Matrix Games Forums

Forums  Register  Login  Photo Gallery  Member List  Search  Calendars  FAQ 

My Profile  Inbox  Address Book  My Subscription  My Forums  Log Out

RE: AI for MWIF - Norway

 
View related threads: (in this forum | in all forums)

Logged in as: Guest
Users viewing this topic: none
  Printable Version
All Forums >> [New Releases from Matrix Games] >> World in Flames >> AI Opponent Discussion >> RE: AI for MWIF - Norway Page: <<   < prev  1 [2] 3 4   next >   >>
Login
Message << Older Topic   Newer Topic >>
RE: AI for MWIF - Norway - 8/21/2008 4:49:12 AM   
paulderynck


Posts: 8201
Joined: 3/24/2007
From: Canada
Status: offline

quote:

ORIGINAL: Shannon V. OKeets


quote:

ORIGINAL: Froonp

Here is my example :

By RAW, Istanbul is not invadable from the East Med (it has no all sea hexside that touches the East Med).
By the MWiF logic, it is (It has an all sea hexside, and the hex is adjacent to the East Med).




Yes.

But in my opinion RAW is very strange here.

It says Istanbul is a port on the East Med but not invadable from the East Med. I can understand when London can not be invaded because the long estuary enables the defenders to stop an invasion fleet far from the hex that London occupies. That doesn't appear to be the case with Istanbul.

I guess my main question is "What is being simulated here by placing the sea area boundary so carefully?" Assuming that this was done intentionally, is it worth the effort to duplicate in MWIF?

IMO it IS important to make it the same as WiFFE. Isn't that one of the "guiding principles" of this undertaking? Especially a hex as important as Istanbul. Certainly I understand the differences in the Pacific (or any other different scale map) as that decision was made a long time ago and had some real attraction in eliminating the scale difference. But just as in the war in China, we all have much to learn about how similar or how different that part of MWiF will play compared to WiFFE. I would like to see the European map work the same way as WiFFE, though.

BTW the Dardanelles are a very big reason for not making Istanbul invadeable from the Eastern Med - hence the Gallipoli campaign in WWI.

_____________________________

Paul

(in reply to Shannon V. OKeets)
Post #: 31
RE: AI for MWIF - Norway - 8/21/2008 5:09:01 AM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: paulderynck


quote:

ORIGINAL: Shannon V. OKeets


quote:

ORIGINAL: Froonp

Here is my example :

By RAW, Istanbul is not invadable from the East Med (it has no all sea hexside that touches the East Med).
By the MWiF logic, it is (It has an all sea hexside, and the hex is adjacent to the East Med).




Yes.

But in my opinion RAW is very strange here.

It says Istanbul is a port on the East Med but not invadable from the East Med. I can understand when London can not be invaded because the long estuary enables the defenders to stop an invasion fleet far from the hex that London occupies. That doesn't appear to be the case with Istanbul.

I guess my main question is "What is being simulated here by placing the sea area boundary so carefully?" Assuming that this was done intentionally, is it worth the effort to duplicate in MWIF?

IMO it IS important to make it the same as WiFFE. Isn't that one of the "guiding principles" of this undertaking? Especially a hex as important as Istanbul. Certainly I understand the differences in the Pacific (or any other different scale map) as that decision was made a long time ago and had some real attraction in eliminating the scale difference. But just as in the war in China, we all have much to learn about how similar or how different that part of MWiF will play compared to WiFFE. I would like to see the European map work the same way as WiFFE, though.

BTW the Dardanelles are a very big reason for not making Istanbul invadeable from the Eastern Med - hence the Gallipoli campaign in WWI.

But you can invade the hex SW of Istanbul, which, being on the Sea of Marmara, is far past the Dardenelles.

I go back to my question of what is being simulated.

To my eye this looks like a rule/exception that affects 1 hex out of 70,200. Or 1 out of 5,000+ if you want to just count coastal hexes. Yet another case where we tell the new player: "Oh, and by the way, for this hex a special rule aplpies."


_____________________________

Steve

Perfection is an elusive goal.

(in reply to paulderynck)
Post #: 32
RE: AI for MWIF - Norway - 8/21/2008 5:53:46 AM   
paulderynck


Posts: 8201
Joined: 3/24/2007
From: Canada
Status: offline

quote:

ORIGINAL: Shannon V. OKeets


quote:

ORIGINAL: paulderynck


quote:

ORIGINAL: Shannon V. OKeets


quote:

ORIGINAL: Froonp

Here is my example :

By RAW, Istanbul is not invadable from the East Med (it has no all sea hexside that touches the East Med).
By the MWiF logic, it is (It has an all sea hexside, and the hex is adjacent to the East Med).




Yes.

But in my opinion RAW is very strange here.

It says Istanbul is a port on the East Med but not invadable from the East Med. I can understand when London can not be invaded because the long estuary enables the defenders to stop an invasion fleet far from the hex that London occupies. That doesn't appear to be the case with Istanbul.

I guess my main question is "What is being simulated here by placing the sea area boundary so carefully?" Assuming that this was done intentionally, is it worth the effort to duplicate in MWIF?

