User avatar
Bohémond
Posts: 2379
Joined: Wed May 14, 2008 8:47 pm

Tue Jan 11, 2011 3:35 am

Hello,

Please find enclose ;

NCP Events RevD zip file in Subfolder Structure
NCP Events xls RevD zip file with revised xls files.
Readme NCP Events Issues RevD txt file

The events for Prussian Campaign scenario have been revised.

The revision B is included.

Regards

Bohémond
Attachments

[The extension txt has been deactivated and can no longer be displayed.]

NCP Events xls RevD.zip
(2.55 MiB) Downloaded 103 times
NCP Events RevD.zip
(448.02 KiB) Downloaded 132 times

User avatar
Gray_Lensman
Posts: 497
Joined: Mon Jun 18, 2007 4:04 am
Location: Who is John Galt?

Tue Jan 11, 2011 3:51 am

deleted

User avatar
Bohémond
Posts: 2379
Joined: Wed May 14, 2008 8:47 pm

Tue Jan 11, 2011 4:35 am

Gray ;

No problem, I keep a track of everything, daily incremental and automatic backup in a external Hard Drive is my friend.


Regards

Bohémond

User avatar
Pocus
Posts: 24759
Joined: Wed Oct 19, 2005 7:37 am
Location: Lyon (France)

Tue Jan 11, 2011 6:11 pm

Bohémond: Do you plan to continue modding (shall I say improving) NCP for long?
Image


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

User avatar
PhilThib
Posts: 13442
Joined: Tue Oct 18, 2005 5:21 pm
Location: Meylan (France)

Tue Jan 11, 2011 6:59 pm

Not that we want you to stop, on the contrary ;) , but just to evaluate the time when we could make an official patch with all your excellent work :coeurs: :cool:
Image

User avatar
Bohémond
Posts: 2379
Joined: Wed May 14, 2008 8:47 pm

Tue Jan 11, 2011 7:31 pm

Hello,

I dont going to ''work'' for a while on Events although I got some preliminary stuff done on 1806 and 1807 sct done.

I will continue to ''work'' on the Units and Models Database issues until the end of the month, including testing.

New parameters, for NCP, like TargetType give a very good result. I want include also the AIaffinity parameter to check if it improves AI.

So let's say end of the month. After this date I will have a ''choix cornélien'' to do ; continue to fix issues in Database, Setup and Events or to redo a whole DB based on template models, so regenerate scenario, etc... . It will depend how new features ( Diplomaticrule, GenMCEvent and others) are nclude in next NCP patch or not.

Let me know

Regards

Bohémond.

User avatar
Ebbingford
Posts: 6059
Joined: Sun Oct 14, 2007 5:22 pm
Location: England

Tue Jan 11, 2011 7:39 pm

Would it be possible to add HQ replacements please. :thumbsup:
I would love to have another bash at the Peninsular scenario, but playing with hardened attrition on for the human only the HQs will vanish. :(
See this thread here..http://www.ageod-forum.com/showthread.php?t=10672

User avatar
Bohémond
Posts: 2379
Joined: Wed May 14, 2008 8:47 pm

Tue Jan 11, 2011 9:28 pm

Ebbingford,

Not before February.
Histoptions.sct ( Buying Replacement by ledger) have some issues to be fixed and need some improvements , like HQ replacements, so it will be done.

And the other hand Peninsular War sceanrio is the hardest scenario to test and need a lot of modding ( DB, events and setup), so maybe it will be the last one to be improved.

Regards

Bohémond

User avatar
Pocus
Posts: 24759
Joined: Wed Oct 19, 2005 7:37 am
Location: Lyon (France)

Wed Jan 12, 2011 4:06 pm

I really advise to get read of HQ sub units, as a mandatory sub unit to create HQ. Without it, the AI is much more efficient.
Image


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

User avatar
Bohémond
Posts: 2379
Joined: Wed May 14, 2008 8:47 pm

Wed Jan 12, 2011 4:13 pm

Pocus,

Sorry but I do not understand the sentence, my english is no so good.

En français , ça donne quoi ?

