree upon."
Napoleon Bonaparte




BOA-AAR: ¡Abajo el imperialismo Británico! (en español)



AGEOD Facebook Fanpage - news & screenshots about the upcoming games!

User avatar
Franciscus
Posts: 4571
Joined: Fri Apr 20, 2007 8:31 pm
Location: Portugal

Thu Oct 15, 2009 4:35 pm

I am glad we all agree :thumbsup: :coeurs: :coeurs:

And now, finally let me give my opinion about the "generic helpers"

Pocus wrote:- generic helper, done by code (logistical helpers like placing a depot)



I am completely in favor of anything that Pocus thinks will help the AI, with some caveats, that for the most part have already been adressed by him (fog of war, etc) or Generalissimo (special rules used by the AI should be "invisible" to the player).
Regarding for instance depots and supply - it would definitely be much better if some mechanism exists that permits the AI to "inteligently" place depots to keep her forces supplied - even if using "special" easy rules unavailable to the player - than see allways supplied AI armies, immune to attrition.

So, to sum it up: Twist the rules as much as you like Pocus - but don't let us notice it too much ;) - and this would, at least to me, be perfectly acceptable and desirable in a "Normal" AI setting (difficult and very difficult levels could include supply, combat, movement, production, etc, bonus for the AI, and even no fog of war - nevertheless I tend to play always in "normal" ;) )

User avatar
McNaughton
Posts: 2766
Joined: Wed Mar 21, 2007 8:47 pm
Location: Toronto, Canada

Thu Oct 15, 2009 9:06 pm

Acceptable advantages, in my opinion, would be to force/encourage the player to play the same rules as the AI.

If a player strips his garrisions, and throws 100% force in the main front lines (something they cannot do in a player vs. player game, or end up being out-flanked), then they should be 'punished' for it. This is one of the biggest abuses by players in games vs. the AI. When players first start any game, for the first time, they tend to play cautiously, as they do not know the patterns of the AI.

However, given time, they know what the AI will and will not do, therefore will not garrison places that they know the AI will never attack (i.e., in AACW no need to garrison Charleston as no amphibious assault is goign to happen with the Union AI, therefore the South devotes all mobile forces to the front lines, while the AI still garrison's its rear areas).

This means the AI is playing on a different rule set than the player, and the player constantly abuses this. However, if there are appropriate punishments, then players will no longer abuse things (as punishments in place in AACW for promoting people outside of the proper seniority)

Punishments could be...

A) If a major coastal city and region has a weak garrison, and the enemy controls the sealanes, chances are an event triggering an 'amphibious landing' happens.

B) On a much more fairer note, if you fail to garrison your rear-areas, as well as other 'unlikely' avenues, events representing population fears (you are not protecting us properly, national morale declines, desertion increases)

C) Small-scale local operations could result in liberation of captured territories if nor properly garrisoned. When Sherman was making his long march, as he advanced further and further, his force became smaller and smaller, as he had to protect his line of communications, primarily from small local forces. However, players tend to capture a city, and move on with their entire force, leaving only token, if any, force to garrison.

If you do not properly garrison population centres, or major supply lines, then I believe it would be fully appropriate (moreso than option A) to have local partisans triggered to liberate population centres, and cut vulnerable supply lines.

These 'punishments' for abusive human actions (not properly garrisoning) will limit the ability of players to overly concentrate their forces to deal smashing blows on the AI's much smaller force (given the AI tends to properly garrison).

So, instead of focussing on AI bonus', difficulty levels could reflect harsher requirements for the player (easy will have no garrison requirements, while hard games will have severe consquences for poor garrisoning).

User avatar
W.Barksdale
AGEod Grognard
Posts: 916
Joined: Tue Feb 05, 2008 8:17 pm
Location: UK

What!?

Thu Oct 15, 2009 9:45 pm

McNaughton wrote:When Sherman was making his long march, as he advanced further and further, his force became smaller and smaller, as he had to protect his line of communications, primarily from small local forces.