IMO it IS important to make it the same as WiFFE. Isn't that one of the "guiding principles" of this undertaking? Especially a hex as important as Istanbul. Certainly I understand the differences in the Pacific (or any other different scale map) as that decision was made a long time ago and had some real attraction in eliminating the scale difference. But just as in the war in China, we all have much to learn about how similar or how different that part of MWiF will play compared to WiFFE. I would like to see the European map work the same way as WiFFE, though.

BTW the Dardanelles are a very big reason for not making Istanbul invadeable from the Eastern Med - hence the Gallipoli campaign in WWI.

But you can invade the hex SW of Istanbul, which, being on the Sea of Marmara, is far past the Dardenelles.

I go back to my question of what is being simulated.

To my eye this looks like a rule/exception that affects 1 hex out of 70,200. Or 1 out of 5,000+ if you want to just count coastal hexes. Yet another case where we tell the new player: "Oh, and by the way, for this hex a special rule aplpies."


In the Q&A recently completed, Harry (prior to July 4th) stated that Istanbul cannot be invaded from the Eastern Med because there "is no all sea hexside to invade across". So in fact the way it is drawn both in WiFFE and in MWiF meets this condition and should prohibit an invasion from the Eastern Med. Meanwhile on the previous page of this thread, Patrice posted two excellent examples of hexes that are invadeable from two sea zones in WiFFE - which are drawn in MWiF to my eye as invadeable from only one sea zone. (Kristiansand and Liverpool)

So are we confusing the new players, the old players, or everyone?

The definition in RAW for being able to invade has always contained problematic wording - hence the question in the Q&A with 5 sub-parts, but essentially you need to have at least a portion of an all-sea hexside that is "traversable" from the sea zone in which the invasion fleet resides - to qualify as invadeable. As you pointed out, the computer does not "see" the map` - but unfortunately for this purpose, we humans rely on seeing the map.

The bigger question is not "can this specific hex be invaded?" but rather: "how does a player of MWiF know it can or it can't"?

The key hexes are the ones where the sea area boundaries meet (maybe 100 on the European map) and the problem needs to be solved both programatically and visually. If we were starting from scratch I'd suggest a special look-up table for the former and some form of visual indicator for the latter.


_____________________________

Paul

(in reply to Shannon V. OKeets)
Post #: 33
RE: AI for MWIF - Norway - 8/21/2008 6:47:48 AM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline
Paul,

Lookup table - arrrgh. I use them (often) but only as a last resort. A system is better for all concerned.

There already is the ability for the player to see if a hex can be invaded, as I defined it earlier. And placing the cursor over any coastal hex lets you know to which sea areas it is adjacent.

I have on my to-do list to somehow visually toggle on/off invadable hexes from a sea area. The design I have laid out is for the player to right click on, say, the North Sea, and select Invasion Hexes from a popup menu. The coastal hexes that can be invaded from the North Sea would then be highlighted.

I haven't decided on how to highlight them - and I want this to be very temporary, with it being toggled off as soon as the player does something else. My concern is the complexity/density of information in each hex already. I do not want to add to something 'permanent' to that. This is a companion decision to highlighting hexes to which a (land) unit can move. I would use the same 'highlight'.

=====
As an argument in your favor for treating Istanbul as a special hex, you might look at Manila (I think it is Manila). There is a problem with the hex being adjacent to two sea areas but the port only being accessible from one of them. There might be a couple of other hexes with similar problems (northern Japan, Rabual)? I have already resigned myself to having to do something unusual to address those unusual circumstances. If we have a list of 'weird' ports I wouldn't be so upset about writing 'exceptions' code for Istanbul. Not to say I would be happy, just less unhappy.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to paulderynck)
Post #: 34
RE: AI for MWIF - Norway - 8/21/2008 7:31:37 AM   
paulderynck


Posts: 8201
Joined: 3/24/2007
From: Canada
Status: offline
quote:

ORIGINAL: Shannon V. OKeets
There already is the ability for the player to see if a hex can be invaded, as I defined it earlier. And placing the cursor over any coastal hex lets you know to which sea areas it is adjacent.
I

Sorry I've been all through the thread a couple times again and I can't see to what you refer. How do I see if a hex can be invaded currently? I could only find the description of the program's methodology. And does adjacency to two sea areas of an invadeable hex automatically mean I can invade from either? If so, this is a departure from WiFFE and RAW, possibly in a lot of places.

_____________________________

Paul

(in reply to Shannon V. OKeets)
Post #: 35
RE: AI for MWIF - Norway - 8/21/2008 7:46:06 AM   
paulderynck


Posts: 8201
Joined: 3/24/2007
From: Canada
Status: offline
quote:

ORIGINAL: Shannon V. OKeets

As an argument in your favor for treating Istanbul as a special hex, you might look at Manila (I think it is Manila). There is a problem with the hex being adjacent to two sea areas but the port only being accessible from one of them.

In WiFFE, Manila is invadeable from the South China Sea only. In MWiF, to me it is not invadeable at all, if I apply RAW (no full sea hexsides to cross from either sea zone). But I have no issue with that because of the change of scale.