Merci

Bohémond

User avatar
Bohémond
Posts: 2379
Joined: Wed May 14, 2008 8:47 pm

Wed Jan 12, 2011 4:47 pm

Pocus,

Do you mean ''get rid of'' ?

Like in Rop no more HQ units needed to form an army ?

Regards

Bohémond

User avatar
Pocus
Posts: 24759
Joined: Wed Oct 19, 2005 7:37 am
Location: Lyon (France)

Wed Jan 12, 2011 5:00 pm

Indeed.
Image


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

User avatar
Bohémond
Posts: 2379
Joined: Wed May 14, 2008 8:47 pm

Wed Jan 12, 2011 5:04 pm

Ok,

I'm going to look by myself how to do that, if I don't manage to do it, I will ask you.

Regards

Bohémond

User avatar
Ebbingford
Posts: 6059
Joined: Sun Oct 14, 2007 5:22 pm
Location: England

Wed Jan 12, 2011 8:37 pm

Sounds good :thumbsup:

User avatar
Bohémond
Posts: 2379
Joined: Wed May 14, 2008 8:47 pm

Thu Jan 13, 2011 5:34 pm

Hello,

I have search how to form a army in NCP without HQ unit like in ROP. I do not find the way.

The specorder is the same (#16), nothing in settings, nothing different in Parameter for models or units . I think it's hardcored.

Pocus, I'm wrong ?

Regards

Bohémond

User avatar
PhilThib
Posts: 13442
Joined: Tue Oct 18, 2005 5:21 pm
Location: Meylan (France)

Thu Jan 13, 2011 6:51 pm

You are right, it's hard coded...when Pocus has time, some of those parameters will be exported into game rules files :)

One thing someone did (a mod I beleive) was to include all NCP data inside ROP, which allows for this.... ;)
Image

User avatar
Bohémond
Posts: 2379
Joined: Wed May 14, 2008 8:47 pm

Thu Jan 13, 2011 7:16 pm

I need to rework the models and units DB, based on template UID and new parameters , before including all NCP data in ROP mod, that why this mod is on hold for now.

Due to my very poor graphic modding skills, I don't know if final result will looks good.


So let's wait for new rule export for Army without HQ.

Thank you for all your support and advices.

Regards

Bohémond

User avatar
Pocus
Posts: 24759
Joined: Wed Oct 19, 2005 7:37 am
Location: Lyon (France)

Fri Jan 14, 2011 1:59 pm

Hommes de peu de foi, ce n'est pas en dur!

You must specify this line in gamerules.opt

rulCommandChain = 1
Image


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

User avatar
Bohémond
Posts: 2379
Joined: Wed May 14, 2008 8:47 pm

Fri Jan 14, 2011 2:46 pm

Thank you,

Moi j'ai la foi et c'est pourquoi je demande ;

Is there a ''complete'' list of applicable rules ( with some comments'' and can we got it ?

Regards

Bohémond

User avatar
Pocus
Posts: 24759
Joined: Wed Oct 19, 2005 7:37 am
Location: Lyon (France)

Fri Jan 14, 2011 5:35 pm

Not really, because this is too dangerous to mod them 'just to see how it works'. I prefer that someone ask 'I would like to have this rule works as in this other game', and then tell him how to do that. Also, even if the Wiki is often updated, it is not up to date on everything.
Image


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

User avatar
Bohémond
Posts: 2379
Joined: Wed May 14, 2008 8:47 pm

rulCommandChain = 1

Mon Jan 17, 2011 2:15 pm

Hello,

I've done several tests, AI vs AI, this week end.

I think NCP need a lot a DB reworking ( less 3 stars leaders) or setup modifications (SetFormedCmdMax) to work properly and help AI.

So I will not introduce it on my ongoing modifications.

Regards

Bohémond

Omnius
Lieutenant Colonel
Posts: 290
Joined: Tue Jul 03, 2007 2:15 pm
Location: Salinas, CA

1805 Thames or Danube Event Problem

Mon Jan 17, 2011 4:41 pm

Bohemond,
In the 1805 Thames or Danube scenario I found an event problem. When the Prussians Declare War on France the event message that shows up like a newspaper article only shows the actual filename and not the contents. It would be neat to see the actual message. You've made so many corrections not sure if this was already done.