The march to the sea? Are you sure about that? :p apy:
"Tell General Lee that if he wants a bridge of dead Yankees I can furnish him with one."

-General William Barksdale at Fredericksburg

User avatar
arsan
Posts: 6244
Joined: Tue Nov 28, 2006 6:35 pm
Location: Madrid, Spain

Thu Oct 15, 2009 10:14 pm

W.Barksdale wrote:The march to the sea? Are you sure about that? :p apy:


I guess he means the long way from Chattanooga up to the Atlanta. Before he broke loose from supply lines and marched to the sea :)

User avatar
McNaughton
Posts: 2766
Joined: Wed Mar 21, 2007 8:47 pm
Location: Toronto, Canada

Thu Oct 15, 2009 10:16 pm

W.Barksdale wrote:The march to the sea? Are you sure about that? :p apy:


Yup, here's an unreliable post from the wikipedia, but more reliable sources will back it up...

"Opposing forces
Sherman, commanding the Military Division of the Mississippi, did not employ his entire army group in the campaign. Confederate Lt. Gen. John Bell Hood was threatening Sherman's supply line from Chattanooga, and Sherman detached two armies under Maj. Gen. George H. Thomas to deal with Hood in the Franklin-Nashville Campaign. For the Savannah Campaign, Sherman's remaining force of 62,000 men (55,000 infantry, 5,000 cavalry, and 2,000 artillerymen manning 64 guns) was divided into two columns for the march:"

I believe he started the campaign with around 90 000 men, and having to garrison cities like Atlanta, and his supply lines from raiders and Confederate field armies, he had to reduce his battlefield frontline force. Makes one wonder if Hood had not squandered his force, and Joe Johnston was allowed to remain in command, when Sherman entered Georgia force would be near 1:1.

Even when they "lived off the land", they had to protect communication lines. When a military force advances, inevitably their force gets thinner (same as when a force retreats, it rolls up its own line of communications, adding garrisons to the main combat force).

User avatar
W.Barksdale
AGEod Grognard
Posts: 916
Joined: Tue Feb 05, 2008 8:17 pm
Location: UK

Fri Oct 16, 2009 2:22 pm

The W & A railroad yes during Atlanta campaign...but Sherman did not protect anythign marching to the sea. Sorry I thought that's what you meant by long march. :cool:

ps sorry for hijacking the thread :blink:
"Tell General Lee that if he wants a bridge of dead Yankees I can furnish him with one."

-General William Barksdale at Fredericksburg

User avatar
McNaughton
Posts: 2766
Joined: Wed Mar 21, 2007 8:47 pm
Location: Toronto, Canada

Fri Oct 16, 2009 5:24 pm

W.Barksdale wrote:The W & A railroad yes during Atlanta campaign...but Sherman did not protect anythign marching to the sea. Sorry I thought that's what you meant by long march. :cool:

ps sorry for hijacking the thread :blink:


Anyways, the issue of the matter is that players tend to abuse knowledge of the AI in games, which further hurts the AI. Knowing that the AI will not be a threat on your coastline will allow the player to do things beyond honest capabilities. A 'March to the Sea' as what Sherman did is an extreme oddity (if he was stopped before he could get to Savannah, he would have been in big trouble given you cannot forage for ammunition...

The truth of the matter is, that players need to protect their supply lines, and rear areas, for threat of the possibility as well as the reality.

tagwyn
AGEod Guard of Honor
Posts: 1220
Joined: Tue Apr 17, 2007 4:09 pm

Fri Oct 16, 2009 7:14 pm

Pocus: I opine: beating children is NEVER a consideration of any kind. t

tagwyn
AGEod Guard of Honor
Posts: 1220
Joined: Tue Apr 17, 2007 4:09 pm

Fri Oct 16, 2009 7:17 pm

I know "nufin" about "nufin!" with respect to computer design. t

Return to “General discussions”

Who is online

Users browsing this forum: No registered users and 22 guests