Attachment (1)

_____________________________

Paul

(in reply to Shannon V. OKeets)
Post #: 36
RE: AI for MWIF - Norway - 8/21/2008 8:43:02 AM   
Froonp


Posts: 7995
Joined: 10/21/2003
From: Marseilles, France
Status: offline
quote:

ORIGINAL: paulderynck
The key hexes are the ones where the sea area boundaries meet (maybe 100 on the European map) and the problem needs to be solved both programatically and visually. If we were starting from scratch I'd suggest a special look-up table for the former and some form of visual indicator for the latter.

Here are all these hexes on the Euroean map.
Doing this revealed to me that 2 of them were worng (one in Corsica and one in Greece), and I'll modify them today, but also revealed to me that the case of Istanbul seems to be unique on the European Map.




Attachment (1)

(in reply to paulderynck)
Post #: 37
RE: AI for MWIF - Norway - 8/21/2008 8:45:54 AM   
Froonp


Posts: 7995
Joined: 10/21/2003
From: Marseilles, France
Status: offline
quote:

ORIGINAL: Froonp

quote:

ORIGINAL: paulderynck
The key hexes are the ones where the sea area boundaries meet (maybe 100 on the European map) and the problem needs to be solved both programatically and visually. If we were starting from scratch I'd suggest a special look-up table for the former and some form of visual indicator for the latter.

Here are all these hexes on the Euroean map.
Doing this revealed to me that 2 of them were worng (one in Corsica and one in Greece), and I'll modify them today, but also revealed to me that the case of Istanbul seems to be unique on the European Map.




There is the problem of :

Tripoli
Liverpool
Kristiansand
Narvik
Plymouth

Looking like they can only be invaded from 1 sea area when the WiF FE map show that they can be invaded from 2.

(in reply to Froonp)
Post #: 38
RE: AI for MWIF - Norway - 8/21/2008 8:47:40 AM   
Froonp


Posts: 7995
Joined: 10/21/2003
From: Marseilles, France
Status: offline
Here are those hexes for Africa.




Attachment (1)

(in reply to Froonp)
Post #: 39
RE: AI for MWIF - Norway - 8/21/2008 8:48:18 AM   
Froonp


Posts: 7995
Joined: 10/21/2003
From: Marseilles, France
Status: offline
Here are those hexes for Asia :




Attachment (1)

(in reply to Froonp)
Post #: 40
RE: AI for MWIF - Norway - 8/21/2008 9:10:28 AM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: paulderynck

quote:

ORIGINAL: Shannon V. OKeets

As an argument in your favor for treating Istanbul as a special hex, you might look at Manila (I think it is Manila). There is a problem with the hex being adjacent to two sea areas but the port only being accessible from one of them.

In WiFFE, Manila is invadeable from the South China Sea only. In MWiF, to me it is not invadeable at all, if I apply RAW (no full sea hexsides to cross from either sea zone). But I have no issue with that because of the change of scale.




It was Legaspi that I was concerned about.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to paulderynck)
Post #: 41
RE: AI for MWIF - Norway - 8/21/2008 9:21:00 AM   
paulderynck


Posts: 8201
Joined: 3/24/2007
From: Canada
Status: offline
Not a major issue for Africa and Asia due to the scale change IMO. The one hex in Pakistan that is invadeable only from the Persian Gulf in WiFFE is probably replaced by two hexes with the further east one invadeable from both sea zones.


_____________________________

Paul

(in reply to Froonp)
Post #: 42
RE: AI for MWIF - Norway - 8/21/2008 9:29:32 AM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: Froonp

Here are those hexes for Asia :




Patrice,

Sorry, but I can't figure out what are problems and what aren't problems from all these pictures.

---
Paul,

WIFFE players see an all sea hexside bisected by a sea area boundary and say it is 'obvious' that the hex can be invaded from both sea areas. I look at it and see that only half of an all sea hexside is present for each sea area, so neither sea area can invade the hex: because there isn't a 'full' all sea hexside bordering, say, Liverpool, for either of the 2 sea areas.

My point is that what you say is visually clear on the WIF FE maps is just something you have learned to see as an invadable hex. The maps could have been drawn as in MWIF and you would have learned a different pattern as 'invadable': there is an all-sea hexside present and the hex borders two sea areas => invadable.

I intend to use the style developed by Chris for CWIF in MWIF. It is less dependent on the pen of the artist. The WIF FE map sea area boundaries were drawn using a straight line for the most part. You can see the problems this creates in Patrice's examples at Lisbon, Malta, and La Spezia. The line's relationship to the sea area hex dot can have a razor thin separation. At 8 levels of zoom, that isn't going to work for MWIF.

Now this doesn't achieve the same result for Istanbul (comparing WIF FE to MWIF). Are there any other places?

_____________________________

Steve

Perfection is an elusive goal.

(in reply to Froonp)
Post #: 43
RE: AI for MWIF - Norway - 8/21/2008 9:31:07 AM   
paulderynck


Posts: 8201
Joined: 3/24/2007
From: Canada
Status: offline

quote:

ORIGINAL: Shannon V. OKeets


quote:

ORIGINAL: paulderynck

quote:

ORIGINAL: Shannon V. OKeets

As an argument in your favor for treating Istanbul as a special hex, you might look at Manila (I think it is Manila). There is a problem with the hex being adjacent to two sea areas but the port only being accessible from one of them.

In WiFFE, Manila is invadeable from the South China Sea only. In MWiF, to me it is not invadeable at all, if I apply RAW (no full sea hexsides to cross from either sea zone). But I have no issue with that because of the change of scale.