Also when the Prussians declare war I see the message that the Coalition gains 5 NM and have units unlocked only on the French side, the Coalition player doesn't see this and should. When the Prussian Corps unlock the Coalition player only sees the messages that they unlock and that may be correct since the French probably didn't get to know such things so quickly.

Omnius
Lieutenant Colonel
Posts: 290
Joined: Tue Jul 03, 2007 2:15 pm
Location: Salinas, CA

Keep on Correcting

Mon Jan 17, 2011 4:46 pm

Pocus and PhilThib,
As a recent customer who remembers the message about the end of November about how 1.07rc3 should be finalized in a week or so I say let Bohemond keep on correcting. While I am really anxious to have 1.07rc4 I would rather wait knowing that in waiting Bohemond will make NCP a better product with his many database and event corrections. So I'm all for waiting for a new patch. :thumbsup:

Omnius
Lieutenant Colonel
Posts: 290
Joined: Tue Jul 03, 2007 2:15 pm
Location: Salinas, CA

Army HQ's

Mon Jan 17, 2011 4:58 pm

I understand Ebbingford's angst about HQ elements disappearing because there are no army element replacements. I haven't played any of the really long scenarios like the Spanish campaign and can only imagine how difficult it would be to keep army HQ elements alive for 295 turns. Same goes for supply wagon replacements, so few supply replacements yet supply wagons really get trashed when moving around in inclement weather. Once again I shudder to think about playing the 295 turn Spanish campaign trying to keep my supply wagons alive that long. :wacko:

I really wish the scenario designers had put more thought into matching up replacements to what's being used in a scenario. Too often we have to buy militia when drafting from depots only to watch those militia replacements pile up with no use for them as replacements or garrisons while the line infantry go down the garrison black hole. I really liked how Bohemond changed the 1815 scenario so that French garrisons were made up of militia in France. I'd love to see light artillery replacements get used as well as light cavalry.

Interesting that Army HQ's seem to become history in more recent games and that the AI seems to be baffled by them. Actually the AI seems utterly baffled by leaders as I see the AI break up good formations and create really poor ones. I also hate to see the AI take Admirals out of fleets to use them as ground leaders. Or in the 1805 scenarios the AI has a tendency to take a good Russian leader to command the garrison at Zamosc which is totally out of the way. :bonk:

User avatar
Bohémond
Posts: 2379
Joined: Wed May 14, 2008 8:47 pm

RE 1805 Thames or Danube Event Problem

Mon Jan 17, 2011 5:49 pm

Omnius,

Please see my answer in NCP Database Issue Thread

Regards

Bohémond

User avatar
Pocus
Posts: 24759
Joined: Wed Oct 19, 2005 7:37 am
Location: Lyon (France)

Tue Jan 18, 2011 6:35 pm

The AI makes errors and don't play as a good human player, this is true since the beginning of the engine, but she has improved a lot anyway. You should not look too much in the details when dealing with an AI, but to what she manage to achieve... You have to accept that she will make big and stupid mistakes regularly... As long as you get an enjoyable game, this is ok for me.
Image


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

Omnius
Lieutenant Colonel
Posts: 290
Joined: Tue Jul 03, 2007 2:15 pm
Location: Salinas, CA

Making a Better AI

Wed Jan 19, 2011 4:21 pm

Pocus,
I've always said that AI stands for Artificial Ignorance when it comes to computer wargames. I do see some improvement from BoA though as the NCP AI on medium aggressiveness is far less crazy about sending little units all over the place. Although I've been critical of all wargame AI's from every company for decades I'm actually still amazed that they can make units move around in a somewhat coherent manner and sometimes they actually do make some interesting moves, hence why I like watching them play themselves.

Last night I watched the AI's play the 1813 scenario and saw the French AI make two depots on turn 1. I never would have suspected a supply wagon was sitting inside Hanover except for seeing the AI use it to make a depot. Hence my request that you leave the AI vs AI feature in the final patch for NCP and the other older games you're fixing.