It was Legaspi that I was concerned about.

The Legaspi hex is clearly invadeable from either sea zone while its port no longer has direct access to the South China Sea in MWiF. I accept this as the result of the scale change. At least that is what I see if I apply RAW and the recent Q&A to the picture. I'm not clear on what the accessability of the port has to do with invading? I just assume that if you successfully invade from the South China Sea, then you crossed overland to take possession of the port versus having assaulted it on the harbour side.

_____________________________

Paul

(in reply to Shannon V. OKeets)
Post #: 44
RE: AI for MWIF - Norway - 8/21/2008 9:45:36 AM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: paulderynck


quote:

ORIGINAL: Shannon V. OKeets


quote:

ORIGINAL: paulderynck

quote:

ORIGINAL: Shannon V. OKeets

As an argument in your favor for treating Istanbul as a special hex, you might look at Manila (I think it is Manila). There is a problem with the hex being adjacent to two sea areas but the port only being accessible from one of them.

In WiFFE, Manila is invadeable from the South China Sea only. In MWiF, to me it is not invadeable at all, if I apply RAW (no full sea hexsides to cross from either sea zone). But I have no issue with that because of the change of scale.




It was Legaspi that I was concerned about.

The Legaspi hex is clearly invadeable from either sea zone while its port no longer has direct access to the South China Sea in MWiF. I accept this as the result of the scale change. At least that is what I see if I apply RAW and the recent Q&A to the picture. I'm not clear on what the accessability of the port has to do with invading? I just assume that if you successfully invade from the South China Sea, then you crossed overland to take possession of the port versus having assaulted it on the harbour side.

The problem is that the fundamental data is the hex's relationship to sea areas, with the relationship of a port to sea areas being one and the same as the relationship of the hex to the sea areas. So you could move a naval unit into/out of Legaspi from either sea area. That is how the code is written for all ports on the map.

To rewrite the code so the relationship of a port to sea areas is independent of the realtionship of the hex it sits in to those same sea areas would be a lot of work (vast understatement). Instead, I expect to have to have special code that says: "if the port we are examining is Legaspi, then different rules apply, ...". What I want to know is are there other ports for which I have to write special code?

=======

Writing code is non-trivial. As an example, here is one of the 3 main routines for deciding about cooperation between countries. All the comments were added by me.
// ****************************************************************************
function TMinorCountry.CooperatesWith(Cnt: TCountryID): Boolean;
// ****************************************************************************
// CooperatesWith returns whether Cnt cooperates with Self (a minor country).
// This routine is for countries, but in actual use it checks for whether units
// can cooperate.  Since cooperation is always based on country of origin and
// the status of the country (conquered, liberated, etc.), this routine is here
// rather than in WIFUnits.
// ****************************************************************************
var
  CMC: TMajorCountry;
  CountryIn: TCountryID;
begin
  CountryIn := Cnt; // So variable is not changed.
// ****************************************************************************
// For CW member countries, check if CW (major power) cooperates with CountryIn.
// ****************************************************************************
  if (ID in CommonwealthCountries) or (ID = Commonwealth.ID) then
    Result := Commonwealth.CooperatesWith(CountryIn)
// ****************************************************************************
// If checking whether Self cooperates with AOI, reverse the call, to check if
// AOI, (a subcountry), cooperates with Self. 
// ****************************************************************************
  else if CountryIn = AOI.ID then Result := AOI.CooperatesWith(Self.ID)
  else
  begin
// ****************************************************************************
// Commonwealth countries are treated as a one country for cooperation purposes.
// Communist China is distinct when determining cooperation.  Some units belong
// to subcountries and have IDs for that subcountry.  So, HomeCountry converts
// the subcountry IDs into a minor or major country ID.
// ****************************************************************************
    if CountryIn in CommonwealthCountries then
      CountryIn := Commonwealth.ID
    else if CountryIn <> CommunistChina.ID then
      CountryIn := Countries[CountryIn].HomeCountry.ID;
// ****************************************************************************
// In Barbarossa, Italian units are controlled by Germany.  The conversion to
// CMC is added code to handle that unique condition.
// ****************************************************************************
    if (CurrScenario = scBarbarossa) and (Self = Italy) then CMC := Italy
    else CMC := Self.ControllingMajorCountry;
// ****************************************************************************
// We are now ready to check for cooperation.
// ****************************************************************************
   	Result := (ID = CountryIn) or                        // From same country
              (Parent = Countries[CountryIn].Parent) or  // With same parent
              (CMC = Countries[CountryIn]) or     // With controling major power
              ((Self = France) and                // Free France with CW & USA
               VichyFrance.Legal and
               ((CountryIn = UnitedStates.ID) or
                (CountryIn = Commonwealth.ID)));
  end;
end;


_____________________________

Steve

Perfection is an elusive goal.

(in reply to paulderynck)
Post #: 45
RE: AI for MWIF - Norway - 8/21/2008 9:49:49 AM   
paulderynck


Posts: 8201
Joined: 3/24/2007
From: Canada
Status: offline
quote:

ORIGINAL: Shannon V. OKeets

Paul,

WIFFE players see an all sea hexside bisected by a sea area boundary and say it is 'obvious' that the hex can be invaded from both sea areas. I look at it and see that only half of an all sea hexside is present for each sea area, so neither sea area can invade the hex: because there isn't a 'full' all sea hexside bordering, say, Liverpool, for either of the 2 sea areas.