The biggest problem to making a great AI is how do you program in the kind of visuo-spatial awareness we humans have? We look at the game map and can see the relationships between where forces are and where the important VP spots are. We humans can think of a good plan and then as circumstances change we can react to those changes better. Plus we humans can develop long-term plans which the AI really can't. One reason why AI's break down after a few turns and why they really can't handle scenarios like the 295 turn Spanish campaign.

I figure that we'll only ever have a great wargame AI when IBM decides to do unto a wargame what it did to chess with it's Big Blue supercomputer and it's Big Blue chess program where they spent a ton of money programming such a great chess program.

I can see one reason why it's good to use the area map system over a hexagon based on because that means less areas to move to for the AI to calculate.

If there's some things that could be improved it's to make the AI use leaders smarter. Keep Admirals with fleets and Generals with armies. Don't waste away really good corp or division commanders on some out of the way garrison, I see way too many good leaders wasted by the AI on garrisons that aren't important while leaving many ground units leaderless.

The other is that bigger is better, too many times the AI breaks down forces both naval and ground into too many small inconsequestial forces too easily hammered by a bigger force. Sometimes it does do a good job of creating big corps or armies that are quite difficult to defeat. My big leadered fleets make hay against the AI's little leaderless fleets.

Another improvement would be to see the AI do better at recovering cohesion, changing from defense to withdraw to get that extra cohesion recovery boost inside some structure.

I play against the AI when I first get a game and am learning the ropes of the game system and scenario, then I switch to playing both sides solo.

User avatar
Pocus
Posts: 24759
Joined: Wed Oct 19, 2005 7:37 am
Location: Lyon (France)

Wed Jan 19, 2011 5:19 pm

You hit the nail indeed, with "The biggest problem to making a great AI is how do you program in the kind of visuo-spatial awareness we humans have" ... This is indeed the center of the trouble. For me at least. I developed a whole series of 'situational maps' for that, like a threat map, a supply map, an interest map. But I admit that the high level thinking is still not there as I would like... But this can be done, with time, look at the new AI scripting commands used by some in RUS for example.

But, good news... You should see improvements and new optional rules to help the AI in the coming month, because I should be able to spend time on it around March I believe.
Image


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

Omnius
Lieutenant Colonel
Posts: 290
Joined: Tue Jul 03, 2007 2:15 pm
Location: Salinas, CA

Programmer's Holy Grail

Thu Jan 20, 2011 3:37 pm

Pocus,
Yep making a really good AI is the wargame programmer's Holy Grail and so far no one has mastered it. Someday I'd like to be able to turn an AI on to a harder level where it actually plays smarter rather than just having a lot more stuff to throw at us. The AI difficulty buttons always remind me of the old George Carlin joke about the tv brightness button - "I turn the bright button up all the way but the tv never gets any smarter". :wacko:

I'm really glad to hear that you may be able to put some time into improving the AGE engine AI. I know that with a series of games that things improve with each game in that series. Will your AI improvements get into NCP or BoA as well? :cool:

Is there a place where I can get my hands on some kind of manual about the AGE engine scripting language? I'm starting to think that I should learn it since I really like the AGE engine series, despite my complaints, and it would behoove me to learn the scripting language so I can mod scenarios.

Omnius
Lieutenant Colonel
Posts: 290
Joined: Tue Jul 03, 2007 2:15 pm
Location: Salinas, CA

1813 Event Problems

Fri Jan 21, 2011 4:23 pm

Bohemond,
There's a problem on turn 6 of the 1813 scenario where an event triggers but what we see is the filename in the event message and the title of the news article that explains what's up. The event message actually shows uo properly.

Also in the same 1813 scenario there are 3 supply wagons that come on for the Coalition. One is a Prussian supply wagon that shows up at Koenigsberg but the event message incorrectly labels it an Austrian supply instead of Prussian.

Return to “Help to improve NCP!”

Who is online

Users browsing this forum: No registered users and 1 guest