My point is that what you say is visually clear on the WIF FE maps is just something you have learned to see as an invadable hex. The maps could have been drawn as in MWIF and you would have learned a different pattern as 'invadable': there is an all-sea hexside present and the hex borders two sea areas => invadable.

I intend to use the style developed by Chris for CWIF in MWIF. It is less dependent on the pen of the artist. The WIF FE map sea area boundaries were drawn using a straight line for the most part. You can see the problems this creates in Patrice's examples at Lisbon, Malta, and La Spezia. The line's relationship to the sea area hex dot can have a razor thin separation. At 8 levels of zoom, that isn't going to work for MWIF.

Now this doesn't achieve the same result for Istanbul (comparing WIF FE to MWIF). Are there any other places?

OK I see your point. The first task is to see how many hexes are like Istanbul. The next is to decide if it is worthwhile making them exceptions (I would like to say that it is). The final task is how to convey to the players, whether old or new, the mindset of viewing the MWiF map in a way that swiftly and accurately imparts the concept of invadeability.

We are in Task 1 and I'd like to help with it. I still think we only need to do this for the part of the map corresponding to the WiFFE European map (other than assuring Task 3 works everywhere). I'll look for any other hexes but I'm inhibited from access to the WiFFE maps unitil we move at the end of next week.

_____________________________

Paul

(in reply to Shannon V. OKeets)
Post #: 46
RE: AI for MWIF - Norway - 8/21/2008 10:04:31 AM   
paulderynck


Posts: 8201
Joined: 3/24/2007
From: Canada
Status: offline
quote:

ORIGINAL: Shannon V. OKeets
The problem is that the fundamental data is the hex's relationship to sea areas, with the relationship of a port to sea areas being one and the same as the relationship of the hex to the sea areas. So you could move a naval unit into/out of Legaspi from either sea area. That is how the code is written for all ports on the map.

To rewrite the code so the relationship of a port to sea areas is independent of the realtionship of the hex it sits in to those same sea areas would be a lot of work (vast understatement). Instead, I expect to have to have special code that says: "if the port we are examining is Legaspi, then different rules apply, ...". What I want to know is are there other ports for which I have to write special code?

Panama would be the first one that comes to mind as it was featured in the Q&A when it was established that "most" ports do in fact have access to all adjacent sea zones in WiFFE. Unfortunately the scale change in the Pacific may have created many more of these exceptions (as is the case for Legaspi).

I'll review the MWiF maps and start making a list (gasp) of any "Legaspi"s I find.

_____________________________

Paul

(in reply to Shannon V. OKeets)
Post #: 47
RE: AI for MWIF - Norway - 8/21/2008 5:26:12 PM   
Froonp


Posts: 7995
Joined: 10/21/2003
From: Marseilles, France
Status: offline
quote:

ORIGINAL: paulderynck
The Legaspi hex is clearly invadeable from either sea zone while its port no longer has direct access to the South China Sea in MWiF.

No, the port still have access to the South China Sea.
It is the same for the Major Port of Auckland (NZ) that has access to 2 Sea Areas while being physicaly placed on only the New Zealand Coast.

Those ports had access to 2 Sea Areas in WiF FE, they had to have access to the same 2 Sea Areas in MWiF.

You just have to use some WiFZen to accept it, imagining that the ships go around the peninsula's tip to go to the other Sea Area.

(in reply to paulderynck)
Post #: 48
RE: AI for MWIF - Norway - 8/21/2008 5:28:27 PM   
Froonp


Posts: 7995
Joined: 10/21/2003
From: Marseilles, France
Status: offline
quote:

ORIGINAL: Shannon V. OKeets
Sorry, but I can't figure out what are problems and what aren't problems from all these pictures.

These are just comparisons of what the Sea Area Boundary connection with the land is in Wif FE and MWiF.
I can send you the full sized picture (it did not fit the forums) so you see it better.
It shows that in 99% of the cases, there are no problems.

Istanbul is the only one I found at the European Scale.

(in reply to Shannon V. OKeets)
Post #: 49
RE: AI for MWIF - Norway - 8/21/2008 5:31:39 PM   
Froonp


Posts: 7995
Joined: 10/21/2003
From: Marseilles, France
Status: offline
quote:

ORIGINAL: paulderynck
Panama would be the first one that comes to mind as it was featured in the Q&A when it was established that "most" ports do in fact have access to all adjacent sea zones in WiFFE. Unfortunately the scale change in the Pacific may have created many more of these exceptions (as is the case for Legaspi).

I'll review the MWiF maps and start making a list (gasp) of any "Legaspi"s I find.

There are none, the CWiF map was faithfull to the WiF FE map in this regards, this is something I have checked some months (years ?) ago when I was making the mockup for the coastlines.

Legaspi is on 2 Sea Areas as I told you. Look at it ingame.

(in reply to paulderynck)
Post #: 50
RE: AI for MWIF - Norway - 8/21/2008 7:16:03 PM   
paulderynck


Posts: 8201
Joined: 3/24/2007
From: Canada
Status: offline

quote:

ORIGINAL: Froonp

quote:

ORIGINAL: paulderynck
Panama would be the first one that comes to mind as it was featured in the Q&A when it was established that "most" ports do in fact have access to all adjacent sea zones in WiFFE. Unfortunately the scale change in the Pacific may have created many more of these exceptions (as is the case for Legaspi).

I'll review the MWiF maps and start making a list (gasp) of any "Legaspi"s I find.

There are none, the CWiF map was faithfull to the WiF FE map in this regards, this is something I have checked some months (years ?) ago when I was making the mockup for the coastlines.

Legaspi is on 2 Sea Areas as I told you. Look at it ingame.

This is really counter-intuitive in view of the discussion about the port of Panama in the recent Q&A. How do I as a player know these are two (or more) sea zone ports?

IMO it would be better to move the port symbol 1 hex to the southeast and have the blue line bisect it or redraw the blue line as below:





Attachment (1)

_____________________________

Paul

(in reply to Froonp)
Post #: 51
RE: AI for MWIF - Norway - 8/21/2008 7:24:41 PM   
Froonp


Posts: 7995
Joined: 10/21/2003
From: Marseilles, France
Status: offline
quote:

ORIGINAL: paulderynck
This is really counter-intuitive in view of the discussion about the port of Panama in the recent Q&A. How do I as a player know these are two (or more) sea zone ports?

You see it in multiple ways, first by looking at the information bar that tells you which Sea Area an hex is adjacent to.

quote:

IMO it would be better to move the port symbol 1 hex to the southeast and have the blue line bisect it or redraw the blue line as below:

I don't want to move it, because it is at the right place (see attached map).
And I can't draw the blue line as you drew it. I don't draw blue lines. I tell the game that Legaspi is on 2 Sea Areas, I tell it which Sea Areas, and I tell it the same for the hex SE of Legaspi. Then the game draws the line it wants to draw.





Attachment (1)

(in reply to paulderynck)
Post #: 52
RE: AI for MWIF - Norway - 8/21/2008 7:31:01 PM   
Froonp


Posts: 7995
Joined: 10/21/2003
From: Marseilles, France
Status: offline
With the blow up effect in the scale of the map on the Pacific, this kind of thing is neither surprising nor bothering me. There is no great continent-like mass of land barring the way to ships from Legaspi (or Auckland, NZ) to go into the South China Sea (or Tasman Sea). The way is just 50 miles away.

There are a score of other abstractions in the Sea Areas on the WiF FE map, this is only one more.
For example, was the the Skaggerat the real bottleneck to enter the Baltic, as the WiF FE map shows ? No, it was the straits around Copenhagen. So why is the Sea Area boundary at the wrong place ? Other example, was the Plymouth-Brest the real Bottleneck to pass the British Channel ? No, it rather was the Dover-Calais area, so why is the Sea ARea boundary at the wrong place.

(in reply to Froonp)
Post #: 53
RE: AI for MWIF - Norway - 8/21/2008 7:33:12 PM   
paulderynck


Posts: 8201
Joined: 3/24/2007
From: Canada
Status: offline
quote:

ORIGINAL: Shannon V. OKeets

I intend to use the style developed by Chris for CWIF in MWIF. It is less dependent on the pen of the artist. The WIF FE map sea area boundaries were drawn using a straight line for the most part. You can see the problems this creates in Patrice's examples at Lisbon, Malta, and La Spezia. The line's relationship to the sea area hex dot can have a razor thin separation. At 8 levels of zoom, that isn't going to work for MWIF.

Now this doesn't achieve the same result for Istanbul (comparing WIF FE to MWIF). Are there any other places?

I've been thinking about this some more since last night, and since there are exceptions to the blue line following the hex border, why can't the blue line be redrawn to fix the impression of non-invadeability for the veteran WiFFErs? After all they will be the corps of the initial buyers and the Newbies have to learn the system regardless.

Is there something about the blue line that affects the program's decision on invadeability? If not then, the lines can be corrected for Tripoli, Liverpool, Kristiansand, and Plymouth. It can be redrawn to make MWiF match WiFFE for Narvik and it is already correct for Istanbul.

quote:

ORIGINAL: Shannon V. OKeets
The problem is that the fundamental data is the hex's relationship to sea areas, with the relationship of a port to sea areas being one and the same as the relationship of the hex to the sea areas. So you could move a naval unit into/out of Legaspi from either sea area. That is how the code is written for all ports on the map.

I hope you aren't saying that the presence of a port is used to determine invadeability somehow? There are many ports that cannot be invaded - Trieste for example.



_____________________________

Paul

(in reply to Shannon V. OKeets)
Post #: 54
RE: AI for MWIF - Norway - 8/21/2008 9:17:18 PM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: paulderynck

quote:

ORIGINAL: Shannon V. OKeets

I intend to use the style developed by Chris for CWIF in MWIF. It is less dependent on the pen of the artist. The WIF FE map sea area boundaries were drawn using a straight line for the most part. You can see the problems this creates in Patrice's examples at Lisbon, Malta, and La Spezia. The line's relationship to the sea area hex dot can have a razor thin separation. At 8 levels of zoom, that isn't going to work for MWIF.

Now this doesn't achieve the same result for Istanbul (comparing WIF FE to MWIF). Are there any other places?

I've been thinking about this some more since last night, and since there are exceptions to the blue line following the hex border, why can't the blue line be redrawn to fix the impression of non-invadeability for the veteran WiFFErs? After all they will be the corps of the initial buyers and the Newbies have to learn the system regardless.

Is there something about the blue line that affects the program's decision on invadeability? If not then, the lines can be corrected for Tripoli, Liverpool, Kristiansand, and Plymouth. It can be redrawn to make MWiF match WiFFE for Narvik and it is already correct for Istanbul.

quote:

ORIGINAL: Shannon V. OKeets
The problem is that the fundamental data is the hex's relationship to sea areas, with the relationship of a port to sea areas being one and the same as the relationship of the hex to the sea areas. So you could move a naval unit into/out of Legaspi from either sea area. That is how the code is written for all ports on the map.

I hope you aren't saying that the presence of a port is used to determine invadeability somehow? There are many ports that cannot be invaded - Trieste for example.



As Patrice, described, the program draws the sea area boundaries (calling them blue lines can be misleading) based on the data. It was a real nightmare to get that code correct: 60-80 hours. I had tried to straighten it out in 2005 but I didn't understand enough about the CWIF code to figure out what was/should be coded. It was only earlier this year (2008) that I revisited that code and was able to make it do what I wanted it to do. Legaspi, the northern Japanese port, and the islands around Rabual were the hardest to get right.

At no time does the program know what the land mass outlines/graphics look like. It only knows that an island is an island if the data says there are all sea hexes/all sea hexsides on all 6 sides of a hex. Add data entries that Moscow is surrounded by all sea hexsides, and the program will consider Moscow an island, regardless of what the graphics look like.

Ports are data that provide added functionality to a hex. A hex's relationship to sea areas is data driven. If the data says a port is in a hex, then the hex recieves the 'benefits' of being a port. The same is true of cities, factories, resources, etc.

Ports have zero interaction with invasions - there are no rules one way or another (aside from notional unit stuff). It is the hex that is invaded and it is the hex's relationship to sea areas that determines whether the hex can be invaded from sea areas or not.

===
If you and Patrice can work this out, with hopefully the involvement of others (Norman42, Brian Brian, and others: you know who you are), I will defer to your group consensus. Right now I am deep into the code for destroying units - the 27 different places in the sequence of play where that occurs. That requires/consumes most of my time and concentration.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to paulderynck)
Post #: 55
RE: AI for MWIF - Norway - 8/21/2008 9:50:24 PM   
brian brian

 

Posts: 3191
Joined: 11/16/2005
Status: offline
There is a hex in paper WiF:FE that takes some getting used to in this regard, that being Batavia. The port symbol is only one one zone, but you can use the port to enter two zones. How about this for a solution - make Legaspi the example in the tutorial. It could be the new Batavia. Think carefully about trying to explain it to a new player without it looking very contradictory to several other one-zone ports, like Cebu only a few hexes away.

As for drawing the zone boundaries, like I said about Liverpool last year, if it ain't broke, don't fix it. The current paper system works perfectly fine. Why confuse the experienced players to no purpose? I _could_ right-click the map to identify the invadable hexes and then scratch my head about why Liverpool is invadable from the Faeroes Gap. Or I could just look at the boundary drawn to the center of the hex. For new players, however you solve this has to be taught in a tutorial anyway.

Look how this question re-appeared ... over a strategy question after looking at a screen shot. I think something that could be conveyed by a mere visual is far more elegant than making it into something that has to be conveyed by another mouse-click. There are going to be thousands of mouse-clicks in this game already.

(in reply to Shannon V. OKeets)
Post #: 56
RE: AI for MWIF - Norway - 8/21/2008 10:03:07 PM   
Froonp


Posts: 7995
Joined: 10/21/2003
From: Marseilles, France
Status: offline
Steve, in a similar way to what we do for the lake and river graphics, that we draw in hex where the data show no lake or river, wouldn't it be possible to have a CSV map data file for drawing exceptional Sea Area boundaries connections to some exceptional land hexes ?

The CSV file would show the coordinates of the hex, and then a 1-12 number that would "connect" the Sea Area border to a place on the hexagon placed at the 1-12 O'Clock positions on its cirumference (12 being the top position and 6 the bottom one) (similar to how cities, ports, symbols are shown inside the hex, but here it would be on the cirumference).

That way we could have the Sea Area boundary connect to pretty much the place we would prefer.

For example, for Plymouth, I would connect it to position 7 (middle of the SW hexside).

(in reply to brian brian)
Post #: 57
RE: AI for MWIF - Norway - 8/21/2008 10:13:14 PM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: brian brian

There is a hex in paper WiF:FE that takes some getting used to in this regard, that being Batavia. The port symbol is only one one zone, but you can use the port to enter two zones. How about this for a solution - make Legaspi the example in the tutorial. It could be the new Batavia. Think carefully about trying to explain it to a new player without it looking very contradictory to several other one-zone ports, like Cebu only a few hexes away.

As for drawing the zone boundaries, like I said about Liverpool last year, if it ain't broke, don't fix it. The current paper system works perfectly fine. Why confuse the experienced players to no purpose? I _could_ right-click the map to identify the invadable hexes and then scratch my head about why Liverpool is invadable from the Faeroes Gap. Or I could just look at the boundary drawn to the center of the hex. For new players, however you solve this has to be taught in a tutorial anyway.

Look how this question re-appeared ... over a strategy question after looking at a screen shot. I think something that could be conveyed by a mere visual is far more elegant than making it into something that has to be conveyed by another mouse-click. There are going to be thousands of mouse-clicks in this game already.


Ah, but it is broke from my point of view. That is because the drawing of the sea boundaries in WIF FE is done by the hand of an artist - it is not data driven. I have no interest in writing code (or creating a lookup table) so each of the sea area boundaries are drawn "just so".

For instance, I already have 300+ lines of code that detemine which exact pixel within a hex's rectangle is the vertex of the hexagon, for each of the 6 vertices, for each of the 8 zoom levels. That enables the program to draw the hexes so they look like a hexagon lattice. It is really ugly code, but essential to render a hex map. And it has the benefit of working for every hex (there is some exception code for handling the top and bottom of the map).

Drawing each sea area boundary "by hand" for all 8 zoom levels, ain't gonna happen.

Let me reiterate my point with another example, the rivers that are drawn on the map have no effect on game play. They are just pretty pictures. What determines the ability of units to receive the benefits of a river defense is the data file that defines which hexsides have rivers separating two hexes. It took a lot of work by Patrice and myself to make sure that the river pictures and the data file matched (7000+ entries).


_____________________________

Steve

Perfection is an elusive goal.

(in reply to brian brian)
Post #: 58
RE: AI for MWIF - Norway - 8/21/2008 10:18:08 PM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: Froonp

Steve, in a similar way to what we do for the lake and river graphics, that we draw in hex where the data show no lake or river, wouldn't it be possible to have a CSV map data file for drawing exceptional Sea Area boundaries connections to some exceptional land hexes ?

The CSV file would show the coordinates of the hex, and then a 1-12 number that would "connect" the Sea Area border to a place on the hexagon placed at the 1-12 O'Clock positions on its cirumference (12 being the top position and 6 the bottom one) (similar to how cities, ports, symbols are shown inside the hex, but here it would be on the cirumference).

That way we could have the Sea Area boundary connect to pretty much the place we would prefer.

For example, for Plymouth, I would connect it to position 7 (middle of the SW hexside).

If you look at the WIF FE map you will see that the sea area boundaries do more than just 'arrive' at the coastal hex. They also traverse all the open sea hexes and go around islands (e.g., the Pacific map). What you are proposing is a look up table, that would be of monstrous size.

Furthermore, I have no interest in rewriting code that is both complete and accurate just to duplicate the visual of hand drawn graphics.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to Froonp)
Post #: 59
RE: AI for MWIF - Norway - 8/21/2008 10:26:36 PM   
Froonp


Posts: 7995
Joined: 10/21/2003
From: Marseilles, France
Status: offline
quote:

ORIGINAL: Shannon V. OKeets
quote:

ORIGINAL: Froonp

Steve, in a similar way to what we do for the lake and river graphics, that we draw in hex where the data show no lake or river, wouldn't it be possible to have a CSV map data file for drawing exceptional Sea Area boundaries connections to some exceptional land hexes ?

The CSV file would show the coordinates of the hex, and then a 1-12 number that would "connect" the Sea Area border to a place on the hexagon placed at the 1-12 O'Clock positions on its cirumference (12 being the top position and 6 the bottom one) (similar to how cities, ports, symbols are shown inside the hex, but here it would be on the cirumference).

That way we could have the Sea Area boundary connect to pretty much the place we would prefer.

For example, for Plymouth, I would connect it to position 7 (middle of the SW hexside).

If you look at the WIF FE map you will see that the sea area boundaries do more than just 'arrive' at the coastal hex. They also traverse all the open sea hexes and go around islands (e.g., the Pacific map). What you are proposing is a look up table, that would be of monstrous size.

Furthermore, I have no interest in rewriting code that is both complete and accurate just to duplicate the visual of hand drawn graphics.

I am not proposing that the whole Sea Area boundary be defined in a CSV file, only a dozen or so of Sea Area Boundaries connections to selected hexes.

The game would see that Plymouth ought to have a Sea Ara Boundary traced so that it joins its initial at sea starting position to the position 7 on the hexagon of Plymouth. The rest of the Sea Area boundary would not change. It is similar to how railways flow. They are automatic, but in some selected hexes we made it connect to a specific place in the hex.

(in reply to Shannon V. OKeets)
Post #: 60
Page:   <<   < prev  1 [2] 3 4   next >   >>
All Forums >> [New Releases from Matrix Games] >> World in Flames >> AI Opponent Discussion >> RE: AI for MWIF - Norway Page: <<   < prev  1 [2] 3 4   next >   >>
Jump to:





New Messages No New Messages
Hot Topic w/ New Messages Hot Topic w/o New Messages
Locked w/ New Messages Locked w/o New Messages
 Post New Thread
 Reply to Message
 Post New Poll
 Submit Vote
 Delete My Own Post
 Delete My Own Thread
 Rate Posts


Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI

1.406